I've rejoined the Tory Party to partake in this leadership contest.
Truss for me, us republican Tories need to stick together.
How long until you qualify to vote in the leadership election - as that could easily be early January at the present rate.
Three months. But I may get an immediate vote, I was a member for 23 years, and my membership only lapsed last year.
FWIW - I think Boris Johnson is secure until May at least, the locals might be the tipping point if it is a very bad night.
Still has to be a danger that there is a meltdown over COVID restrictions. They kicked the can today, but seems like there isn't much collective responsibility in the cabinet.
Actually I think Boris Johnson has played a blinder today.
If Omicron is as bad as SAGE think it will be, he'll put the blame squarely on the cabinet refuseniks.
Eh ?
He's the PM. If Omicron goes badly, he goes down with the ship.
He will at a general election, but in terms of internal Conservative Party politics, it will benefit him.
At Lingfield this afternoon enjoying a moderate afternoon's jump racing, I saw two drones in the distance presumably carrying real time pictures of the action for exchange players.
I believe there was a bill put forward to legislate against these drones which can be a hazard and a danger - one of the issues is where the landowner has given consent for the operation from their field and garden.
I've not followed the ins and outs but I believe the Government has effectively blocked the proposed legislation by not providing debating time.
Coincidentally, a similar proposal in Ireland looks likely to be passed by the Dáil.
Curiously, I was discussing this with my devil this afternoon. Aircraft have the right to overfly private property by statute but I am not aware of any such provision for drones. They can overfly public land, provided they are not constituting a hazard but not private land. If we are ever to get to the point where Amazon are going to replace white van man with a drone this is going to need to change. Your point suggests how difficult that legislation might be.
Fraser Nelson: "So we have an asteroid that may hit the Earth?" Scientist: "Yes. And we have no idea how likely it is - only that it's heading towards us. NORAD were too busy tracking Santa Claus." Fraser Nelson: "But it may miss." Scientist: "Yes." Fraser Nelson: "So why are you only modelling what will happen if it hits?" Scientist: "Because the decision-makers need to consider what to do if the worst comes to the worst." Fraser Nelson: "But they might not have to do anything if it doesn't hit." Scientist: "But it may. And they need to think about what they'd do." Fraser Nelson: "Why didn't you model the fact it might miss?" Scientist: "Because that doesn't really help the decision-makers."
That Fraser Nelson article in the Spectator is really a whole load of nothing IMO. What the scientist said makes sense.
What a load of crap.
If the government are considering locking us down because of the virus then they need to know what's likely to happen with the virus. If the models say that the NHS isn't likely to be overwhelmed but those models are disregarded in favour of those that say it is, then that's operating with false information.
It's not (see TimT's reply).
It is.
If the government is weighing up their response then they need the full information. ' If the full information says for instance there's a 99.9% chance that the NHS won't be overwhelmed, but there's a 0.1% chance that it is - then do you seriously think the government should only be shown the 0.1% scenario without any qualification of caveat or rating of how likely it is?
They should get the full information, and be allowed to judge with full knowledge whether the risk of these so-called "never events" are worth acting over or not. If they don't have the full information, then they can't weigh that up.
"If the full information says for instance there's a 99.9% chance that the NHS won't be overwhelmed, but there's a 0.1% chance that it is"
But that's not what Nelson was talking about (although he moved onto that at the end). He was talking about a lack of a model that replicated some of JP Morgan's modelling, not the probabilities of any scenario.
Nelson was talking about modelling that used data that the scientists had recognised. But since this model didn't give "the right" answer it was disregarded.
If you decide in advance to disregard all models that don't give a certain outcome, then you've prejudiced your work in advance.
I might suggest you reread what I (and especially TimT) have written.
There is no point in presenting reasonable scenarios where the decision-makers need to do nothing, because that's pointless. The decision-makers need to know the scenarios where they may have to do things, so they can consider them.
The probabilities may come later.
I assumed when Max and others were going on about the Nelson article, that the scientist had said something outrageous. Instead he said something utterly sensible, and Nelson has either misunderstood the point of the modelling, or is deliberately shit-stirring. The scientist should have explained a little better, though.
No this is utterly wrong. If you have a whole range of possible scenarios and you only present the ones that result in action then you are taking the responsibility for decision making away from the politicians and giving it to the unelected scientists. By removing the scenarios where nothing needs to be done you are forcing the politicians into a position where they either do something - even if it is probably unnecessary - or they can be accused of ignoring the evidence as it was presented to them. It is absolutely vital that the advisors present all possible scenarios and weigh them for the decision makers.
It is not utterly wrong. It is presenting the scenarios that need stuff doing about them.
"Present all possible scenarios"
And how many different scenarios is that? Ten? a hundred? A thousand?
The scientists are saying: "these are the sh*t scenarios. These are the ones that, if they occur, we will need to take actions on." Presenting scenarios that need no action is pointless - because there are no actions required. It doesn't mean the politicians are being told these will happen, or that these are the only scenarios. They are the ones to worry about.
It's not that they're saying these are the only scenarios that may happen; or the probabilities of them happening; they're the ones the politicians really need to worry about.
It's simple stuff.
FPT
The problem is that right now people think the most optomistic scenario from SAGE is for a peak of 3000 admissions a day, as that is their lowest model. Except other models suggest this is NOT the lowest possible. The narrative is running that if we do nothing, we will definitely see a minimum of 3000 admissions a day. It’s dishonest. I’m not blaming the scientists entirely here, I want to know what instructions they are getting.
Then I'd blame those telling the people that this is the issue without a fair grasp on what SAGE and SPI-M-O are doing, or their terms of reference, or the specifics of decision making during a period of ignorance but with the possibility of a Never Event (as described by TimT).
That would be the attention-seekers of the media and some Twitter commentators. Who grab stuff, skim it for an attention-grabbing takeaway line, and plaster it everywhere they can.
Why use knowingly incorrect data points for model inputs, though? However you cut it that's just bad practice and now they've managed to lose their last ounce of credibility with the politicians. The fault lies with whoever is asking for them to do that, sure, but also with them for not having the professional integrity to flag that up as bad practice. They're being asked to put their finger on the scale so someone can use their findings to push the lockdown decision making process forwards. As it turned out Cabinet took exception to being presented data that has no basis in reality.
Which knowingly incorrect parameters? Looking at the Warwick paper, they had several possible Ve figures against potential levels of immune escape and booster effectiveness (high to low), and it was written and updated on the 11th of December.
TimT explained the parameters of the decision-making process in conditions of high uncertainty and with credible possibilities of a Never Event, and it follows exactly that process.
So it could either be that process, with the criteria of a Never Event as laid down over a decade ago in the last Labour Government, or its a cabal of evil scientists or a secret civil servant with a fetish for lockdowns who’s not been sprung yet.
I hate lockdowns. I personally incline to the view that there will not be a need for them. But I agree that the probabilities involved are still unclear, and that they should ensure that they have the options on the table, together with what they would or would not do, as well as the costs for them, all clearly laid out.
And, of course, as Paul Mainwood drily observed a few days ago: if we do want to adjust for a variable severity, we can simply do the less-than-arcane procedure of “multiply by a constant” against the outcomes.
For one of the three vaccine efficacy tables, they used the one not based on observation and used the 85% which used an additional 6x reduction in antibody binding efficiency. The second and slightly more contentious one is not using our own study data which gives Omicron a different manifestation of symptoms than Delta, get the central forecast simply assumes that Omicron manifests identical symptoms to Delta.
The issue I have with the kind of modelling here is that Omicron isn't an unknown unknown. We know things about it, we know fairly well what vaccine efficacy is, we know with a reasonable degree of certainty that it manifests as an upper respiratory tract virus than a lower respiratory tract one like Delta.
It's simply not right to ask the decision makers to form an opinion on the right way forwards if the best available data isn't being used and instead guesswork that we know to be worse is being used. I understand the need for expediency and that in a situation where exponential growth is a factor acting late can mean it's too late to act. I do understand that they are under pressure. However, there's simply no need to put their finger on the scales, whoever asked them to do that is trying to force a decision on lockdown and it's spectacularly back fired on them because now the politicians have lost all confidence in the modelling. The 7k central forecast from the summer had already hurt their credibility now it's completely destroyed because they have been found to be fiddling the numbers for some as yet unknown DoH bod who wants to guide the process in a specific direction.
On the whole idea of the "never scenario" it's fanciful that anything that comes with COVID is now in that category.
I also agree that we probably won't need a lockdown. My issue is that the data for that scenario isn't being presented to the politicians when it absolutely should be. It's not an unlikely scenario that Omicron is a paper tiger, yet our politicians are given that option because the unnamed party in the DoH has deemed it irrelevant to the decision making process and asked the modellers not to give their findings for that scenario. It would be a piece of piss to do it as well.
There was never a “7k central forecast” in the summer as anyone who ever read the fucking document in question knew.
I've rejoined the Tory Party to partake in this leadership contest.
Truss for me, us republican Tories need to stick together.
How long until you qualify to vote in the leadership election - as that could easily be early January at the present rate.
Three months. But I may get an immediate vote, I was a member for 23 years, and my membership only lapsed last year.
FWIW - I think Boris Johnson is secure until May at least, the locals might be the tipping point if it is a very bad night.
Still has to be a danger that there is a meltdown over COVID restrictions. They kicked the can today, but seems like there isn't much collective responsibility in the cabinet.
Actually I think Boris Johnson has played a blinder today.
If Omicron is as bad as SAGE think it will be, he'll put the blame squarely on the cabinet refuseniks.
With Covid, it is important to remember one thing. It relies on people participating in the testing system, and self isolating to avoid transmitting the virus. This despite the fact that, for large numbers of people, self isolation is a major deprivation of liberty and financial inconvenience.
Eventually, people will stop submitting to the testing system. And that will be the end of Covid. It becomes something that is just passed on to the healthcare system to deal with.
Looking at some of the comments today, I don't think we are far off this point.
And the irony is they will stop submitting whilst the health experts and politicians demand ever more.
Not enough to test when you're feeling symptoms defined as covid symptoms. Not enough to test when you are feeling any sort of illness at all, even if you're sure it's hayfever because its the middle of summer and the grass pollen is off the scale, it could be covid! Not enough to test twice per week. Not enough to test every day, or every time you leave the house - at some point someone will suggest we all test two or three times per day, I feel sure of it.
They already do! Trafford public health - who I'm sure aren't alone in this - require every household member of a child in education to take an LFT, every day of the Christmas holidays. That's 70 tests over the course of Christmas. They can, quite frankly, fuck off. I don't mind doing an LFT test. I took one the other day when seeing my sister-in-law, who was flying to Australia the following day with her new baby and partner (who had not seen her own parents in two years.) I obviously didn't want to jeopardise that. (It surprised my father in law when my mother in law LFTd him - he had never heard of the process, despite working in a university. ) But testing every day is just overkill.
Fraser Nelson: "So we have an asteroid that may hit the Earth?" Scientist: "Yes. And we have no idea how likely it is - only that it's heading towards us. NORAD were too busy tracking Santa Claus." Fraser Nelson: "But it may miss." Scientist: "Yes." Fraser Nelson: "So why are you only modelling what will happen if it hits?" Scientist: "Because the decision-makers need to consider what to do if the worst comes to the worst." Fraser Nelson: "But they might not have to do anything if it doesn't hit." Scientist: "But it may. And they need to think about what they'd do." Fraser Nelson: "Why didn't you model the fact it might miss?" Scientist: "Because that doesn't really help the decision-makers."
That Fraser Nelson article in the Spectator is really a whole load of nothing IMO. What the scientist said makes sense.
What a load of crap.
If the government are considering locking us down because of the virus then they need to know what's likely to happen with the virus. If the models say that the NHS isn't likely to be overwhelmed but those models are disregarded in favour of those that say it is, then that's operating with false information.
It's not (see TimT's reply).
It is.
If the government is weighing up their response then they need the full information. ' If the full information says for instance there's a 99.9% chance that the NHS won't be overwhelmed, but there's a 0.1% chance that it is - then do you seriously think the government should only be shown the 0.1% scenario without any qualification of caveat or rating of how likely it is?
They should get the full information, and be allowed to judge with full knowledge whether the risk of these so-called "never events" are worth acting over or not. If they don't have the full information, then they can't weigh that up.
"If the full information says for instance there's a 99.9% chance that the NHS won't be overwhelmed, but there's a 0.1% chance that it is"
But that's not what Nelson was talking about (although he moved onto that at the end). He was talking about a lack of a model that replicated some of JP Morgan's modelling, not the probabilities of any scenario.
Nelson was talking about modelling that used data that the scientists had recognised. But since this model didn't give "the right" answer it was disregarded.
If you decide in advance to disregard all models that don't give a certain outcome, then you've prejudiced your work in advance.
I might suggest you reread what I (and especially TimT) have written.
There is no point in presenting reasonable scenarios where the decision-makers need to do nothing, because that's pointless. The decision-makers need to know the scenarios where they may have to do things, so they can consider them.
The probabilities may come later.
I assumed when Max and others were going on about the Nelson article, that the scientist had said something outrageous. Instead he said something utterly sensible, and Nelson has either misunderstood the point of the modelling, or is deliberately shit-stirring. The scientist should have explained a little better, though.
No this is utterly wrong. If you have a whole range of possible scenarios and you only present the ones that result in action then you are taking the responsibility for decision making away from the politicians and giving it to the unelected scientists. By removing the scenarios where nothing needs to be done you are forcing the politicians into a position where they either do something - even if it is probably unnecessary - or they can be accused of ignoring the evidence as it was presented to them. It is absolutely vital that the advisors present all possible scenarios and weigh them for the decision makers.
It is not utterly wrong. It is presenting the scenarios that need stuff doing about them.
"Present all possible scenarios"
And how many different scenarios is that? Ten? a hundred? A thousand?
The scientists are saying: "these are the sh*t scenarios. These are the ones that, if they occur, we will need to take actions on." Presenting scenarios that need no action is pointless - because there are no actions required. It doesn't mean the politicians are being told these will happen, or that these are the only scenarios. They are the ones to worry about.
It's not that they're saying these are the only scenarios that may happen; or the probabilities of them happening; they're the ones the politicians really need to worry about.
It's simple stuff.
FPT
The problem is that right now people think the most optomistic scenario from SAGE is for a peak of 3000 admissions a day, as that is their lowest model. Except other models suggest this is NOT the lowest possible. The narrative is running that if we do nothing, we will definitely see a minimum of 3000 admissions a day. It’s dishonest. I’m not blaming the scientists entirely here, I want to know what instructions they are getting.
Then I'd blame those telling the people that this is the issue without a fair grasp on what SAGE and SPI-M-O are doing, or their terms of reference, or the specifics of decision making during a period of ignorance but with the possibility of a Never Event (as described by TimT).
That would be the attention-seekers of the media and some Twitter commentators. Who grab stuff, skim it for an attention-grabbing takeaway line, and plaster it everywhere they can.
Why use knowingly incorrect data points for model inputs, though? However you cut it that's just bad practice and now they've managed to lose their last ounce of credibility with the politicians. The fault lies with whoever is asking for them to do that, sure, but also with them for not having the professional integrity to flag that up as bad practice. They're being asked to put their finger on the scale so someone can use their findings to push the lockdown decision making process forwards. As it turned out Cabinet took exception to being presented data that has no basis in reality.
Which knowingly incorrect parameters? Looking at the Warwick paper, they had several possible Ve figures against potential levels of immune escape and booster effectiveness (high to low), and it was written and updated on the 11th of December.
TimT explained the parameters of the decision-making process in conditions of high uncertainty and with credible possibilities of a Never Event, and it follows exactly that process.
So it could either be that process, with the criteria of a Never Event as laid down over a decade ago in the last Labour Government, or its a cabal of evil scientists or a secret civil servant with a fetish for lockdowns who’s not been sprung yet.
I hate lockdowns. I personally incline to the view that there will not be a need for them. But I agree that the probabilities involved are still unclear, and that they should ensure that they have the options on the table, together with what they would or would not do, as well as the costs for them, all clearly laid out.
And, of course, as Paul Mainwood drily observed a few days ago: if we do want to adjust for a variable severity, we can simply do the less-than-arcane procedure of “multiply by a constant” against the outcomes.
For one of the three vaccine efficacy tables, they used the one not based on observation and used the 85% which used an additional 6x reduction in antibody binding efficiency. The second and slightly more contentious one is not using our own study data which gives Omicron a different manifestation of symptoms than Delta, get the central forecast simply assumes that Omicron manifests identical symptoms to Delta.
The issue I have with the kind of modelling here is that Omicron isn't an unknown unknown. We know things about it, we know fairly well what vaccine efficacy is, we know with a reasonable degree of certainty that it manifests as an upper respiratory tract virus than a lower respiratory tract one like Delta.
It's simply not right to ask the decision makers to form an opinion on the right way forwards if the best available data isn't being used and instead guesswork that we know to be worse is being used. I understand the need for expediency and that in a situation where exponential growth is a factor acting late can mean it's too late to act. I do understand that they are under pressure. However, there's simply no need to put their finger on the scales, whoever asked them to do that is trying to force a decision on lockdown and it's spectacularly back fired on them because now the politicians have lost all confidence in the modelling. The 7k central forecast from the summer had already hurt their credibility now it's completely destroyed because they have been found to be fiddling the numbers for some as yet unknown DoH bod who wants to guide the process in a specific direction.
On the whole idea of the "never scenario" it's fanciful that anything that comes with COVID is now in that category.
I also agree that we probably won't need a lockdown. My issue is that the data for that scenario isn't being presented to the politicians when it absolutely should be. It's not an unlikely scenario that Omicron is a paper tiger, yet our politicians are given that option because the unnamed party in the DoH has deemed it irrelevant to the decision making process and asked the modellers not to give their findings for that scenario. It would be a piece of piss to do it as well.
There was never a “7k central forecast” in the summer as anyone who ever read the fucking document in question knew.
True, but why have it in there then? Just like these unlikely to pass scenarios all it does is sap their credibility. You've fairly outlined why the modelling isn't bad, yet the politicians are making decisions based on 600-6000 deaths per day.
I've rejoined the Tory Party to partake in this leadership contest.
Truss for me, us republican Tories need to stick together.
How long until you qualify to vote in the leadership election - as that could easily be early January at the present rate.
Three months. But I may get an immediate vote, I was a member for 23 years, and my membership only lapsed last year.
FWIW - I think Boris Johnson is secure until May at least, the locals might be the tipping point if it is a very bad night.
Still has to be a danger that there is a meltdown over COVID restrictions. They kicked the can today, but seems like there isn't much collective responsibility in the cabinet.
Actually I think Boris Johnson has played a blinder today.
If Omicron is as bad as SAGE think it will be, he'll put the blame squarely on the cabinet refuseniks.
According to @JosiasJessop we will never know how bad SAGE think it will be, as they're under no obligation or responsibility to say how it will be.
Their only responsibility apparently is to forward the worst case catastrophising they can come up with, while anything that they think will happen should be binned if it isn't a catastrophe.
So maybe no action is entirely the right response to how bad they think it will be and we should have that attitude towards SAGE indefinitely going forwards because apparently they'll never tell us if things are going to be OK in their expectations.
At this rate, everybody in London will have had Omicron by then.
If more Londoners got vaccinated and had their boosters like the rest of the UK are, then Omicron would be less of a problem in the capital
There was a suggestion the other day - which was my theory - that the reason vaccinations are so low in London is that yhe position estimates are wrong. Many of those people simply aren't there any more. They can't get vaxxed or catch or spread the virus.
Good header, Richard. I don’t believe for a moment that disliking the direction of travel on Covid was the real reason for Frost’s resignation. Rather he realised that he wasn’t going to be able to maintain the bluster any longer.
He has said that pretty clearly and unambiguously. I'd have guessed otherwise, but as he's said it so clearly I think I have to believe what he says.
Many cabinet ministers have said one thing is the reason for their resignation when it's actually another. In any case, normally there's more than one reason. As we saw today, the Cabinet have blocked any restrictions. Moreover, Frost's departure seems to have been agreed some time ago, when this wasn't on the horizon.
What people say and what they actually do are two different things.
Give me 'many' examples then.
Every cabinet minister who ever wrote "to spend more time with my family"?
I mean, these people entered politics specifically to get away from their families.
I thought Fowler coined that one?
And to fair to him, Fiona was giving him shit about him never seeing his kids
I've rejoined the Tory Party to partake in this leadership contest.
Truss for me, us republican Tories need to stick together.
How long until you qualify to vote in the leadership election - as that could easily be early January at the present rate.
Three months. But I may get an immediate vote, I was a member for 23 years, and my membership only lapsed last year.
FWIW - I think Boris Johnson is secure until May at least, the locals might be the tipping point if it is a very bad night.
Still has to be a danger that there is a meltdown over COVID restrictions. They kicked the can today, but seems like there isn't much collective responsibility in the cabinet.
Actually I think Boris Johnson has played a blinder today.
If Omicron is as bad as SAGE think it will be, he'll put the blame squarely on the cabinet refuseniks.
According to @JosiasJessop we will never know how bad SAGE think it will be, as they're under no obligation or responsibility to say how it will be.
Their only responsibility apparently is to forward the worst case catastrophising they can come up with, while anything that they think will happen should be binned if it isn't a catastrophe.
So maybe no action is entirely the right response to how bad they think it will be and we should have that attitude towards SAGE indefinitely going forwards because apparently they'll never tell us if things are going to be OK in their expectations.
I really, really think you ned to go back to the last thread, read, and think. That's not what I'm saying at all.
The wide open stable door is under review, hour by hour, and if someone in the village says they've seen a horse, maybe one that looks like ours, then we won't hesitate to close that stable door in the nick of two weeks too late.
I've rejoined the Tory Party to partake in this leadership contest.
Truss for me, us republican Tories need to stick together.
How long until you qualify to vote in the leadership election - as that could easily be early January at the present rate.
Three months. But I may get an immediate vote, I was a member for 23 years, and my membership only lapsed last year.
FWIW - I think Boris Johnson is secure until May at least, the locals might be the tipping point if it is a very bad night.
Still has to be a danger that there is a meltdown over COVID restrictions. They kicked the can today, but seems like there isn't much collective responsibility in the cabinet.
Actually I think Boris Johnson has played a blinder today.
If Omicron is as bad as SAGE think it will be, he'll put the blame squarely on the cabinet refuseniks.
Eh ?
He's the PM. If Omicron goes badly, he goes down with the ship.
Rationally, I agree.
But this way, if things do go pear-shaped, It's Not Boris's Fault.
Pathetic adolescent attitude, but that's the most useful framework to understand the PM sometimes.
Very good news the Cabinet resisted putting new restrictions in place. I also note we tipped over 50% of 12+ with booster jabs, so that's progressing well.
Do we think they'll try again tomorrow to get Cabinet approval? I struggle to see how sufficient new data will be available then, which would rule out restrictions until after they're able to recall Parliament post-Christmas?
Fingers crossed London (non-incidental) hospitalisation data this week is not alarming, and we may just escape another lockdown.
Maybe, we certainly cannot increase at that rate for long. But in a week or so's time hospital admissions will be at least double what they are now, maybe more.
I don't see how the thread header makes logical sense. The impending collapse of Brexit (which the long-lived amongst us will no doubt still be hearing about in 3022) would affect Truss once she was leader, not before. I have some nice bets on her and it does make some sense to lay it off (if I can work it out) but this thread is by no means a persuasive argument for doing so.
I don't see how the thread header makes logical sense. The impending collapse of Brexit (which the long-lived amongst us will no doubt still be hearing about in 3022) would affect Truss once she was leader, not before. I have some nice bets on her and it does make some sense to lay it off (if I can work it out) but this thread is by no means a persuasive argument for doing so.
You gonna tell us why Boris Johnson is actually a great PM like last time?
I've rejoined the Tory Party to partake in this leadership contest.
Truss for me, us republican Tories need to stick together.
How long until you qualify to vote in the leadership election - as that could easily be early January at the present rate.
Three months. But I may get an immediate vote, I was a member for 23 years, and my membership only lapsed last year.
FWIW - I think Boris Johnson is secure until May at least, the locals might be the tipping point if it is a very bad night.
Still has to be a danger that there is a meltdown over COVID restrictions. They kicked the can today, but seems like there isn't much collective responsibility in the cabinet.
Actually I think Boris Johnson has played a blinder today.
If Omicron is as bad as SAGE think it will be, he'll put the blame squarely on the cabinet refuseniks.
According to @JosiasJessop we will never know how bad SAGE think it will be, as they're under no obligation or responsibility to say how it will be.
Their only responsibility apparently is to forward the worst case catastrophising they can come up with, while anything that they think will happen should be binned if it isn't a catastrophe.
So maybe no action is entirely the right response to how bad they think it will be and we should have that attitude towards SAGE indefinitely going forwards because apparently they'll never tell us if things are going to be OK in their expectations.
I really, really think you ned to go back to the last thread, read, and think. That's not what I'm saying at all.
Please don't misrepresent me.
You were saying, were you not, that only the catastrophic scenarios should be presented?
The scenario of what they think will happen, using the most accurate data available, as explained well by @MaxPB above was not in the SAGE scenarios. Why not? Why should they not present what they think will happen, using the best available data?
There isn't one "never event" in question here, there's two. And under the principle that advisors advise and ministers decide, the best available modelling should be presented without prejudice to what outcome it gives so that the ministers can make an informed decision.
If you wish to provide worst case scenarios AS WELL as central scenarios that's entirely reasonable. And since we have two worst cases here (an NHS collapse, and an unnecessary lockdown) it would make sense to show what assumptions are required for scenarios for both of those.
Instead the results from one catastrophe has been cherry picked without the central scenarios being shown, or the other side.
Maybe, we certainly cannot increase at that rate for long. But in a week or so's time hospital admissions will be at least double what they are now, maybe more.
Will they ?
It depends on who is getting infected.
And every day more vaccines come into effect.
What we can say is likely is that Omicron will burn through the anti-vaxxers quickly.
For one of the three vaccine efficacy tables, they used the one not based on observation and used the 85% which used an additional 6x reduction in antibody binding efficiency. The second and slightly more contentious one is not using our own study data which gives Omicron a different manifestation of symptoms than Delta, get the central forecast simply assumes that Omicron manifests identical symptoms to Delta.
In slightly slower time: the Ve figures you refer to were the ones published in the UKHSA report on the 16th, I presume? If so, there is an obvious reason why they would not be included in a report of the 11th of December, which instead presented a range of possibilities. Which did include the one that actually came up, I note.
The issue I have with the kind of modelling here is that Omicron isn't an unknown unknown. We know things about it, we know fairly well what vaccine efficacy is, we know with a reasonable degree of certainty that it manifests as an upper respiratory tract virus than a lower respiratory tract one like Delta.
I also noted that: 1 - The findings on it replicating in the bronchi rather than the lungs were published a few days ago. Once more, well after the 11th. 2 - We still don’t know the effective outcome of that; we have some ideas but not solid facts
It's simply not right to ask the decision makers to form an opinion on the right way forwards if the best available data isn't being used and instead guesswork that we know to be worse is being used. I understand the need for expediency and that in a situation where exponential growth is a factor acting late can mean it's too late to act. I do understand that they are under pressure. However, there's simply no need to put their finger on the scales, whoever asked them to do that is trying to force a decision on lockdown and it's spectacularly back fired on them because now the politicians have lost all confidence in the modelling. The 7k central forecast from the summer had already hurt their credibility now it's completely destroyed because they have been found to be fiddling the numbers for some as yet unknown DoH bod who wants to guide the process in a specific direction.
It’s also implausible to ask for data from the 16th and 18th of December to be in a report dated the 11th. One would assume that newer information is also mentioned during the briefings.
And anyone claiming “7k central forecast” can only be getting the information from media and commentators rather than actually reading the reports in question. Otherwise they’d know what was actually presented and projected. (In summary: a 1.1 R value as the most likely under many scenarios (the lowest figure), an included 1.5 R value projection, as Scotland had seen such an R value in the several days just leading up to the date of presentation, with this being the highest plausible route (and the one quoted by the Mail, Christopher Snowdon, and others as “central projection), and by obvious request, an R value of 2.0 which they dismissed as extremely unlikely except if waning was far greater than expected or for an unknown variant.
I also agree that we probably won't need a lockdown. My issue is that the data for that scenario isn't being presented to the politicians when it absolutely should be. It's not an unlikely scenario that Omicron is a paper tiger, yet our politicians are given that option because the unnamed party in the DoH has deemed it irrelevant to the decision making process and asked the modellers not to give their findings for that scenario. It would be a piece of piss to do it as well.
It’s been explained that it’s not rogue scientists trying to hijack the nation, nor deep cover civil servants in the Health Department with a fetish for lockdowns dictating the parameters. But it’s obviously no point in discussing further.
At this rate, everybody in London will have had Omicron by then.
If more Londoners got vaccinated and had their boosters like the rest of the UK are, then Omicron would be less of a problem in the capital
It must be a concern for you and your family, for whom London would normally be somewhere you would regularly visit. Do you still go into town, or stay safely in Essex?
Fraser Nelson: "So we have an asteroid that may hit the Earth?" Scientist: "Yes. And we have no idea how likely it is - only that it's heading towards us. NORAD were too busy tracking Santa Claus." Fraser Nelson: "But it may miss." Scientist: "Yes." Fraser Nelson: "So why are you only modelling what will happen if it hits?" Scientist: "Because the decision-makers need to consider what to do if the worst comes to the worst." Fraser Nelson: "But they might not have to do anything if it doesn't hit." Scientist: "But it may. And they need to think about what they'd do." Fraser Nelson: "Why didn't you model the fact it might miss?" Scientist: "Because that doesn't really help the decision-makers."
That Fraser Nelson article in the Spectator is really a whole load of nothing IMO. What the scientist said makes sense.
What a load of crap.
If the government are considering locking us down because of the virus then they need to know what's likely to happen with the virus. If the models say that the NHS isn't likely to be overwhelmed but those models are disregarded in favour of those that say it is, then that's operating with false information.
It's not (see TimT's reply).
It is.
If the government is weighing up their response then they need the full information. ' If the full information says for instance there's a 99.9% chance that the NHS won't be overwhelmed, but there's a 0.1% chance that it is - then do you seriously think the government should only be shown the 0.1% scenario without any qualification of caveat or rating of how likely it is?
They should get the full information, and be allowed to judge with full knowledge whether the risk of these so-called "never events" are worth acting over or not. If they don't have the full information, then they can't weigh that up.
"If the full information says for instance there's a 99.9% chance that the NHS won't be overwhelmed, but there's a 0.1% chance that it is"
But that's not what Nelson was talking about (although he moved onto that at the end). He was talking about a lack of a model that replicated some of JP Morgan's modelling, not the probabilities of any scenario.
Nelson was talking about modelling that used data that the scientists had recognised. But since this model didn't give "the right" answer it was disregarded.
If you decide in advance to disregard all models that don't give a certain outcome, then you've prejudiced your work in advance.
I might suggest you reread what I (and especially TimT) have written.
There is no point in presenting reasonable scenarios where the decision-makers need to do nothing, because that's pointless. The decision-makers need to know the scenarios where they may have to do things, so they can consider them.
The probabilities may come later.
I assumed when Max and others were going on about the Nelson article, that the scientist had said something outrageous. Instead he said something utterly sensible, and Nelson has either misunderstood the point of the modelling, or is deliberately shit-stirring. The scientist should have explained a little better, though.
No this is utterly wrong. If you have a whole range of possible scenarios and you only present the ones that result in action then you are taking the responsibility for decision making away from the politicians and giving it to the unelected scientists. By removing the scenarios where nothing needs to be done you are forcing the politicians into a position where they either do something - even if it is probably unnecessary - or they can be accused of ignoring the evidence as it was presented to them. It is absolutely vital that the advisors present all possible scenarios and weigh them for the decision makers.
It is not utterly wrong. It is presenting the scenarios that need stuff doing about them.
"Present all possible scenarios"
And how many different scenarios is that? Ten? a hundred? A thousand?
The scientists are saying: "these are the sh*t scenarios. These are the ones that, if they occur, we will need to take actions on." Presenting scenarios that need no action is pointless - because there are no actions required. It doesn't mean the politicians are being told these will happen, or that these are the only scenarios. They are the ones to worry about.
It's not that they're saying these are the only scenarios that may happen; or the probabilities of them happening; they're the ones the politicians really need to worry about.
It's simple stuff.
FPT
The problem is that right now people think the most optomistic scenario from SAGE is for a peak of 3000 admissions a day, as that is their lowest model. Except other models suggest this is NOT the lowest possible. The narrative is running that if we do nothing, we will definitely see a minimum of 3000 admissions a day. It’s dishonest. I’m not blaming the scientists entirely here, I want to know what instructions they are getting.
Then I'd blame those telling the people that this is the issue without a fair grasp on what SAGE and SPI-M-O are doing, or their terms of reference, or the specifics of decision making during a period of ignorance but with the possibility of a Never Event (as described by TimT).
That would be the attention-seekers of the media and some Twitter commentators. Who grab stuff, skim it for an attention-grabbing takeaway line, and plaster it everywhere they can.
Why use knowingly incorrect data points for model inputs, though? However you cut it that's just bad practice and now they've managed to lose their last ounce of credibility with the politicians. The fault lies with whoever is asking for them to do that, sure, but also with them for not having the professional integrity to flag that up as bad practice. They're being asked to put their finger on the scale so someone can use their findings to push the lockdown decision making process forwards. As it turned out Cabinet took exception to being presented data that has no basis in reality.
Which knowingly incorrect parameters? Looking at the Warwick paper, they had several possible Ve figures against potential levels of immune escape and booster effectiveness (high to low), and it was written and updated on the 11th of December.
TimT explained the parameters of the decision-making process in conditions of high uncertainty and with credible possibilities of a Never Event, and it follows exactly that process.
So it could either be that process, with the criteria of a Never Event as laid down over a decade ago in the last Labour Government, or its a cabal of evil scientists or a secret civil servant with a fetish for lockdowns who’s not been sprung yet.
I hate lockdowns. I personally incline to the view that there will not be a need for them. But I agree that the probabilities involved are still unclear, and that they should ensure that they have the options on the table, together with what they would or would not do, as well as the costs for them, all clearly laid out.
And, of course, as Paul Mainwood drily observed a few days ago: if we do want to adjust for a variable severity, we can simply do the less-than-arcane procedure of “multiply by a constant” against the outcomes.
For one of the three vaccine efficacy tables, they used the one not based on observation and used the 85% which used an additional 6x reduction in antibody binding efficiency. The second and slightly more contentious one is not using our own study data which gives Omicron a different manifestation of symptoms than Delta, get the central forecast simply assumes that Omicron manifests identical symptoms to Delta.
The issue I have with the kind of modelling here is that Omicron isn't an unknown unknown. We know things about it, we know fairly well what vaccine efficacy is, we know with a reasonable degree of certainty that it manifests as an upper respiratory tract virus than a lower respiratory tract one like Delta.
It's simply not right to ask the decision makers to form an opinion on the right way forwards if the best available data isn't being used and instead guesswork that we know to be worse is being used. I understand the need for expediency and that in a situation where exponential growth is a factor acting late can mean it's too late to act. I do understand that they are under pressure. However, there's simply no need to put their finger on the scales, whoever asked them to do that is trying to force a decision on lockdown and it's spectacularly back fired on them because now the politicians have lost all confidence in the modelling. The 7k central forecast from the summer had already hurt their credibility now it's completely destroyed because they have been found to be fiddling the numbers for some as yet unknown DoH bod who wants to guide the process in a specific direction.
On the whole idea of the "never scenario" it's fanciful that anything that comes with COVID is now in that category.
I also agree that we probably won't need a lockdown. My issue is that the data for that scenario isn't being presented to the politicians when it absolutely should be. It's not an unlikely scenario that Omicron is a paper tiger, yet our politicians are given that option because the unnamed party in the DoH has deemed it irrelevant to the decision making process and asked the modellers not to give their findings for that scenario. It would be a piece of piss to do it as well.
There was never a “7k central forecast” in the summer as anyone who ever read the fucking document in question knew.
True, but why have it in there then? Just like these unlikely to pass scenarios all it does is sap their credibility. You've fairly outlined why the modelling isn't bad, yet the politicians are making decisions based on 600-6000 deaths per day.
It was in there, as it made clear, because they’d just had a surge of R=1.5 in Scotland and therefore couldn’t rule it out as a scenario for the UK as a whole. So they put it as the highest plausible scenario.
Considering it plausible because it had literally just been happening in Scotland, although they were pointing more at the R=1.1 route.
As they said at the time and in the explanation of what each of the scenarios were.
I get why the media ignored the explanation (attention-grabbing) but if politicians who are being presented with the scenarios and having the experts answer questions on hand and still be unable to understand, then such idiots shouldn’t be anywhere near any position of power or decision-making.
NEW: Matt Hancock is in the cheese and wine photo from the Downing Street garden, Telegraph can reveal.
As mentioned earlier, all this talk of cheese and wine is making me think of Mike Leigh's Abigail's Party. The only possible candidate for Beverley is Priti Patel, in some sort of shockingly loud dress and offering cocktails and cheese sticks, and the taciturn man stewing angrily in the corner would have to be Lord Frost.
With Covid, it is important to remember one thing. It relies on people participating in the testing system, and self isolating to avoid transmitting the virus. This despite the fact that, for large numbers of people, self isolation is a major deprivation of liberty and financial inconvenience.
Eventually, people will stop submitting to the testing system. And that will be the end of Covid. It becomes something that is just passed on to the healthcare system to deal with.
Looking at some of the comments today, I don't think we are far off this point.
And the irony is they will stop submitting whilst the health experts and politicians demand ever more.
Not enough to test when you're feeling symptoms defined as covid symptoms. Not enough to test when you are feeling any sort of illness at all, even if you're sure it's hayfever because its the middle of summer and the grass pollen is off the scale, it could be covid! Not enough to test twice per week. Not enough to test every day, or every time you leave the house - at some point someone will suggest we all test two or three times per day, I feel sure of it.
They already do! Trafford public health - who I'm sure aren't alone in this - require every household member of a child in education to take an LFT, every day of the Christmas holidays. That's 70 tests over the course of Christmas. They can, quite frankly, fuck off. I don't mind doing an LFT test. I took one the other day when seeing my sister-in-law, who was flying to Australia the following day with her new baby and partner (who had not seen her own parents in two years.) I obviously didn't want to jeopardise that. (It surprised my father in law when my mother in law LFTd him - he had never heard of the process, despite working in a university. ) But testing every day is just overkill.
I can understand your feelings. Taking a regular LFT to protect family and friends is one thing. Taking one to satisfy some bureaucrats and teachers union reps is totally unacceptable.
I've rejoined the Tory Party to partake in this leadership contest.
Truss for me, us republican Tories need to stick together.
How long until you qualify to vote in the leadership election - as that could easily be early January at the present rate.
Three months. But I may get an immediate vote, I was a member for 23 years, and my membership only lapsed last year.
FWIW - I think Boris Johnson is secure until May at least, the locals might be the tipping point if it is a very bad night.
Still has to be a danger that there is a meltdown over COVID restrictions. They kicked the can today, but seems like there isn't much collective responsibility in the cabinet.
Actually I think Boris Johnson has played a blinder today.
If Omicron is as bad as SAGE think it will be, he'll put the blame squarely on the cabinet refuseniks.
According to @JosiasJessop we will never know how bad SAGE think it will be, as they're under no obligation or responsibility to say how it will be.
Their only responsibility apparently is to forward the worst case catastrophising they can come up with, while anything that they think will happen should be binned if it isn't a catastrophe.
So maybe no action is entirely the right response to how bad they think it will be and we should have that attitude towards SAGE indefinitely going forwards because apparently they'll never tell us if things are going to be OK in their expectations.
I really, really think you ned to go back to the last thread, read, and think. That's not what I'm saying at all.
Please don't misrepresent me.
You were saying, were you not, that only the catastrophic scenarios should be presented?
The scenario of what they think will happen, using the most accurate data available, as explained well by @MaxPB above was not in the SAGE scenarios. Why not? Why should they not present what they think will happen, using the best available data?
There isn't one "never event" in question here, there's two. And under the principle that advisors advise and ministers decide, the best available modelling should be presented without prejudice to what outcome it gives so that the ministers can make an informed decision.
If you wish to provide worst case scenarios AS WELL as central scenarios that's entirely reasonable. And since we have two worst cases here (an NHS collapse, and an unnecessary lockdown) it would make sense to show what assumptions are required for scenarios for both of those.
Instead the results from one catastrophe has been cherry picked without the central scenarios being shown, or the other side.
You and Max keep using the word "scenarios" wrong (or at least in a sense I have never seen it used, in modelling terms). The "scenarios" are things like: Do Nothing, Impose Some Restrictions, Full Lockdown. The complaint you are making seems to be that the model input parameters are biased towards pessimism, because the modellers are excluding data sources which would result in more optimistic selections.
There is no value in modelling scenarios in which we remove restrictions, because no-one is suggesting that. There is a good argument (an excellent one, actually) that we should be modelling based on realistic, rather than pessimistic, parameters, because otherwise we are overstating the chances of Bad Outcomes, and hence giving them too much weight when making policy decisions. However, it is wrong to say that they aren't capturing non-catastrophic scenarios, because the models are stochastic by nature, and are designed to model all possible future outcomes. In addition, if the secondary uncertainty around the mean parameters is high enough, you are also capturing the possibility that the more favourable view of the parameters is actually true. It's just that, because the mean parameters are artificially pessimistic, the midpoint of the ranges produced, and - especially - the worst case outcomes, will look worse than if you picked something more "realistic".
I've rejoined the Tory Party to partake in this leadership contest.
Truss for me, us republican Tories need to stick together.
How long until you qualify to vote in the leadership election - as that could easily be early January at the present rate.
Three months. But I may get an immediate vote, I was a member for 23 years, and my membership only lapsed last year.
FWIW - I think Boris Johnson is secure until May at least, the locals might be the tipping point if it is a very bad night.
Still has to be a danger that there is a meltdown over COVID restrictions. They kicked the can today, but seems like there isn't much collective responsibility in the cabinet.
Actually I think Boris Johnson has played a blinder today.
If Omicron is as bad as SAGE think it will be, he'll put the blame squarely on the cabinet refuseniks.
According to @JosiasJessop we will never know how bad SAGE think it will be, as they're under no obligation or responsibility to say how it will be.
Their only responsibility apparently is to forward the worst case catastrophising they can come up with, while anything that they think will happen should be binned if it isn't a catastrophe.
So maybe no action is entirely the right response to how bad they think it will be and we should have that attitude towards SAGE indefinitely going forwards because apparently they'll never tell us if things are going to be OK in their expectations.
I really, really think you ned to go back to the last thread, read, and think. That's not what I'm saying at all.
Please don't misrepresent me.
You were saying, were you not, that only the catastrophic scenarios should be presented?
The scenario of what they think will happen, using the most accurate data available, as explained well by @MaxPB above was not in the SAGE scenarios. Why not? Why should they not present what they think will happen, using the best available data?
There isn't one "never event" in question here, there's two. And under the principle that advisors advise and ministers decide, the best available modelling should be presented without prejudice to what outcome it gives so that the ministers can make an informed decision.
If you wish to provide worst case scenarios AS WELL as central scenarios that's entirely reasonable. And since we have two worst cases here (an NHS collapse, and an unnecessary lockdown) it would make sense to show what assumptions are required for scenarios for both of those.
Instead the results from one catastrophe has been cherry picked without the central scenarios being shown, or the other side.
No, that's not what I said. And I really, really, really, think you need to go back and *think* about what I, TimT, Andy Cooke (I think) and others wrote in opposition to you.
As for 'never' events: that depends on perspective. Someone who doesn't care how many people die from Covid as long as they can lead a normal life might have a different view from someone who thinks it is wise to try to reduce deaths where possible - and the costs of saving those lives should be weighed up against those lives. But if you go in saying we shouldn't be trying to save any lives, then that's a bogus calculation.
NEW: Matt Hancock is in the cheese and wine photo from the Downing Street garden, Telegraph can reveal.
As mentioned earlier, all this talk of cheese and wine is making me think of Mike Leigh's Abigail's Party. The only possible candidate for Beverley is Priti Patel, in some sort of shockingly loud dress and offering cocktails and cheese sticks, and the taciturn man stewing angrily in the corner would have to be Lord Frost.
At this rate, everybody in London will have had Omicron by then.
If more Londoners got vaccinated and had their boosters like the rest of the UK are, then Omicron would be less of a problem in the capital
It must be a concern for you and your family, for whom London would normally be somewhere you would regularly visit. Do you still go into town, or stay safely in Essex?
My wife and I have both been double jabbed and had our boosters, we are not too bothered, indeed we went up the Shard last night and for cocktails and snacks at the Shangri-La hotel. Albeit normally I work in London but have only been going in a couple of times a month and now fulltime wfh again.
My sister, who lives in outer London in Beckenham, has unfortunately caught Omicron and been isolating but has also had her booster and hopes to join us for boxing day. We will be spending Christmas with our parents in Kent
I've rejoined the Tory Party to partake in this leadership contest.
Truss for me, us republican Tories need to stick together.
How long until you qualify to vote in the leadership election - as that could easily be early January at the present rate.
Three months. But I may get an immediate vote, I was a member for 23 years, and my membership only lapsed last year.
FWIW - I think Boris Johnson is secure until May at least, the locals might be the tipping point if it is a very bad night.
Still has to be a danger that there is a meltdown over COVID restrictions. They kicked the can today, but seems like there isn't much collective responsibility in the cabinet.
Actually I think Boris Johnson has played a blinder today.
If Omicron is as bad as SAGE think it will be, he'll put the blame squarely on the cabinet refuseniks.
According to @JosiasJessop we will never know how bad SAGE think it will be, as they're under no obligation or responsibility to say how it will be.
Their only responsibility apparently is to forward the worst case catastrophising they can come up with, while anything that they think will happen should be binned if it isn't a catastrophe.
So maybe no action is entirely the right response to how bad they think it will be and we should have that attitude towards SAGE indefinitely going forwards because apparently they'll never tell us if things are going to be OK in their expectations.
I really, really think you ned to go back to the last thread, read, and think. That's not what I'm saying at all.
Please don't misrepresent me.
You were saying, were you not, that only the catastrophic scenarios should be presented?
The scenario of what they think will happen, using the most accurate data available, as explained well by @MaxPB above was not in the SAGE scenarios. Why not? Why should they not present what they think will happen, using the best available data?
There isn't one "never event" in question here, there's two. And under the principle that advisors advise and ministers decide, the best available modelling should be presented without prejudice to what outcome it gives so that the ministers can make an informed decision.
If you wish to provide worst case scenarios AS WELL as central scenarios that's entirely reasonable. And since we have two worst cases here (an NHS collapse, and an unnecessary lockdown) it would make sense to show what assumptions are required for scenarios for both of those.
Instead the results from one catastrophe has been cherry picked without the central scenarios being shown, or the other side.
You and Max keep using the word "scenarios" wrong (or at least in a sense I have never seen it used, in modelling terms). The "scenarios" are things like: Do Nothing, Impose Some Restrictions, Full Lockdown. The complaint you are making seems to be that the model input parameters are biased towards pessimism, because the modellers are excluding data sources which would result in more optimistic selections.
There is no value in modelling scenarios in which we remove restrictions, because no-one is suggesting that. There is a good argument (an excellent one, actually) that we should be modelling based on realistic, rather than pessimistic, parameters, because otherwise we are overstating the chances of Bad Outcomes, and hence giving them too much weight when making policy decisions. However, it is wrong to say that they aren't capturing non-catastrophic scenarios, because the models are stochastic by nature, and are designed to model all possible future outcomes. In addition, if the secondary uncertainty around the mean parameters is high enough, you are also capturing the possibility that the more favourable view of the parameters is actually true. It's just that, because the mean parameters are artificially pessimistic, the midpoint of the ranges produced, and - especially - the worst case outcomes, will look worse than if you picked something more "realistic".
The advisors should present all possible likely outcomes of their models including ones where no further action is needed or restrictions are completely lifted - so long as any of their models indicate these outcomes. They should advise which of the modelled outcomes are more likely or less likely but they should certainly not be excluding outcomes because they are not worst case. It is for elected politicians to make decisions about which models they wish to consider but they need to be shown the full selection and then make an informed decision on that basis.
Only showing the worst outcomes is the scientists forcing the politicians to follow a particular preferred route even if that outcome is unlikely and irrespective of what the costs are to the country. It is a ludicrous way to manage a pandemic and all it will do is destroy any last vestiges of trust in the modelling or the advice.
At this rate, everybody in London will have had Omicron by then.
If more Londoners got vaccinated and had their boosters like the rest of the UK are, then Omicron would be less of a problem in the capital
It must be a concern for you and your family, for whom London would normally be somewhere you would regularly visit. Do you still go into town, or stay safely in Essex?
My wife and I have both been double jabbed and had our boosters, we are not too bothered, indeed we went up the Shard last night and for cocktails and snacks at the Shangri-La hotel. Albeit normally I work in London but have only been going in a couple of times a month and now fulltime wfh again.
My sister, who lives in outer London in Beckenham, has unfortunately caught Omicron and been isolating but has also had her booster and hopes to join us for boxing day. We will be spending Christmas with our parents in Kent
Stay safe and Happy Christmas to you and yours (and to all PBers).
What better time to announce them than at midnight?
Visions of the Drake emerging from his coffin, blood dripping from his fangs, with a maniacal laugh. Hah hah hah hah hah, you’re all locked down and can’t escape!
Fraser Nelson: "So we have an asteroid that may hit the Earth?" Scientist: "Yes. And we have no idea how likely it is - only that it's heading towards us. NORAD were too busy tracking Santa Claus." Fraser Nelson: "But it may miss." Scientist: "Yes." Fraser Nelson: "So why are you only modelling what will happen if it hits?" Scientist: "Because the decision-makers need to consider what to do if the worst comes to the worst." Fraser Nelson: "But they might not have to do anything if it doesn't hit." Scientist: "But it may. And they need to think about what they'd do." Fraser Nelson: "Why didn't you model the fact it might miss?" Scientist: "Because that doesn't really help the decision-makers."
That Fraser Nelson article in the Spectator is really a whole load of nothing IMO. What the scientist said makes sense.
What a load of crap.
If the government are considering locking us down because of the virus then they need to know what's likely to happen with the virus. If the models say that the NHS isn't likely to be overwhelmed but those models are disregarded in favour of those that say it is, then that's operating with false information.
It's not (see TimT's reply).
It is.
If the government is weighing up their response then they need the full information. ' If the full information says for instance there's a 99.9% chance that the NHS won't be overwhelmed, but there's a 0.1% chance that it is - then do you seriously think the government should only be shown the 0.1% scenario without any qualification of caveat or rating of how likely it is?
They should get the full information, and be allowed to judge with full knowledge whether the risk of these so-called "never events" are worth acting over or not. If they don't have the full information, then they can't weigh that up.
"If the full information says for instance there's a 99.9% chance that the NHS won't be overwhelmed, but there's a 0.1% chance that it is"
But that's not what Nelson was talking about (although he moved onto that at the end). He was talking about a lack of a model that replicated some of JP Morgan's modelling, not the probabilities of any scenario.
Nelson was talking about modelling that used data that the scientists had recognised. But since this model didn't give "the right" answer it was disregarded.
If you decide in advance to disregard all models that don't give a certain outcome, then you've prejudiced your work in advance.
I might suggest you reread what I (and especially TimT) have written.
There is no point in presenting reasonable scenarios where the decision-makers need to do nothing, because that's pointless. The decision-makers need to know the scenarios where they may have to do things, so they can consider them.
The probabilities may come later.
I assumed when Max and others were going on about the Nelson article, that the scientist had said something outrageous. Instead he said something utterly sensible, and Nelson has either misunderstood the point of the modelling, or is deliberately shit-stirring. The scientist should have explained a little better, though.
No this is utterly wrong. If you have a whole range of possible scenarios and you only present the ones that result in action then you are taking the responsibility for decision making away from the politicians and giving it to the unelected scientists. By removing the scenarios where nothing needs to be done you are forcing the politicians into a position where they either do something - even if it is probably unnecessary - or they can be accused of ignoring the evidence as it was presented to them. It is absolutely vital that the advisors present all possible scenarios and weigh them for the decision makers.
It is not utterly wrong. It is presenting the scenarios that need stuff doing about them.
"Present all possible scenarios"
And how many different scenarios is that? Ten? a hundred? A thousand?
The scientists are saying: "these are the sh*t scenarios. These are the ones that, if they occur, we will need to take actions on." Presenting scenarios that need no action is pointless - because there are no actions required. It doesn't mean the politicians are being told these will happen, or that these are the only scenarios. They are the ones to worry about.
It's not that they're saying these are the only scenarios that may happen; or the probabilities of them happening; they're the ones the politicians really need to worry about.
It's simple stuff.
FPT
The problem is that right now people think the most optomistic scenario from SAGE is for a peak of 3000 admissions a day, as that is their lowest model. Except other models suggest this is NOT the lowest possible. The narrative is running that if we do nothing, we will definitely see a minimum of 3000 admissions a day. It’s dishonest. I’m not blaming the scientists entirely here, I want to know what instructions they are getting.
Then I'd blame those telling the people that this is the issue without a fair grasp on what SAGE and SPI-M-O are doing, or their terms of reference, or the specifics of decision making during a period of ignorance but with the possibility of a Never Event (as described by TimT).
That would be the attention-seekers of the media and some Twitter commentators. Who grab stuff, skim it for an attention-grabbing takeaway line, and plaster it everywhere they can.
Why use knowingly incorrect data points for model inputs, though? However you cut it that's just bad practice and now they've managed to lose their last ounce of credibility with the politicians. The fault lies with whoever is asking for them to do that, sure, but also with them for not having the professional integrity to flag that up as bad practice. They're being asked to put their finger on the scale so someone can use their findings to push the lockdown decision making process forwards. As it turned out Cabinet took exception to being presented data that has no basis in reality.
Which knowingly incorrect parameters? Looking at the Warwick paper, they had several possible Ve figures against potential levels of immune escape and booster effectiveness (high to low), and it was written and updated on the 11th of December.
TimT explained the parameters of the decision-making process in conditions of high uncertainty and with credible possibilities of a Never Event, and it follows exactly that process.
So it could either be that process, with the criteria of a Never Event as laid down over a decade ago in the last Labour Government, or its a cabal of evil scientists or a secret civil servant with a fetish for lockdowns who’s not been sprung yet.
I hate lockdowns. I personally incline to the view that there will not be a need for them. But I agree that the probabilities involved are still unclear, and that they should ensure that they have the options on the table, together with what they would or would not do, as well as the costs for them, all clearly laid out.
And, of course, as Paul Mainwood drily observed a few days ago: if we do want to adjust for a variable severity, we can simply do the less-than-arcane procedure of “multiply by a constant” against the outcomes.
For one of the three vaccine efficacy tables, they used the one not based on observation and used the 85% which used an additional 6x reduction in antibody binding efficiency. The second and slightly more contentious one is not using our own study data which gives Omicron a different manifestation of symptoms than Delta, get the central forecast simply assumes that Omicron manifests identical symptoms to Delta.
The issue I have with the kind of modelling here is that Omicron isn't an unknown unknown. We know things about it, we know fairly well what vaccine efficacy is, we know with a reasonable degree of certainty that it manifests as an upper respiratory tract virus than a lower respiratory tract one like Delta.
It's simply not right to ask the decision makers to form an opinion on the right way forwards if the best available data isn't being used and instead guesswork that we know to be worse is being used. I understand the need for expediency and that in a situation where exponential growth is a factor acting late can mean it's too late to act. I do understand that they are under pressure. However, there's simply no need to put their finger on the scales, whoever asked them to do that is trying to force a decision on lockdown and it's spectacularly back fired on them because now the politicians have lost all confidence in the modelling. The 7k central forecast from the summer had already hurt their credibility now it's completely destroyed because they have been found to be fiddling the numbers for some as yet unknown DoH bod who wants to guide the process in a specific direction.
On the whole idea of the "never scenario" it's fanciful that anything that comes with COVID is now in that category.
I also agree that we probably won't need a lockdown. My issue is that the data for that scenario isn't being presented to the politicians when it absolutely should be. It's not an unlikely scenario that Omicron is a paper tiger, yet our politicians are given that option because the unnamed party in the DoH has deemed it irrelevant to the decision making process and asked the modellers not to give their findings for that scenario. It would be a piece of piss to do it as well.
There was never a “7k central forecast” in the summer as anyone who ever read the fucking document in question knew.
True, but why have it in there then? Just like these unlikely to pass scenarios all it does is sap their credibility. You've fairly outlined why the modelling isn't bad, yet the politicians are making decisions based on 600-6000 deaths per day.
It was in there, as it made clear, because they’d just had a surge of R=1.5 in Scotland and therefore couldn’t rule it out as a scenario for the UK as a whole. So they put it as the highest plausible scenario.
Considering it plausible because it had literally just been happening in Scotland, although they were pointing more at the R=1.1 route.
As they said at the time and in the explanation of what each of the scenarios were.
I get why the media ignored the explanation (attention-grabbing) but if politicians who are being presented with the scenarios and having the experts answer questions on hand and still be unable to understand, then such idiots shouldn’t be anywhere near any position of power or decision-making.
But a sustained 1.1 also didn't happen and a sustained 1.5 was never plausible, it was, in fact, implausible yet presented as something to plan against.
We're lucky that the politicians ignored the summer forecasts because they'd have insisted on silly measures like indoor distancing and vaccine passports which have been of no help in Europe and in fact may have been counter productive as it hasn't allowed them to build a natural immunity wall in the unvaccinated.
As to your other points, your faith and trust in people who you have no possible way of removing from the levers of power is worrying. Politicians may be flawed (and this set particularly so) yet we still have some way of making the decision on whether or not we want to do what they say. The unnamed bod in the DoH putting his fingers on the scale and presenting the politicians with a range unlikely scenarios and not also presenting the likely scenarios, well we have no way of removing this person from power, no way of holding them accountable to us.
That makes me instinctively wary of of them and sceptical about what their motives are. I know what Labour's motives are, I know what the Tories motives are, they told me at length in ab election campaign. We have an unelected tyranny of experts, maybe you're happy to live like that, I'm not. Quite frankly, our views on this differ too much, I'm extremely worried that the UK and its people are too happy to simply hand our freedoms to these unknown experts with no guarantees they'll be happy to give them back again.
So yes, colour me very sceptical about anyone who is pushing a very narrow set of possible outcomes while excluding other more favourable ones. There's simply no need for it.
As to the other point, I understand the data came later than the report, that's a reason to tell the decision makers - "we have new key data inputs for our model, please await our new results while we rerun the scenarios with the new data" not a reason to run with the outdated model and present simply incorrect scenarios that will now never come to pass.
I've rejoined the Tory Party to partake in this leadership contest.
Truss for me, us republican Tories need to stick together.
How long until you qualify to vote in the leadership election - as that could easily be early January at the present rate.
Three months. But I may get an immediate vote, I was a member for 23 years, and my membership only lapsed last year.
FWIW - I think Boris Johnson is secure until May at least, the locals might be the tipping point if it is a very bad night.
Still has to be a danger that there is a meltdown over COVID restrictions. They kicked the can today, but seems like there isn't much collective responsibility in the cabinet.
Actually I think Boris Johnson has played a blinder today.
If Omicron is as bad as SAGE think it will be, he'll put the blame squarely on the cabinet refuseniks.
According to @JosiasJessop we will never know how bad SAGE think it will be, as they're under no obligation or responsibility to say how it will be.
Their only responsibility apparently is to forward the worst case catastrophising they can come up with, while anything that they think will happen should be binned if it isn't a catastrophe.
So maybe no action is entirely the right response to how bad they think it will be and we should have that attitude towards SAGE indefinitely going forwards because apparently they'll never tell us if things are going to be OK in their expectations.
I really, really think you ned to go back to the last thread, read, and think. That's not what I'm saying at all.
Please don't misrepresent me.
You were saying, were you not, that only the catastrophic scenarios should be presented?
The scenario of what they think will happen, using the most accurate data available, as explained well by @MaxPB above was not in the SAGE scenarios. Why not? Why should they not present what they think will happen, using the best available data?
There isn't one "never event" in question here, there's two. And under the principle that advisors advise and ministers decide, the best available modelling should be presented without prejudice to what outcome it gives so that the ministers can make an informed decision.
If you wish to provide worst case scenarios AS WELL as central scenarios that's entirely reasonable. And since we have two worst cases here (an NHS collapse, and an unnecessary lockdown) it would make sense to show what assumptions are required for scenarios for both of those.
Instead the results from one catastrophe has been cherry picked without the central scenarios being shown, or the other side.
You and Max keep using the word "scenarios" wrong (or at least in a sense I have never seen it used, in modelling terms). The "scenarios" are things like: Do Nothing, Impose Some Restrictions, Full Lockdown. The complaint you are making seems to be that the model input parameters are biased towards pessimism, because the modellers are excluding data sources which would result in more optimistic selections.
There is no value in modelling scenarios in which we remove restrictions, because no-one is suggesting that. There is a good argument (an excellent one, actually) that we should be modelling based on realistic, rather than pessimistic, parameters, because otherwise we are overstating the chances of Bad Outcomes, and hence giving them too much weight when making policy decisions. However, it is wrong to say that they aren't capturing non-catastrophic scenarios, because the models are stochastic by nature, and are designed to model all possible future outcomes. In addition, if the secondary uncertainty around the mean parameters is high enough, you are also capturing the possibility that the more favourable view of the parameters is actually true. It's just that, because the mean parameters are artificially pessimistic, the midpoint of the ranges produced, and - especially - the worst case outcomes, will look worse than if you picked something more "realistic".
No you've missed the point. Absolutely I 100% agree that the scenarios should be on things like "do nothing" etc.
The problem isn't that the optimistic scenarios have been excluded, the problem is that the baseline scenarios have been excluded. As per Fraser Nelson's article the evidence on the severity of Omicron, which is already available from South Africa and has been used in the JP Morgan modelling wasn't used by SAGE.
That scenario was excluded as Medley said because the results didn't fit. They were only presenting catastrophe scenarios, so baseline scenarios (not optimistic ones) were excluded if they didn't show a catastrophe.
There's more other false and dodgy assumptions as Max has gone into great detail with.
This isn't about saying we should be cherry-picking optimistic or pessimistic, we shouldn't be doing either. Reasonable scenarios using accurate data weren't included in the data dump because they didn't present a catastrophe - that's not science that's fiddling with data.
I've rejoined the Tory Party to partake in this leadership contest.
Truss for me, us republican Tories need to stick together.
How long until you qualify to vote in the leadership election - as that could easily be early January at the present rate.
Three months. But I may get an immediate vote, I was a member for 23 years, and my membership only lapsed last year.
FWIW - I think Boris Johnson is secure until May at least, the locals might be the tipping point if it is a very bad night.
Still has to be a danger that there is a meltdown over COVID restrictions. They kicked the can today, but seems like there isn't much collective responsibility in the cabinet.
Actually I think Boris Johnson has played a blinder today.
If Omicron is as bad as SAGE think it will be, he'll put the blame squarely on the cabinet refuseniks.
According to @JosiasJessop we will never know how bad SAGE think it will be, as they're under no obligation or responsibility to say how it will be.
Their only responsibility apparently is to forward the worst case catastrophising they can come up with, while anything that they think will happen should be binned if it isn't a catastrophe.
So maybe no action is entirely the right response to how bad they think it will be and we should have that attitude towards SAGE indefinitely going forwards because apparently they'll never tell us if things are going to be OK in their expectations.
I really, really think you ned to go back to the last thread, read, and think. That's not what I'm saying at all.
Please don't misrepresent me.
You were saying, were you not, that only the catastrophic scenarios should be presented?
The scenario of what they think will happen, using the most accurate data available, as explained well by @MaxPB above was not in the SAGE scenarios. Why not? Why should they not present what they think will happen, using the best available data?
There isn't one "never event" in question here, there's two. And under the principle that advisors advise and ministers decide, the best available modelling should be presented without prejudice to what outcome it gives so that the ministers can make an informed decision.
If you wish to provide worst case scenarios AS WELL as central scenarios that's entirely reasonable. And since we have two worst cases here (an NHS collapse, and an unnecessary lockdown) it would make sense to show what assumptions are required for scenarios for both of those.
Instead the results from one catastrophe has been cherry picked without the central scenarios being shown, or the other side.
You and Max keep using the word "scenarios" wrong (or at least in a sense I have never seen it used, in modelling terms). The "scenarios" are things like: Do Nothing, Impose Some Restrictions, Full Lockdown. The complaint you are making seems to be that the model input parameters are biased towards pessimism, because the modellers are excluding data sources which would result in more optimistic selections.
There is no value in modelling scenarios in which we remove restrictions, because no-one is suggesting that. There is a good argument (an excellent one, actually) that we should be modelling based on realistic, rather than pessimistic, parameters, because otherwise we are overstating the chances of Bad Outcomes, and hence giving them too much weight when making policy decisions. However, it is wrong to say that they aren't capturing non-catastrophic scenarios, because the models are stochastic by nature, and are designed to model all possible future outcomes. In addition, if the secondary uncertainty around the mean parameters is high enough, you are also capturing the possibility that the more favourable view of the parameters is actually true. It's just that, because the mean parameters are artificially pessimistic, the midpoint of the ranges produced, and - especially - the worst case outcomes, will look worse than if you picked something more "realistic".
The advisors should present all possible likely outcomes of their models including ones where no further action is needed or restrictions are completely lifted - so long as any of their models indicate these outcomes. They should advise which of the modelled outcomes are more likely or less likely but they should certainly not be excluding outcomes because they are not worst case. It is for elected politicians to make decisions about which models they wish to consider but they need to be shown the full selection and then make an informed decision on that basis.
Only showing the worst outcomes is the scientists forcing the politicians to follow a particular preferred route even if that outcome is unlikely and irrespective of what the costs are to the country. It is a ludicrous way to manage a pandemic and all it will do is destroy any last vestiges of trust in the modelling or the advice.
That's not how the models work (to my understanding). You feed in a set of inputs, which includes a particular policy decision, like Do Nothing, or Impose Lockdown. The model spits out a set of probability-weighted outcomes, and the outputs are summarised and presented to policymakers. All possible outcomes are included which are relevant to those inputs. You can't get the model to tell you whether you should take action or not, because that's an input to the model, not an outcome.
Thanks, Richard, really good piece. She's a lay alright. Set up to fail - in the crazies eyes - on NI. And in any case she's rather vacuous. Parties tend to veer away from what they've just had so I expect BJ's replacement, when it happens, to be a fairly solid thinker.
Would one describe a man as being "vacuous"?
Yep - you on Covid.
Hey don't shoot the messenger. When I see sexism I call it out.
Maybe, we certainly cannot increase at that rate for long. But in a week or so's time hospital admissions will be at least double what they are now, maybe more.
Will they?
The data from Guateng is that hospitalisations of people *for* CV19/Omicron are pretty low.
The other day, I mentioned an anti-vaxxer I know from school. This is his latest missive:
"Dr Robert Malone, inventor of MRNA and DNA vaccines, tell people NOT to vaccinate their children with the Covid-19 vaccine for at least 5 years, until the research and testing has been done:"
Followed by a link to a TikTok video.
Facebook have added a 'Covid vaccines work' banner underneath the post.
From my (limited) knowledge, there's a fair bit wrong in it.
He didn't invent mRNA vaccines at all. Katarin Kariko did.
An elderly acquaintance has not only refused to have vaccinations but refuses to have a lateral flow test - "I read somewhere that they contain nasty particles". Her carer developed Covid this morning, so she is taking precautions: "I breathe in essential oils and then sneeze out the nasty stuff".
There are a lot of delusions out there, urged on by nutters on the internet.
I've rejoined the Tory Party to partake in this leadership contest.
Truss for me, us republican Tories need to stick together.
How long until you qualify to vote in the leadership election - as that could easily be early January at the present rate.
Three months. But I may get an immediate vote, I was a member for 23 years, and my membership only lapsed last year.
FWIW - I think Boris Johnson is secure until May at least, the locals might be the tipping point if it is a very bad night.
Still has to be a danger that there is a meltdown over COVID restrictions. They kicked the can today, but seems like there isn't much collective responsibility in the cabinet.
Actually I think Boris Johnson has played a blinder today.
If Omicron is as bad as SAGE think it will be, he'll put the blame squarely on the cabinet refuseniks.
According to @JosiasJessop we will never know how bad SAGE think it will be, as they're under no obligation or responsibility to say how it will be.
Their only responsibility apparently is to forward the worst case catastrophising they can come up with, while anything that they think will happen should be binned if it isn't a catastrophe.
So maybe no action is entirely the right response to how bad they think it will be and we should have that attitude towards SAGE indefinitely going forwards because apparently they'll never tell us if things are going to be OK in their expectations.
I really, really think you ned to go back to the last thread, read, and think. That's not what I'm saying at all.
Please don't misrepresent me.
You were saying, were you not, that only the catastrophic scenarios should be presented?
The scenario of what they think will happen, using the most accurate data available, as explained well by @MaxPB above was not in the SAGE scenarios. Why not? Why should they not present what they think will happen, using the best available data?
There isn't one "never event" in question here, there's two. And under the principle that advisors advise and ministers decide, the best available modelling should be presented without prejudice to what outcome it gives so that the ministers can make an informed decision.
If you wish to provide worst case scenarios AS WELL as central scenarios that's entirely reasonable. And since we have two worst cases here (an NHS collapse, and an unnecessary lockdown) it would make sense to show what assumptions are required for scenarios for both of those.
Instead the results from one catastrophe has been cherry picked without the central scenarios being shown, or the other side.
You and Max keep using the word "scenarios" wrong (or at least in a sense I have never seen it used, in modelling terms). The "scenarios" are things like: Do Nothing, Impose Some Restrictions, Full Lockdown. The complaint you are making seems to be that the model input parameters are biased towards pessimism, because the modellers are excluding data sources which would result in more optimistic selections.
There is no value in modelling scenarios in which we remove restrictions, because no-one is suggesting that. There is a good argument (an excellent one, actually) that we should be modelling based on realistic, rather than pessimistic, parameters, because otherwise we are overstating the chances of Bad Outcomes, and hence giving them too much weight when making policy decisions. However, it is wrong to say that they aren't capturing non-catastrophic scenarios, because the models are stochastic by nature, and are designed to model all possible future outcomes. In addition, if the secondary uncertainty around the mean parameters is high enough, you are also capturing the possibility that the more favourable view of the parameters is actually true. It's just that, because the mean parameters are artificially pessimistic, the midpoint of the ranges produced, and - especially - the worst case outcomes, will look worse than if you picked something more "realistic".
No you've missed the point. Absolutely I 100% agree that the scenarios should be on things like "do nothing" etc.
The problem isn't that the optimistic scenarios have been excluded, the problem is that the baseline scenarios have been excluded. As per Fraser Nelson's article the evidence on the severity of Omicron, which is already available from South Africa and has been used in the JP Morgan modelling wasn't used by SAGE.
That scenario was excluded as Medley said because the results didn't fit. They were only presenting catastrophe scenarios, so baseline scenarios (not optimistic ones) were excluded if they didn't show a catastrophe.
There's more other false and dodgy assumptions as Max has gone into great detail with.
This isn't about saying we should be cherry-picking optimistic or pessimistic, we shouldn't be doing either. Reasonable scenarios using accurate data weren't included in the data dump because they didn't present a catastrophe - that's not science that's fiddling with data.
OK, we're more or less agreeing now, except you're still stuck on using "scenarios" to mean "input parameters". Omicron being less severe than Delta is not a scenario; it's something the modellers have to make a call on either way, and they have chosen to downweight the evidence that it might be more benign, and cover that by putting in a large secondary uncertainty around their mean estimate. That is not the same as "excluding baseline scenarios", whatever that means. The wide ranges given show that all future outcomes are still included, it's just that the probabilities are different than if they'd not downweighted the evidence for Omicron being benign.
I've rejoined the Tory Party to partake in this leadership contest.
Truss for me, us republican Tories need to stick together.
How long until you qualify to vote in the leadership election - as that could easily be early January at the present rate.
Three months. But I may get an immediate vote, I was a member for 23 years, and my membership only lapsed last year.
FWIW - I think Boris Johnson is secure until May at least, the locals might be the tipping point if it is a very bad night.
Still has to be a danger that there is a meltdown over COVID restrictions. They kicked the can today, but seems like there isn't much collective responsibility in the cabinet.
Actually I think Boris Johnson has played a blinder today.
If Omicron is as bad as SAGE think it will be, he'll put the blame squarely on the cabinet refuseniks.
According to @JosiasJessop we will never know how bad SAGE think it will be, as they're under no obligation or responsibility to say how it will be.
Their only responsibility apparently is to forward the worst case catastrophising they can come up with, while anything that they think will happen should be binned if it isn't a catastrophe.
So maybe no action is entirely the right response to how bad they think it will be and we should have that attitude towards SAGE indefinitely going forwards because apparently they'll never tell us if things are going to be OK in their expectations.
I really, really think you ned to go back to the last thread, read, and think. That's not what I'm saying at all.
Please don't misrepresent me.
You were saying, were you not, that only the catastrophic scenarios should be presented?
The scenario of what they think will happen, using the most accurate data available, as explained well by @MaxPB above was not in the SAGE scenarios. Why not? Why should they not present what they think will happen, using the best available data?
There isn't one "never event" in question here, there's two. And under the principle that advisors advise and ministers decide, the best available modelling should be presented without prejudice to what outcome it gives so that the ministers can make an informed decision.
If you wish to provide worst case scenarios AS WELL as central scenarios that's entirely reasonable. And since we have two worst cases here (an NHS collapse, and an unnecessary lockdown) it would make sense to show what assumptions are required for scenarios for both of those.
Instead the results from one catastrophe has been cherry picked without the central scenarios being shown, or the other side.
You and Max keep using the word "scenarios" wrong (or at least in a sense I have never seen it used, in modelling terms). The "scenarios" are things like: Do Nothing, Impose Some Restrictions, Full Lockdown. The complaint you are making seems to be that the model input parameters are biased towards pessimism, because the modellers are excluding data sources which would result in more optimistic selections.
There is no value in modelling scenarios in which we remove restrictions, because no-one is suggesting that. There is a good argument (an excellent one, actually) that we should be modelling based on realistic, rather than pessimistic, parameters, because otherwise we are overstating the chances of Bad Outcomes, and hence giving them too much weight when making policy decisions. However, it is wrong to say that they aren't capturing non-catastrophic scenarios, because the models are stochastic by nature, and are designed to model all possible future outcomes. In addition, if the secondary uncertainty around the mean parameters is high enough, you are also capturing the possibility that the more favourable view of the parameters is actually true. It's just that, because the mean parameters are artificially pessimistic, the midpoint of the ranges produced, and - especially - the worst case outcomes, will look worse than if you picked something more "realistic".
The advisors should present all possible likely outcomes of their models including ones where no further action is needed or restrictions are completely lifted - so long as any of their models indicate these outcomes. They should advise which of the modelled outcomes are more likely or less likely but they should certainly not be excluding outcomes because they are not worst case. It is for elected politicians to make decisions about which models they wish to consider but they need to be shown the full selection and then make an informed decision on that basis.
Only showing the worst outcomes is the scientists forcing the politicians to follow a particular preferred route even if that outcome is unlikely and irrespective of what the costs are to the country. It is a ludicrous way to manage a pandemic and all it will do is destroy any last vestiges of trust in the modelling or the advice.
That's not how the models work (to my understanding). You feed in a set of inputs, which includes a particular policy decision, like Do Nothing, or Impose Lockdown. The model spits out a set of probability-weighted outcomes, and the outputs are summarised and presented to policymakers. All possible outcomes are included which are relevant to those inputs. You can't get the model to tell you whether you should take action or not, because that's an input to the model, not an outcome.
Deciding what answer you want before you do the modelling is not science. If that is what they are doing then we should chuck the whole lot out.
I've rejoined the Tory Party to partake in this leadership contest.
Truss for me, us republican Tories need to stick together.
How long until you qualify to vote in the leadership election - as that could easily be early January at the present rate.
Three months. But I may get an immediate vote, I was a member for 23 years, and my membership only lapsed last year.
FWIW - I think Boris Johnson is secure until May at least, the locals might be the tipping point if it is a very bad night.
Still has to be a danger that there is a meltdown over COVID restrictions. They kicked the can today, but seems like there isn't much collective responsibility in the cabinet.
Actually I think Boris Johnson has played a blinder today.
If Omicron is as bad as SAGE think it will be, he'll put the blame squarely on the cabinet refuseniks.
According to @JosiasJessop we will never know how bad SAGE think it will be, as they're under no obligation or responsibility to say how it will be.
Their only responsibility apparently is to forward the worst case catastrophising they can come up with, while anything that they think will happen should be binned if it isn't a catastrophe.
So maybe no action is entirely the right response to how bad they think it will be and we should have that attitude towards SAGE indefinitely going forwards because apparently they'll never tell us if things are going to be OK in their expectations.
I really, really think you ned to go back to the last thread, read, and think. That's not what I'm saying at all.
Please don't misrepresent me.
You were saying, were you not, that only the catastrophic scenarios should be presented?
The scenario of what they think will happen, using the most accurate data available, as explained well by @MaxPB above was not in the SAGE scenarios. Why not? Why should they not present what they think will happen, using the best available data?
There isn't one "never event" in question here, there's two. And under the principle that advisors advise and ministers decide, the best available modelling should be presented without prejudice to what outcome it gives so that the ministers can make an informed decision.
If you wish to provide worst case scenarios AS WELL as central scenarios that's entirely reasonable. And since we have two worst cases here (an NHS collapse, and an unnecessary lockdown) it would make sense to show what assumptions are required for scenarios for both of those.
Instead the results from one catastrophe has been cherry picked without the central scenarios being shown, or the other side.
No, that's not what I said. And I really, really, really, think you need to go back and *think* about what I, TimT, Andy Cooke (I think) and others wrote in opposition to you.
As for 'never' events: that depends on perspective. Someone who doesn't care how many people die from Covid as long as they can lead a normal life might have a different view from someone who thinks it is wise to try to reduce deaths where possible - and the costs of saving those lives should be weighed up against those lives. But if you go in saying we shouldn't be trying to save any lives, then that's a bogus calculation.
It is what you were saying. If I'm wrong, then say how I'm wrong, just because you don't like getting called out.
Its for the ministers to decide what outcome they take, not the scientists. That's why they should be given the full picture in order to make an educated decision, not half a picture presenting only the extremely gloomiest scenarios.
In good science you present all the relevant information, not just that which suits your agenda. That includes showing the scenarios [and their assumptions] that mean no action is necessary.
I've rejoined the Tory Party to partake in this leadership contest.
Truss for me, us republican Tories need to stick together.
How long until you qualify to vote in the leadership election - as that could easily be early January at the present rate.
Three months. But I may get an immediate vote, I was a member for 23 years, and my membership only lapsed last year.
FWIW - I think Boris Johnson is secure until May at least, the locals might be the tipping point if it is a very bad night.
Still has to be a danger that there is a meltdown over COVID restrictions. They kicked the can today, but seems like there isn't much collective responsibility in the cabinet.
Actually I think Boris Johnson has played a blinder today.
If Omicron is as bad as SAGE think it will be, he'll put the blame squarely on the cabinet refuseniks.
According to @JosiasJessop we will never know how bad SAGE think it will be, as they're under no obligation or responsibility to say how it will be.
Their only responsibility apparently is to forward the worst case catastrophising they can come up with, while anything that they think will happen should be binned if it isn't a catastrophe.
So maybe no action is entirely the right response to how bad they think it will be and we should have that attitude towards SAGE indefinitely going forwards because apparently they'll never tell us if things are going to be OK in their expectations.
I really, really think you ned to go back to the last thread, read, and think. That's not what I'm saying at all.
Please don't misrepresent me.
You were saying, were you not, that only the catastrophic scenarios should be presented?
The scenario of what they think will happen, using the most accurate data available, as explained well by @MaxPB above was not in the SAGE scenarios. Why not? Why should they not present what they think will happen, using the best available data?
There isn't one "never event" in question here, there's two. And under the principle that advisors advise and ministers decide, the best available modelling should be presented without prejudice to what outcome it gives so that the ministers can make an informed decision.
If you wish to provide worst case scenarios AS WELL as central scenarios that's entirely reasonable. And since we have two worst cases here (an NHS collapse, and an unnecessary lockdown) it would make sense to show what assumptions are required for scenarios for both of those.
Instead the results from one catastrophe has been cherry picked without the central scenarios being shown, or the other side.
You and Max keep using the word "scenarios" wrong (or at least in a sense I have never seen it used, in modelling terms). The "scenarios" are things like: Do Nothing, Impose Some Restrictions, Full Lockdown. The complaint you are making seems to be that the model input parameters are biased towards pessimism, because the modellers are excluding data sources which would result in more optimistic selections.
There is no value in modelling scenarios in which we remove restrictions, because no-one is suggesting that. There is a good argument (an excellent one, actually) that we should be modelling based on realistic, rather than pessimistic, parameters, because otherwise we are overstating the chances of Bad Outcomes, and hence giving them too much weight when making policy decisions. However, it is wrong to say that they aren't capturing non-catastrophic scenarios, because the models are stochastic by nature, and are designed to model all possible future outcomes. In addition, if the secondary uncertainty around the mean parameters is high enough, you are also capturing the possibility that the more favourable view of the parameters is actually true. It's just that, because the mean parameters are artificially pessimistic, the midpoint of the ranges produced, and - especially - the worst case outcomes, will look worse than if you picked something more "realistic".
The advisors should present all possible likely outcomes of their models including ones where no further action is needed or restrictions are completely lifted - so long as any of their models indicate these outcomes. They should advise which of the modelled outcomes are more likely or less likely but they should certainly not be excluding outcomes because they are not worst case. It is for elected politicians to make decisions about which models they wish to consider but they need to be shown the full selection and then make an informed decision on that basis.
Only showing the worst outcomes is the scientists forcing the politicians to follow a particular preferred route even if that outcome is unlikely and irrespective of what the costs are to the country. It is a ludicrous way to manage a pandemic and all it will do is destroy any last vestiges of trust in the modelling or the advice.
Part of the problem is that "there are a range of possible outcomes" is a little too nuanced for either politicians or journalists.
I've rejoined the Tory Party to partake in this leadership contest.
Truss for me, us republican Tories need to stick together.
How long until you qualify to vote in the leadership election - as that could easily be early January at the present rate.
Three months. But I may get an immediate vote, I was a member for 23 years, and my membership only lapsed last year.
FWIW - I think Boris Johnson is secure until May at least, the locals might be the tipping point if it is a very bad night.
Still has to be a danger that there is a meltdown over COVID restrictions. They kicked the can today, but seems like there isn't much collective responsibility in the cabinet.
Actually I think Boris Johnson has played a blinder today.
If Omicron is as bad as SAGE think it will be, he'll put the blame squarely on the cabinet refuseniks.
According to @JosiasJessop we will never know how bad SAGE think it will be, as they're under no obligation or responsibility to say how it will be.
Their only responsibility apparently is to forward the worst case catastrophising they can come up with, while anything that they think will happen should be binned if it isn't a catastrophe.
So maybe no action is entirely the right response to how bad they think it will be and we should have that attitude towards SAGE indefinitely going forwards because apparently they'll never tell us if things are going to be OK in their expectations.
I really, really think you ned to go back to the last thread, read, and think. That's not what I'm saying at all.
Please don't misrepresent me.
You were saying, were you not, that only the catastrophic scenarios should be presented?
The scenario of what they think will happen, using the most accurate data available, as explained well by @MaxPB above was not in the SAGE scenarios. Why not? Why should they not present what they think will happen, using the best available data?
There isn't one "never event" in question here, there's two. And under the principle that advisors advise and ministers decide, the best available modelling should be presented without prejudice to what outcome it gives so that the ministers can make an informed decision.
If you wish to provide worst case scenarios AS WELL as central scenarios that's entirely reasonable. And since we have two worst cases here (an NHS collapse, and an unnecessary lockdown) it would make sense to show what assumptions are required for scenarios for both of those.
Instead the results from one catastrophe has been cherry picked without the central scenarios being shown, or the other side.
You and Max keep using the word "scenarios" wrong (or at least in a sense I have never seen it used, in modelling terms). The "scenarios" are things like: Do Nothing, Impose Some Restrictions, Full Lockdown. The complaint you are making seems to be that the model input parameters are biased towards pessimism, because the modellers are excluding data sources which would result in more optimistic selections.
There is no value in modelling scenarios in which we remove restrictions, because no-one is suggesting that. There is a good argument (an excellent one, actually) that we should be modelling based on realistic, rather than pessimistic, parameters, because otherwise we are overstating the chances of Bad Outcomes, and hence giving them too much weight when making policy decisions. However, it is wrong to say that they aren't capturing non-catastrophic scenarios, because the models are stochastic by nature, and are designed to model all possible future outcomes. In addition, if the secondary uncertainty around the mean parameters is high enough, you are also capturing the possibility that the more favourable view of the parameters is actually true. It's just that, because the mean parameters are artificially pessimistic, the midpoint of the ranges produced, and - especially - the worst case outcomes, will look worse than if you picked something more "realistic".
The advisors should present all possible likely outcomes of their models including ones where no further action is needed or restrictions are completely lifted - so long as any of their models indicate these outcomes. They should advise which of the modelled outcomes are more likely or less likely but they should certainly not be excluding outcomes because they are not worst case. It is for elected politicians to make decisions about which models they wish to consider but they need to be shown the full selection and then make an informed decision on that basis.
Only showing the worst outcomes is the scientists forcing the politicians to follow a particular preferred route even if that outcome is unlikely and irrespective of what the costs are to the country. It is a ludicrous way to manage a pandemic and all it will do is destroy any last vestiges of trust in the modelling or the advice.
Part of the problem is that "there are a range of possible outcomes" is a little too nuanced for either politicians or journalists.
In which case they are not fit to be doing their jobs. Mind you in both cases we pretty much already knew that.
I've rejoined the Tory Party to partake in this leadership contest.
Truss for me, us republican Tories need to stick together.
How long until you qualify to vote in the leadership election - as that could easily be early January at the present rate.
Three months. But I may get an immediate vote, I was a member for 23 years, and my membership only lapsed last year.
FWIW - I think Boris Johnson is secure until May at least, the locals might be the tipping point if it is a very bad night.
Still has to be a danger that there is a meltdown over COVID restrictions. They kicked the can today, but seems like there isn't much collective responsibility in the cabinet.
Actually I think Boris Johnson has played a blinder today.
If Omicron is as bad as SAGE think it will be, he'll put the blame squarely on the cabinet refuseniks.
According to @JosiasJessop we will never know how bad SAGE think it will be, as they're under no obligation or responsibility to say how it will be.
Their only responsibility apparently is to forward the worst case catastrophising they can come up with, while anything that they think will happen should be binned if it isn't a catastrophe.
So maybe no action is entirely the right response to how bad they think it will be and we should have that attitude towards SAGE indefinitely going forwards because apparently they'll never tell us if things are going to be OK in their expectations.
I really, really think you ned to go back to the last thread, read, and think. That's not what I'm saying at all.
Please don't misrepresent me.
You were saying, were you not, that only the catastrophic scenarios should be presented?
The scenario of what they think will happen, using the most accurate data available, as explained well by @MaxPB above was not in the SAGE scenarios. Why not? Why should they not present what they think will happen, using the best available data?
There isn't one "never event" in question here, there's two. And under the principle that advisors advise and ministers decide, the best available modelling should be presented without prejudice to what outcome it gives so that the ministers can make an informed decision.
If you wish to provide worst case scenarios AS WELL as central scenarios that's entirely reasonable. And since we have two worst cases here (an NHS collapse, and an unnecessary lockdown) it would make sense to show what assumptions are required for scenarios for both of those.
Instead the results from one catastrophe has been cherry picked without the central scenarios being shown, or the other side.
You and Max keep using the word "scenarios" wrong (or at least in a sense I have never seen it used, in modelling terms). The "scenarios" are things like: Do Nothing, Impose Some Restrictions, Full Lockdown. The complaint you are making seems to be that the model input parameters are biased towards pessimism, because the modellers are excluding data sources which would result in more optimistic selections.
There is no value in modelling scenarios in which we remove restrictions, because no-one is suggesting that. There is a good argument (an excellent one, actually) that we should be modelling based on realistic, rather than pessimistic, parameters, because otherwise we are overstating the chances of Bad Outcomes, and hence giving them too much weight when making policy decisions. However, it is wrong to say that they aren't capturing non-catastrophic scenarios, because the models are stochastic by nature, and are designed to model all possible future outcomes. In addition, if the secondary uncertainty around the mean parameters is high enough, you are also capturing the possibility that the more favourable view of the parameters is actually true. It's just that, because the mean parameters are artificially pessimistic, the midpoint of the ranges produced, and - especially - the worst case outcomes, will look worse than if you picked something more "realistic".
No you've missed the point. Absolutely I 100% agree that the scenarios should be on things like "do nothing" etc.
The problem isn't that the optimistic scenarios have been excluded, the problem is that the baseline scenarios have been excluded. As per Fraser Nelson's article the evidence on the severity of Omicron, which is already available from South Africa and has been used in the JP Morgan modelling wasn't used by SAGE.
That scenario was excluded as Medley said because the results didn't fit. They were only presenting catastrophe scenarios, so baseline scenarios (not optimistic ones) were excluded if they didn't show a catastrophe.
There's more other false and dodgy assumptions as Max has gone into great detail with.
This isn't about saying we should be cherry-picking optimistic or pessimistic, we shouldn't be doing either. Reasonable scenarios using accurate data weren't included in the data dump because they didn't present a catastrophe - that's not science that's fiddling with data.
OK, we're more or less agreeing now, except you're still stuck on using "scenarios" to mean "input parameters". Omicron being less severe than Delta is not a scenario; it's something the modellers have to make a call on either way, and they have chosen to downweight the evidence that it might be more benign, and cover that by putting in a large secondary uncertainty around their mean estimate. That is not the same as "excluding baseline scenarios", whatever that means. The wide ranges given show that all future outcomes are still included, it's just that the probabilities are different than if they'd not downweighted the evidence for Omicron being benign.
It is a scenario, a plausible one. The point of having a logistic regression model is that you can change multiple inputs and see what happens in each different scenario. The government modellers have been asked by some unknown party to not create scenarios where Omicron doesn't cause the NHS to collapse despite there being many plausible, likely I'd say at this point, ones where it won't.
One gets the impression Liz Truss is particularly disliked by Remainers because she used to be a Remainer herself and then switched to Leave as soon as the result was in.
I've rejoined the Tory Party to partake in this leadership contest.
Truss for me, us republican Tories need to stick together.
How long until you qualify to vote in the leadership election - as that could easily be early January at the present rate.
Three months. But I may get an immediate vote, I was a member for 23 years, and my membership only lapsed last year.
FWIW - I think Boris Johnson is secure until May at least, the locals might be the tipping point if it is a very bad night.
Still has to be a danger that there is a meltdown over COVID restrictions. They kicked the can today, but seems like there isn't much collective responsibility in the cabinet.
Actually I think Boris Johnson has played a blinder today.
If Omicron is as bad as SAGE think it will be, he'll put the blame squarely on the cabinet refuseniks.
According to @JosiasJessop we will never know how bad SAGE think it will be, as they're under no obligation or responsibility to say how it will be.
Their only responsibility apparently is to forward the worst case catastrophising they can come up with, while anything that they think will happen should be binned if it isn't a catastrophe.
So maybe no action is entirely the right response to how bad they think it will be and we should have that attitude towards SAGE indefinitely going forwards because apparently they'll never tell us if things are going to be OK in their expectations.
I really, really think you ned to go back to the last thread, read, and think. That's not what I'm saying at all.
Please don't misrepresent me.
You were saying, were you not, that only the catastrophic scenarios should be presented?
The scenario of what they think will happen, using the most accurate data available, as explained well by @MaxPB above was not in the SAGE scenarios. Why not? Why should they not present what they think will happen, using the best available data?
There isn't one "never event" in question here, there's two. And under the principle that advisors advise and ministers decide, the best available modelling should be presented without prejudice to what outcome it gives so that the ministers can make an informed decision.
If you wish to provide worst case scenarios AS WELL as central scenarios that's entirely reasonable. And since we have two worst cases here (an NHS collapse, and an unnecessary lockdown) it would make sense to show what assumptions are required for scenarios for both of those.
Instead the results from one catastrophe has been cherry picked without the central scenarios being shown, or the other side.
You and Max keep using the word "scenarios" wrong (or at least in a sense I have never seen it used, in modelling terms). The "scenarios" are things like: Do Nothing, Impose Some Restrictions, Full Lockdown. The complaint you are making seems to be that the model input parameters are biased towards pessimism, because the modellers are excluding data sources which would result in more optimistic selections.
There is no value in modelling scenarios in which we remove restrictions, because no-one is suggesting that. There is a good argument (an excellent one, actually) that we should be modelling based on realistic, rather than pessimistic, parameters, because otherwise we are overstating the chances of Bad Outcomes, and hence giving them too much weight when making policy decisions. However, it is wrong to say that they aren't capturing non-catastrophic scenarios, because the models are stochastic by nature, and are designed to model all possible future outcomes. In addition, if the secondary uncertainty around the mean parameters is high enough, you are also capturing the possibility that the more favourable view of the parameters is actually true. It's just that, because the mean parameters are artificially pessimistic, the midpoint of the ranges produced, and - especially - the worst case outcomes, will look worse than if you picked something more "realistic".
The advisors should present all possible likely outcomes of their models including ones where no further action is needed or restrictions are completely lifted - so long as any of their models indicate these outcomes. They should advise which of the modelled outcomes are more likely or less likely but they should certainly not be excluding outcomes because they are not worst case. It is for elected politicians to make decisions about which models they wish to consider but they need to be shown the full selection and then make an informed decision on that basis.
Only showing the worst outcomes is the scientists forcing the politicians to follow a particular preferred route even if that outcome is unlikely and irrespective of what the costs are to the country. It is a ludicrous way to manage a pandemic and all it will do is destroy any last vestiges of trust in the modelling or the advice.
Part of the problem is that "there are a range of possible outcomes" is a little too nuanced for either politicians or journalists.
Particularly when presumably the modellers (quite reasonably) shrug their shoulders and go "those are some possible outcomes but it's not our job to pick which one we're going to react to, mate".
Meanwhile, in Holland it is illegal to go to a pub or a restaurant, shop for clothes or even (and I find this hard to believe) hardware, which has never been the case in the UK.
I've rejoined the Tory Party to partake in this leadership contest.
Truss for me, us republican Tories need to stick together.
How long until you qualify to vote in the leadership election - as that could easily be early January at the present rate.
Three months. But I may get an immediate vote, I was a member for 23 years, and my membership only lapsed last year.
FWIW - I think Boris Johnson is secure until May at least, the locals might be the tipping point if it is a very bad night.
Still has to be a danger that there is a meltdown over COVID restrictions. They kicked the can today, but seems like there isn't much collective responsibility in the cabinet.
Actually I think Boris Johnson has played a blinder today.
If Omicron is as bad as SAGE think it will be, he'll put the blame squarely on the cabinet refuseniks.
According to @JosiasJessop we will never know how bad SAGE think it will be, as they're under no obligation or responsibility to say how it will be.
Their only responsibility apparently is to forward the worst case catastrophising they can come up with, while anything that they think will happen should be binned if it isn't a catastrophe.
So maybe no action is entirely the right response to how bad they think it will be and we should have that attitude towards SAGE indefinitely going forwards because apparently they'll never tell us if things are going to be OK in their expectations.
I really, really think you ned to go back to the last thread, read, and think. That's not what I'm saying at all.
Please don't misrepresent me.
You were saying, were you not, that only the catastrophic scenarios should be presented?
The scenario of what they think will happen, using the most accurate data available, as explained well by @MaxPB above was not in the SAGE scenarios. Why not? Why should they not present what they think will happen, using the best available data?
There isn't one "never event" in question here, there's two. And under the principle that advisors advise and ministers decide, the best available modelling should be presented without prejudice to what outcome it gives so that the ministers can make an informed decision.
If you wish to provide worst case scenarios AS WELL as central scenarios that's entirely reasonable. And since we have two worst cases here (an NHS collapse, and an unnecessary lockdown) it would make sense to show what assumptions are required for scenarios for both of those.
Instead the results from one catastrophe has been cherry picked without the central scenarios being shown, or the other side.
You and Max keep using the word "scenarios" wrong (or at least in a sense I have never seen it used, in modelling terms). The "scenarios" are things like: Do Nothing, Impose Some Restrictions, Full Lockdown. The complaint you are making seems to be that the model input parameters are biased towards pessimism, because the modellers are excluding data sources which would result in more optimistic selections.
There is no value in modelling scenarios in which we remove restrictions, because no-one is suggesting that. There is a good argument (an excellent one, actually) that we should be modelling based on realistic, rather than pessimistic, parameters, because otherwise we are overstating the chances of Bad Outcomes, and hence giving them too much weight when making policy decisions. However, it is wrong to say that they aren't capturing non-catastrophic scenarios, because the models are stochastic by nature, and are designed to model all possible future outcomes. In addition, if the secondary uncertainty around the mean parameters is high enough, you are also capturing the possibility that the more favourable view of the parameters is actually true. It's just that, because the mean parameters are artificially pessimistic, the midpoint of the ranges produced, and - especially - the worst case outcomes, will look worse than if you picked something more "realistic".
No you've missed the point. Absolutely I 100% agree that the scenarios should be on things like "do nothing" etc.
The problem isn't that the optimistic scenarios have been excluded, the problem is that the baseline scenarios have been excluded. As per Fraser Nelson's article the evidence on the severity of Omicron, which is already available from South Africa and has been used in the JP Morgan modelling wasn't used by SAGE.
That scenario was excluded as Medley said because the results didn't fit. They were only presenting catastrophe scenarios, so baseline scenarios (not optimistic ones) were excluded if they didn't show a catastrophe.
There's more other false and dodgy assumptions as Max has gone into great detail with.
This isn't about saying we should be cherry-picking optimistic or pessimistic, we shouldn't be doing either. Reasonable scenarios using accurate data weren't included in the data dump because they didn't present a catastrophe - that's not science that's fiddling with data.
OK, we're more or less agreeing now, except you're still stuck on using "scenarios" to mean "input parameters". Omicron being less severe than Delta is not a scenario; it's something the modellers have to make a call on either way, and they have chosen to downweight the evidence that it might be more benign, and cover that by putting in a large secondary uncertainty around their mean estimate. That is not the same as "excluding baseline scenarios", whatever that means. The wide ranges given show that all future outcomes are still included, it's just that the probabilities are different than if they'd not downweighted the evidence for Omicron being benign.
The problem with that being that they are acting as if there is no downside to excluding those more optimistic inputs. But by doing so they are making the social, economic, mental and health consequences all the worse because they are prioritising combating a possibly non existent threat over the other health aspects.
I've rejoined the Tory Party to partake in this leadership contest.
Truss for me, us republican Tories need to stick together.
How long until you qualify to vote in the leadership election - as that could easily be early January at the present rate.
Three months. But I may get an immediate vote, I was a member for 23 years, and my membership only lapsed last year.
FWIW - I think Boris Johnson is secure until May at least, the locals might be the tipping point if it is a very bad night.
Still has to be a danger that there is a meltdown over COVID restrictions. They kicked the can today, but seems like there isn't much collective responsibility in the cabinet.
Actually I think Boris Johnson has played a blinder today.
If Omicron is as bad as SAGE think it will be, he'll put the blame squarely on the cabinet refuseniks.
According to @JosiasJessop we will never know how bad SAGE think it will be, as they're under no obligation or responsibility to say how it will be.
Their only responsibility apparently is to forward the worst case catastrophising they can come up with, while anything that they think will happen should be binned if it isn't a catastrophe.
So maybe no action is entirely the right response to how bad they think it will be and we should have that attitude towards SAGE indefinitely going forwards because apparently they'll never tell us if things are going to be OK in their expectations.
I really, really think you ned to go back to the last thread, read, and think. That's not what I'm saying at all.
Please don't misrepresent me.
You were saying, were you not, that only the catastrophic scenarios should be presented?
The scenario of what they think will happen, using the most accurate data available, as explained well by @MaxPB above was not in the SAGE scenarios. Why not? Why should they not present what they think will happen, using the best available data?
There isn't one "never event" in question here, there's two. And under the principle that advisors advise and ministers decide, the best available modelling should be presented without prejudice to what outcome it gives so that the ministers can make an informed decision.
If you wish to provide worst case scenarios AS WELL as central scenarios that's entirely reasonable. And since we have two worst cases here (an NHS collapse, and an unnecessary lockdown) it would make sense to show what assumptions are required for scenarios for both of those.
Instead the results from one catastrophe has been cherry picked without the central scenarios being shown, or the other side.
You and Max keep using the word "scenarios" wrong (or at least in a sense I have never seen it used, in modelling terms). The "scenarios" are things like: Do Nothing, Impose Some Restrictions, Full Lockdown. The complaint you are making seems to be that the model input parameters are biased towards pessimism, because the modellers are excluding data sources which would result in more optimistic selections.
There is no value in modelling scenarios in which we remove restrictions, because no-one is suggesting that. There is a good argument (an excellent one, actually) that we should be modelling based on realistic, rather than pessimistic, parameters, because otherwise we are overstating the chances of Bad Outcomes, and hence giving them too much weight when making policy decisions. However, it is wrong to say that they aren't capturing non-catastrophic scenarios, because the models are stochastic by nature, and are designed to model all possible future outcomes. In addition, if the secondary uncertainty around the mean parameters is high enough, you are also capturing the possibility that the more favourable view of the parameters is actually true. It's just that, because the mean parameters are artificially pessimistic, the midpoint of the ranges produced, and - especially - the worst case outcomes, will look worse than if you picked something more "realistic".
The advisors should present all possible likely outcomes of their models including ones where no further action is needed or restrictions are completely lifted - so long as any of their models indicate these outcomes. They should advise which of the modelled outcomes are more likely or less likely but they should certainly not be excluding outcomes because they are not worst case. It is for elected politicians to make decisions about which models they wish to consider but they need to be shown the full selection and then make an informed decision on that basis.
Only showing the worst outcomes is the scientists forcing the politicians to follow a particular preferred route even if that outcome is unlikely and irrespective of what the costs are to the country. It is a ludicrous way to manage a pandemic and all it will do is destroy any last vestiges of trust in the modelling or the advice.
That's not how the models work (to my understanding). You feed in a set of inputs, which includes a particular policy decision, like Do Nothing, or Impose Lockdown. The model spits out a set of probability-weighted outcomes, and the outputs are summarised and presented to policymakers. All possible outcomes are included which are relevant to those inputs. You can't get the model to tell you whether you should take action or not, because that's an input to the model, not an outcome.
Deciding what answer you want before you do the modelling is not science. If that is what they are doing then we should chuck the whole lot out.
The purpose of the modelling is to answer questions like "If we do X, what will happen?" If you model that for several different values of X, you can then decide which one to do.
Fraser Nelson: "So we have an asteroid that may hit the Earth?" Scientist: "Yes. And we have no idea how likely it is - only that it's heading towards us. NORAD were too busy tracking Santa Claus." Fraser Nelson: "But it may miss." Scientist: "Yes." Fraser Nelson: "So why are you only modelling what will happen if it hits?" Scientist: "Because the decision-makers need to consider what to do if the worst comes to the worst." Fraser Nelson: "But they might not have to do anything if it doesn't hit." Scientist: "But it may. And they need to think about what they'd do." Fraser Nelson: "Why didn't you model the fact it might miss?" Scientist: "Because that doesn't really help the decision-makers."
That Fraser Nelson article in the Spectator is really a whole load of nothing IMO. What the scientist said makes sense.
What a load of crap.
If the government are considering locking us down because of the virus then they need to know what's likely to happen with the virus. If the models say that the NHS isn't likely to be overwhelmed but those models are disregarded in favour of those that say it is, then that's operating with false information.
It's not (see TimT's reply).
It is.
If the government is weighing up their response then they need the full information. ' If the full information says for instance there's a 99.9% chance that the NHS won't be overwhelmed, but there's a 0.1% chance that it is - then do you seriously think the government should only be shown the 0.1% scenario without any qualification of caveat or rating of how likely it is?
They should get the full information, and be allowed to judge with full knowledge whether the risk of these so-called "never events" are worth acting over or not. If they don't have the full information, then they can't weigh that up.
"If the full information says for instance there's a 99.9% chance that the NHS won't be overwhelmed, but there's a 0.1% chance that it is"
But that's not what Nelson was talking about (although he moved onto that at the end). He was talking about a lack of a model that replicated some of JP Morgan's modelling, not the probabilities of any scenario.
Nelson was talking about modelling that used data that the scientists had recognised. But since this model didn't give "the right" answer it was disregarded.
If you decide in advance to disregard all models that don't give a certain outcome, then you've prejudiced your work in advance.
I might suggest you reread what I (and especially TimT) have written.
There is no point in presenting reasonable scenarios where the decision-makers need to do nothing, because that's pointless. The decision-makers need to know the scenarios where they may have to do things, so they can consider them.
The probabilities may come later.
I assumed when Max and others were going on about the Nelson article, that the scientist had said something outrageous. Instead he said something utterly sensible, and Nelson has either misunderstood the point of the modelling, or is deliberately shit-stirring. The scientist should have explained a little better, though.
No this is utterly wrong. If you have a whole range of possible scenarios and you only present the ones that result in action then you are taking the responsibility for decision making away from the politicians and giving it to the unelected scientists. By removing the scenarios where nothing needs to be done you are forcing the politicians into a position where they either do something - even if it is probably unnecessary - or they can be accused of ignoring the evidence as it was presented to them. It is absolutely vital that the advisors present all possible scenarios and weigh them for the decision makers.
It is not utterly wrong. It is presenting the scenarios that need stuff doing about them.
"Present all possible scenarios"
And how many different scenarios is that? Ten? a hundred? A thousand?
The scientists are saying: "these are the sh*t scenarios. These are the ones that, if they occur, we will need to take actions on." Presenting scenarios that need no action is pointless - because there are no actions required. It doesn't mean the politicians are being told these will happen, or that these are the only scenarios. They are the ones to worry about.
It's not that they're saying these are the only scenarios that may happen; or the probabilities of them happening; they're the ones the politicians really need to worry about.
It's simple stuff.
FPT
The problem is that right now people think the most optomistic scenario from SAGE is for a peak of 3000 admissions a day, as that is their lowest model. Except other models suggest this is NOT the lowest possible. The narrative is running that if we do nothing, we will definitely see a minimum of 3000 admissions a day. It’s dishonest. I’m not blaming the scientists entirely here, I want to know what instructions they are getting.
Then I'd blame those telling the people that this is the issue without a fair grasp on what SAGE and SPI-M-O are doing, or their terms of reference, or the specifics of decision making during a period of ignorance but with the possibility of a Never Event (as described by TimT).
That would be the attention-seekers of the media and some Twitter commentators. Who grab stuff, skim it for an attention-grabbing takeaway line, and plaster it everywhere they can.
Why use knowingly incorrect data points for model inputs, though? However you cut it that's just bad practice and now they've managed to lose their last ounce of credibility with the politicians. The fault lies with whoever is asking for them to do that, sure, but also with them for not having the professional integrity to flag that up as bad practice. They're being asked to put their finger on the scale so someone can use their findings to push the lockdown decision making process forwards. As it turned out Cabinet took exception to being presented data that has no basis in reality.
Which knowingly incorrect parameters? Looking at the Warwick paper, they had several possible Ve figures against potential levels of immune escape and booster effectiveness (high to low), and it was written and updated on the 11th of December.
TimT explained the parameters of the decision-making process in conditions of high uncertainty and with credible possibilities of a Never Event, and it follows exactly that process.
So it could either be that process, with the criteria of a Never Event as laid down over a decade ago in the last Labour Government, or its a cabal of evil scientists or a secret civil servant with a fetish for lockdowns who’s not been sprung yet.
I hate lockdowns. I personally incline to the view that there will not be a need for them. But I agree that the probabilities involved are still unclear, and that they should ensure that they have the options on the table, together with what they would or would not do, as well as the costs for them, all clearly laid out.
And, of course, as Paul Mainwood drily observed a few days ago: if we do want to adjust for a variable severity, we can simply do the less-than-arcane procedure of “multiply by a constant” against the outcomes.
For one of the three vaccine efficacy tables, they used the one not based on observation and used the 85% which used an additional 6x reduction in antibody binding efficiency. The second and slightly more contentious one is not using our own study data which gives Omicron a different manifestation of symptoms than Delta, get the central forecast simply assumes that Omicron manifests identical symptoms to Delta.
The issue I have with the kind of modelling here is that Omicron isn't an unknown unknown. We know things about it, we know fairly well what vaccine efficacy is, we know with a reasonable degree of certainty that it manifests as an upper respiratory tract virus than a lower respiratory tract one like Delta.
It's simply not right to ask the decision makers to form an opinion on the right way forwards if the best available data isn't being used and instead guesswork that we know to be worse is being used. I understand the need for expediency and that in a situation where exponential growth is a factor acting late can mean it's too late to act. I do understand that they are under pressure. However, there's simply no need to put their finger on the scales, whoever asked them to do that is trying to force a decision on lockdown and it's spectacularly back fired on them because now the politicians have lost all confidence in the modelling. The 7k central forecast from the summer had already hurt their credibility now it's completely destroyed because they have been found to be fiddling the numbers for some as yet unknown DoH bod who wants to guide the process in a specific direction.
On the whole idea of the "never scenario" it's fanciful that anything that comes with COVID is now in that category.
I also agree that we probably won't need a lockdown. My issue is that the data for that scenario isn't being presented to the politicians when it absolutely should be. It's not an unlikely scenario that Omicron is a paper tiger, yet our politicians are given that option because the unnamed party in the DoH has deemed it irrelevant to the decision making process and asked the modellers not to give their findings for that scenario. It would be a piece of piss to do it as well.
There was never a “7k central forecast” in the summer as anyone who ever read the fucking document in question knew.
True, but why have it in there then? Just like these unlikely to pass scenarios all it does is sap their credibility. You've fairly outlined why the modelling isn't bad, yet the politicians are making decisions based on 600-6000 deaths per day.
It was in there, as it made clear, because they’d just had a surge of R=1.5 in Scotland and therefore couldn’t rule it out as a scenario for the UK as a whole. So they put it as the highest plausible scenario.
Considering it plausible because it had literally just been happening in Scotland, although they were pointing more at the R=1.1 route.
As they said at the time and in the explanation of what each of the scenarios were.
I get why the media ignored the explanation (attention-grabbing) but if politicians who are being presented with the scenarios and having the experts answer questions on hand and still be unable to understand, then such idiots shouldn’t be anywhere near any position of power or decision-making.
But a sustained 1.1 also didn't happen and a sustained 1.5 was never plausible, it was, in fact, implausible yet presented as something to plan against.
We're lucky that the politicians ignored the summer forecasts because they'd have insisted on silly measures like indoor distancing and vaccine passports which have been of no help in Europe and in fact may have been counter productive as it hasn't allowed them to build a natural immunity wall in the unvaccinated.
As to your other points, your faith and trust in people who you have no possible way of removing from the levers of power is worrying. Politicians may be flawed (and this set particularly so) yet we still have some way of making the decision on whether or not we want to do what they say. The unnamed bod in the DoH putting his fingers on the scale and presenting the politicians with a range unlikely scenarios and not also presenting the likely scenarios, well we have no way of removing this person from power, no way of holding them accountable to us.
That makes me instinctively wary of of them and sceptical about what their motives are. I know what Labour's motives are, I know what the Tories motives are, they told me at length in ab election campaign. We have an unelected tyranny of experts, maybe you're happy to live like that, I'm not. Quite frankly, our views on this differ too much, I'm extremely worried that the UK and its people are too happy to simply hand our freedoms to these unknown experts with no guarantees they'll be happy to give them back again.
So yes, colour me very sceptical about anyone who is pushing a very narrow set of possible outcomes while excluding other more favourable ones. There's simply no need for it.
As to the other point, I understand the data came later than the report, that's a reason to tell the decision makers - "we have new key data inputs for our model, please await our new results while we rerun the scenarios with the new data" not a reason to run with the outdated model and present simply incorrect scenarios that will now never come to pass.
An excellent post Max. Some people have not seen how some humans can be motivated by darkness, with the combination of power and ego being intoxicating. If you are a moral person trying to always do their best it can be a shock to learn not everyone is like that, and that corruption and casual threats of violence can be found in quite surprising places.
What concerns me, is we don’t even know the name of the person or persons “putting their fingers on the scale”, much less take a view on their motivation. They may just be a bit misguided but trying to what’s right, or poor communicators. But they may not.
I've rejoined the Tory Party to partake in this leadership contest.
Truss for me, us republican Tories need to stick together.
How long until you qualify to vote in the leadership election - as that could easily be early January at the present rate.
Three months. But I may get an immediate vote, I was a member for 23 years, and my membership only lapsed last year.
FWIW - I think Boris Johnson is secure until May at least, the locals might be the tipping point if it is a very bad night.
Still has to be a danger that there is a meltdown over COVID restrictions. They kicked the can today, but seems like there isn't much collective responsibility in the cabinet.
Actually I think Boris Johnson has played a blinder today.
If Omicron is as bad as SAGE think it will be, he'll put the blame squarely on the cabinet refuseniks.
According to @JosiasJessop we will never know how bad SAGE think it will be, as they're under no obligation or responsibility to say how it will be.
Their only responsibility apparently is to forward the worst case catastrophising they can come up with, while anything that they think will happen should be binned if it isn't a catastrophe.
So maybe no action is entirely the right response to how bad they think it will be and we should have that attitude towards SAGE indefinitely going forwards because apparently they'll never tell us if things are going to be OK in their expectations.
I really, really think you ned to go back to the last thread, read, and think. That's not what I'm saying at all.
Please don't misrepresent me.
You were saying, were you not, that only the catastrophic scenarios should be presented?
The scenario of what they think will happen, using the most accurate data available, as explained well by @MaxPB above was not in the SAGE scenarios. Why not? Why should they not present what they think will happen, using the best available data?
There isn't one "never event" in question here, there's two. And under the principle that advisors advise and ministers decide, the best available modelling should be presented without prejudice to what outcome it gives so that the ministers can make an informed decision.
If you wish to provide worst case scenarios AS WELL as central scenarios that's entirely reasonable. And since we have two worst cases here (an NHS collapse, and an unnecessary lockdown) it would make sense to show what assumptions are required for scenarios for both of those.
Instead the results from one catastrophe has been cherry picked without the central scenarios being shown, or the other side.
You and Max keep using the word "scenarios" wrong (or at least in a sense I have never seen it used, in modelling terms). The "scenarios" are things like: Do Nothing, Impose Some Restrictions, Full Lockdown. The complaint you are making seems to be that the model input parameters are biased towards pessimism, because the modellers are excluding data sources which would result in more optimistic selections.
There is no value in modelling scenarios in which we remove restrictions, because no-one is suggesting that. There is a good argument (an excellent one, actually) that we should be modelling based on realistic, rather than pessimistic, parameters, because otherwise we are overstating the chances of Bad Outcomes, and hence giving them too much weight when making policy decisions. However, it is wrong to say that they aren't capturing non-catastrophic scenarios, because the models are stochastic by nature, and are designed to model all possible future outcomes. In addition, if the secondary uncertainty around the mean parameters is high enough, you are also capturing the possibility that the more favourable view of the parameters is actually true. It's just that, because the mean parameters are artificially pessimistic, the midpoint of the ranges produced, and - especially - the worst case outcomes, will look worse than if you picked something more "realistic".
No you've missed the point. Absolutely I 100% agree that the scenarios should be on things like "do nothing" etc.
The problem isn't that the optimistic scenarios have been excluded, the problem is that the baseline scenarios have been excluded. As per Fraser Nelson's article the evidence on the severity of Omicron, which is already available from South Africa and has been used in the JP Morgan modelling wasn't used by SAGE.
That scenario was excluded as Medley said because the results didn't fit. They were only presenting catastrophe scenarios, so baseline scenarios (not optimistic ones) were excluded if they didn't show a catastrophe.
There's more other false and dodgy assumptions as Max has gone into great detail with.
This isn't about saying we should be cherry-picking optimistic or pessimistic, we shouldn't be doing either. Reasonable scenarios using accurate data weren't included in the data dump because they didn't present a catastrophe - that's not science that's fiddling with data.
OK, we're more or less agreeing now, except you're still stuck on using "scenarios" to mean "input parameters". Omicron being less severe than Delta is not a scenario; it's something the modellers have to make a call on either way, and they have chosen to downweight the evidence that it might be more benign, and cover that by putting in a large secondary uncertainty around their mean estimate. That is not the same as "excluding baseline scenarios", whatever that means. The wide ranges given show that all future outcomes are still included, it's just that the probabilities are different than if they'd not downweighted the evidence for Omicron being benign.
The various input parameters were presented as scenarios and talked about as such so yes that's what we're talking about.
The evidence shows that it might be more benign so that input parameter should be presented as one of the scenarios shouldn't it?
Similarly with VE they've only presented more negative figures than what the evidence presents.
When questioned on why not to present the potentially more accurate and more scientific scenario of input parameters using the best known data, its being excluded based upon the results.
It is not scientific to keep changing your input parameters until you get an output that you are happy with. You might as well just say "I want a lockdown" and not bother with any models if you want to do that. Create your parameters, run your models and show the outcomes regardless of their outputs. If they're negative, they're negative. If they're positive, they're positive. If they're neutral, they're neutral. At least then you're not forcing your own agenda onto the science, let the science speak for itself.
Meanwhile, in Holland, it is illegal to go to a pub or a restaurant, shop for clothes or even (and I find this hard to believe) hardware, which has never been the case in the UK.
Actually Brexit is not an issue for many especially if the NI Protocol can be resolved
One gets the impression Liz Truss is particularly disliked by Remainers because she used to be a Remainer herself and then switched to Leave as soon as the result was in.
I've rejoined the Tory Party to partake in this leadership contest.
Truss for me, us republican Tories need to stick together.
How long until you qualify to vote in the leadership election - as that could easily be early January at the present rate.
Three months. But I may get an immediate vote, I was a member for 23 years, and my membership only lapsed last year.
FWIW - I think Boris Johnson is secure until May at least, the locals might be the tipping point if it is a very bad night.
Still has to be a danger that there is a meltdown over COVID restrictions. They kicked the can today, but seems like there isn't much collective responsibility in the cabinet.
Actually I think Boris Johnson has played a blinder today.
If Omicron is as bad as SAGE think it will be, he'll put the blame squarely on the cabinet refuseniks.
According to @JosiasJessop we will never know how bad SAGE think it will be, as they're under no obligation or responsibility to say how it will be.
Their only responsibility apparently is to forward the worst case catastrophising they can come up with, while anything that they think will happen should be binned if it isn't a catastrophe.
So maybe no action is entirely the right response to how bad they think it will be and we should have that attitude towards SAGE indefinitely going forwards because apparently they'll never tell us if things are going to be OK in their expectations.
I really, really think you ned to go back to the last thread, read, and think. That's not what I'm saying at all.
Please don't misrepresent me.
You were saying, were you not, that only the catastrophic scenarios should be presented?
The scenario of what they think will happen, using the most accurate data available, as explained well by @MaxPB above was not in the SAGE scenarios. Why not? Why should they not present what they think will happen, using the best available data?
There isn't one "never event" in question here, there's two. And under the principle that advisors advise and ministers decide, the best available modelling should be presented without prejudice to what outcome it gives so that the ministers can make an informed decision.
If you wish to provide worst case scenarios AS WELL as central scenarios that's entirely reasonable. And since we have two worst cases here (an NHS collapse, and an unnecessary lockdown) it would make sense to show what assumptions are required for scenarios for both of those.
Instead the results from one catastrophe has been cherry picked without the central scenarios being shown, or the other side.
You and Max keep using the word "scenarios" wrong (or at least in a sense I have never seen it used, in modelling terms). The "scenarios" are things like: Do Nothing, Impose Some Restrictions, Full Lockdown. The complaint you are making seems to be that the model input parameters are biased towards pessimism, because the modellers are excluding data sources which would result in more optimistic selections.
There is no value in modelling scenarios in which we remove restrictions, because no-one is suggesting that. There is a good argument (an excellent one, actually) that we should be modelling based on realistic, rather than pessimistic, parameters, because otherwise we are overstating the chances of Bad Outcomes, and hence giving them too much weight when making policy decisions. However, it is wrong to say that they aren't capturing non-catastrophic scenarios, because the models are stochastic by nature, and are designed to model all possible future outcomes. In addition, if the secondary uncertainty around the mean parameters is high enough, you are also capturing the possibility that the more favourable view of the parameters is actually true. It's just that, because the mean parameters are artificially pessimistic, the midpoint of the ranges produced, and - especially - the worst case outcomes, will look worse than if you picked something more "realistic".
The advisors should present all possible likely outcomes of their models including ones where no further action is needed or restrictions are completely lifted - so long as any of their models indicate these outcomes. They should advise which of the modelled outcomes are more likely or less likely but they should certainly not be excluding outcomes because they are not worst case. It is for elected politicians to make decisions about which models they wish to consider but they need to be shown the full selection and then make an informed decision on that basis.
Only showing the worst outcomes is the scientists forcing the politicians to follow a particular preferred route even if that outcome is unlikely and irrespective of what the costs are to the country. It is a ludicrous way to manage a pandemic and all it will do is destroy any last vestiges of trust in the modelling or the advice.
That's not how the models work (to my understanding). You feed in a set of inputs, which includes a particular policy decision, like Do Nothing, or Impose Lockdown. The model spits out a set of probability-weighted outcomes, and the outputs are summarised and presented to policymakers. All possible outcomes are included which are relevant to those inputs. You can't get the model to tell you whether you should take action or not, because that's an input to the model, not an outcome.
Deciding what answer you want before you do the modelling is not science. If that is what they are doing then we should chuck the whole lot out.
The purpose of the modelling is to answer questions like "If we do X, what will happen?" If you model that for several different values of X, you can then decide which one to do.
How else would you decide?
No, the point of modelling is to say if we do X while variables A, B, and C are true, false or in between what happens. That's the model scenario. No one is running a simple single input linear regression model, it will be at minimum a multiple input logistic regression with reasonable weighting for vaccine efficacy, transmission and virulence/severity.
I've rejoined the Tory Party to partake in this leadership contest.
Truss for me, us republican Tories need to stick together.
How long until you qualify to vote in the leadership election - as that could easily be early January at the present rate.
Three months. But I may get an immediate vote, I was a member for 23 years, and my membership only lapsed last year.
FWIW - I think Boris Johnson is secure until May at least, the locals might be the tipping point if it is a very bad night.
Still has to be a danger that there is a meltdown over COVID restrictions. They kicked the can today, but seems like there isn't much collective responsibility in the cabinet.
Actually I think Boris Johnson has played a blinder today.
If Omicron is as bad as SAGE think it will be, he'll put the blame squarely on the cabinet refuseniks.
According to @JosiasJessop we will never know how bad SAGE think it will be, as they're under no obligation or responsibility to say how it will be.
Their only responsibility apparently is to forward the worst case catastrophising they can come up with, while anything that they think will happen should be binned if it isn't a catastrophe.
So maybe no action is entirely the right response to how bad they think it will be and we should have that attitude towards SAGE indefinitely going forwards because apparently they'll never tell us if things are going to be OK in their expectations.
I really, really think you ned to go back to the last thread, read, and think. That's not what I'm saying at all.
Please don't misrepresent me.
You were saying, were you not, that only the catastrophic scenarios should be presented?
The scenario of what they think will happen, using the most accurate data available, as explained well by @MaxPB above was not in the SAGE scenarios. Why not? Why should they not present what they think will happen, using the best available data?
There isn't one "never event" in question here, there's two. And under the principle that advisors advise and ministers decide, the best available modelling should be presented without prejudice to what outcome it gives so that the ministers can make an informed decision.
If you wish to provide worst case scenarios AS WELL as central scenarios that's entirely reasonable. And since we have two worst cases here (an NHS collapse, and an unnecessary lockdown) it would make sense to show what assumptions are required for scenarios for both of those.
Instead the results from one catastrophe has been cherry picked without the central scenarios being shown, or the other side.
You and Max keep using the word "scenarios" wrong (or at least in a sense I have never seen it used, in modelling terms). The "scenarios" are things like: Do Nothing, Impose Some Restrictions, Full Lockdown. The complaint you are making seems to be that the model input parameters are biased towards pessimism, because the modellers are excluding data sources which would result in more optimistic selections.
There is no value in modelling scenarios in which we remove restrictions, because no-one is suggesting that. There is a good argument (an excellent one, actually) that we should be modelling based on realistic, rather than pessimistic, parameters, because otherwise we are overstating the chances of Bad Outcomes, and hence giving them too much weight when making policy decisions. However, it is wrong to say that they aren't capturing non-catastrophic scenarios, because the models are stochastic by nature, and are designed to model all possible future outcomes. In addition, if the secondary uncertainty around the mean parameters is high enough, you are also capturing the possibility that the more favourable view of the parameters is actually true. It's just that, because the mean parameters are artificially pessimistic, the midpoint of the ranges produced, and - especially - the worst case outcomes, will look worse than if you picked something more "realistic".
The advisors should present all possible likely outcomes of their models including ones where no further action is needed or restrictions are completely lifted - so long as any of their models indicate these outcomes. They should advise which of the modelled outcomes are more likely or less likely but they should certainly not be excluding outcomes because they are not worst case. It is for elected politicians to make decisions about which models they wish to consider but they need to be shown the full selection and then make an informed decision on that basis.
Only showing the worst outcomes is the scientists forcing the politicians to follow a particular preferred route even if that outcome is unlikely and irrespective of what the costs are to the country. It is a ludicrous way to manage a pandemic and all it will do is destroy any last vestiges of trust in the modelling or the advice.
That's not how the models work (to my understanding). You feed in a set of inputs, which includes a particular policy decision, like Do Nothing, or Impose Lockdown. The model spits out a set of probability-weighted outcomes, and the outputs are summarised and presented to policymakers. All possible outcomes are included which are relevant to those inputs. You can't get the model to tell you whether you should take action or not, because that's an input to the model, not an outcome.
Deciding what answer you want before you do the modelling is not science. If that is what they are doing then we should chuck the whole lot out.
The purpose of the modelling is to answer questions like "If we do X, what will happen?" If you model that for several different values of X, you can then decide which one to do.
How else would you decide?
That's one possible purpose of modelling.
Another purpose would be entirely causally different i.e. "what will happen that we might then have to do something about".
I've rejoined the Tory Party to partake in this leadership contest.
Truss for me, us republican Tories need to stick together.
How long until you qualify to vote in the leadership election - as that could easily be early January at the present rate.
Three months. But I may get an immediate vote, I was a member for 23 years, and my membership only lapsed last year.
FWIW - I think Boris Johnson is secure until May at least, the locals might be the tipping point if it is a very bad night.
Still has to be a danger that there is a meltdown over COVID restrictions. They kicked the can today, but seems like there isn't much collective responsibility in the cabinet.
Actually I think Boris Johnson has played a blinder today.
If Omicron is as bad as SAGE think it will be, he'll put the blame squarely on the cabinet refuseniks.
According to @JosiasJessop we will never know how bad SAGE think it will be, as they're under no obligation or responsibility to say how it will be.
Their only responsibility apparently is to forward the worst case catastrophising they can come up with, while anything that they think will happen should be binned if it isn't a catastrophe.
So maybe no action is entirely the right response to how bad they think it will be and we should have that attitude towards SAGE indefinitely going forwards because apparently they'll never tell us if things are going to be OK in their expectations.
I really, really think you ned to go back to the last thread, read, and think. That's not what I'm saying at all.
Please don't misrepresent me.
You were saying, were you not, that only the catastrophic scenarios should be presented?
The scenario of what they think will happen, using the most accurate data available, as explained well by @MaxPB above was not in the SAGE scenarios. Why not? Why should they not present what they think will happen, using the best available data?
There isn't one "never event" in question here, there's two. And under the principle that advisors advise and ministers decide, the best available modelling should be presented without prejudice to what outcome it gives so that the ministers can make an informed decision.
If you wish to provide worst case scenarios AS WELL as central scenarios that's entirely reasonable. And since we have two worst cases here (an NHS collapse, and an unnecessary lockdown) it would make sense to show what assumptions are required for scenarios for both of those.
Instead the results from one catastrophe has been cherry picked without the central scenarios being shown, or the other side.
You and Max keep using the word "scenarios" wrong (or at least in a sense I have never seen it used, in modelling terms). The "scenarios" are things like: Do Nothing, Impose Some Restrictions, Full Lockdown. The complaint you are making seems to be that the model input parameters are biased towards pessimism, because the modellers are excluding data sources which would result in more optimistic selections.
There is no value in modelling scenarios in which we remove restrictions, because no-one is suggesting that. There is a good argument (an excellent one, actually) that we should be modelling based on realistic, rather than pessimistic, parameters, because otherwise we are overstating the chances of Bad Outcomes, and hence giving them too much weight when making policy decisions. However, it is wrong to say that they aren't capturing non-catastrophic scenarios, because the models are stochastic by nature, and are designed to model all possible future outcomes. In addition, if the secondary uncertainty around the mean parameters is high enough, you are also capturing the possibility that the more favourable view of the parameters is actually true. It's just that, because the mean parameters are artificially pessimistic, the midpoint of the ranges produced, and - especially - the worst case outcomes, will look worse than if you picked something more "realistic".
No you've missed the point. Absolutely I 100% agree that the scenarios should be on things like "do nothing" etc.
The problem isn't that the optimistic scenarios have been excluded, the problem is that the baseline scenarios have been excluded. As per Fraser Nelson's article the evidence on the severity of Omicron, which is already available from South Africa and has been used in the JP Morgan modelling wasn't used by SAGE.
That scenario was excluded as Medley said because the results didn't fit. They were only presenting catastrophe scenarios, so baseline scenarios (not optimistic ones) were excluded if they didn't show a catastrophe.
There's more other false and dodgy assumptions as Max has gone into great detail with.
This isn't about saying we should be cherry-picking optimistic or pessimistic, we shouldn't be doing either. Reasonable scenarios using accurate data weren't included in the data dump because they didn't present a catastrophe - that's not science that's fiddling with data.
OK, we're more or less agreeing now, except you're still stuck on using "scenarios" to mean "input parameters". Omicron being less severe than Delta is not a scenario; it's something the modellers have to make a call on either way, and they have chosen to downweight the evidence that it might be more benign, and cover that by putting in a large secondary uncertainty around their mean estimate. That is not the same as "excluding baseline scenarios", whatever that means. The wide ranges given show that all future outcomes are still included, it's just that the probabilities are different than if they'd not downweighted the evidence for Omicron being benign.
It is a scenario, a plausible one. The point of having a logistic regression model is that you can change multiple inputs and see what happens in each different scenario. The government modellers have been asked by some unknown party to not create scenarios where Omicron doesn't cause the NHS to collapse despite there being many plausible, likely I'd say at this point, ones where it won't.
No, I don't think that's true. I think you are basing this on the off-the-cuff Twitter exchange between Fraser Nelson and Graham Medley, and I think this is based on the two sides not understanding what the other means by "scenario".
I've rejoined the Tory Party to partake in this leadership contest.
Truss for me, us republican Tories need to stick together.
How long until you qualify to vote in the leadership election - as that could easily be early January at the present rate.
Three months. But I may get an immediate vote, I was a member for 23 years, and my membership only lapsed last year.
FWIW - I think Boris Johnson is secure until May at least, the locals might be the tipping point if it is a very bad night.
Still has to be a danger that there is a meltdown over COVID restrictions. They kicked the can today, but seems like there isn't much collective responsibility in the cabinet.
Actually I think Boris Johnson has played a blinder today.
If Omicron is as bad as SAGE think it will be, he'll put the blame squarely on the cabinet refuseniks.
According to @JosiasJessop we will never know how bad SAGE think it will be, as they're under no obligation or responsibility to say how it will be.
Their only responsibility apparently is to forward the worst case catastrophising they can come up with, while anything that they think will happen should be binned if it isn't a catastrophe.
So maybe no action is entirely the right response to how bad they think it will be and we should have that attitude towards SAGE indefinitely going forwards because apparently they'll never tell us if things are going to be OK in their expectations.
I really, really think you ned to go back to the last thread, read, and think. That's not what I'm saying at all.
Please don't misrepresent me.
You were saying, were you not, that only the catastrophic scenarios should be presented?
The scenario of what they think will happen, using the most accurate data available, as explained well by @MaxPB above was not in the SAGE scenarios. Why not? Why should they not present what they think will happen, using the best available data?
There isn't one "never event" in question here, there's two. And under the principle that advisors advise and ministers decide, the best available modelling should be presented without prejudice to what outcome it gives so that the ministers can make an informed decision.
If you wish to provide worst case scenarios AS WELL as central scenarios that's entirely reasonable. And since we have two worst cases here (an NHS collapse, and an unnecessary lockdown) it would make sense to show what assumptions are required for scenarios for both of those.
Instead the results from one catastrophe has been cherry picked without the central scenarios being shown, or the other side.
You and Max keep using the word "scenarios" wrong (or at least in a sense I have never seen it used, in modelling terms). The "scenarios" are things like: Do Nothing, Impose Some Restrictions, Full Lockdown. The complaint you are making seems to be that the model input parameters are biased towards pessimism, because the modellers are excluding data sources which would result in more optimistic selections.
There is no value in modelling scenarios in which we remove restrictions, because no-one is suggesting that. There is a good argument (an excellent one, actually) that we should be modelling based on realistic, rather than pessimistic, parameters, because otherwise we are overstating the chances of Bad Outcomes, and hence giving them too much weight when making policy decisions. However, it is wrong to say that they aren't capturing non-catastrophic scenarios, because the models are stochastic by nature, and are designed to model all possible future outcomes. In addition, if the secondary uncertainty around the mean parameters is high enough, you are also capturing the possibility that the more favourable view of the parameters is actually true. It's just that, because the mean parameters are artificially pessimistic, the midpoint of the ranges produced, and - especially - the worst case outcomes, will look worse than if you picked something more "realistic".
The advisors should present all possible likely outcomes of their models including ones where no further action is needed or restrictions are completely lifted - so long as any of their models indicate these outcomes. They should advise which of the modelled outcomes are more likely or less likely but they should certainly not be excluding outcomes because they are not worst case. It is for elected politicians to make decisions about which models they wish to consider but they need to be shown the full selection and then make an informed decision on that basis.
Only showing the worst outcomes is the scientists forcing the politicians to follow a particular preferred route even if that outcome is unlikely and irrespective of what the costs are to the country. It is a ludicrous way to manage a pandemic and all it will do is destroy any last vestiges of trust in the modelling or the advice.
Part of the problem is that "there are a range of possible outcomes" is a little too nuanced for either politicians or journalists.
In presentations I used the diagram type which shows probability as a river branching (what's that called formally?) - width of branch proportionate to the probability.
That seems to embed the probability in non-mathematical minds
I've rejoined the Tory Party to partake in this leadership contest.
Truss for me, us republican Tories need to stick together.
How long until you qualify to vote in the leadership election - as that could easily be early January at the present rate.
Three months. But I may get an immediate vote, I was a member for 23 years, and my membership only lapsed last year.
FWIW - I think Boris Johnson is secure until May at least, the locals might be the tipping point if it is a very bad night.
Still has to be a danger that there is a meltdown over COVID restrictions. They kicked the can today, but seems like there isn't much collective responsibility in the cabinet.
Actually I think Boris Johnson has played a blinder today.
If Omicron is as bad as SAGE think it will be, he'll put the blame squarely on the cabinet refuseniks.
According to @JosiasJessop we will never know how bad SAGE think it will be, as they're under no obligation or responsibility to say how it will be.
Their only responsibility apparently is to forward the worst case catastrophising they can come up with, while anything that they think will happen should be binned if it isn't a catastrophe.
So maybe no action is entirely the right response to how bad they think it will be and we should have that attitude towards SAGE indefinitely going forwards because apparently they'll never tell us if things are going to be OK in their expectations.
I really, really think you ned to go back to the last thread, read, and think. That's not what I'm saying at all.
Please don't misrepresent me.
You were saying, were you not, that only the catastrophic scenarios should be presented?
The scenario of what they think will happen, using the most accurate data available, as explained well by @MaxPB above was not in the SAGE scenarios. Why not? Why should they not present what they think will happen, using the best available data?
There isn't one "never event" in question here, there's two. And under the principle that advisors advise and ministers decide, the best available modelling should be presented without prejudice to what outcome it gives so that the ministers can make an informed decision.
If you wish to provide worst case scenarios AS WELL as central scenarios that's entirely reasonable. And since we have two worst cases here (an NHS collapse, and an unnecessary lockdown) it would make sense to show what assumptions are required for scenarios for both of those.
Instead the results from one catastrophe has been cherry picked without the central scenarios being shown, or the other side.
You and Max keep using the word "scenarios" wrong (or at least in a sense I have never seen it used, in modelling terms). The "scenarios" are things like: Do Nothing, Impose Some Restrictions, Full Lockdown. The complaint you are making seems to be that the model input parameters are biased towards pessimism, because the modellers are excluding data sources which would result in more optimistic selections.
There is no value in modelling scenarios in which we remove restrictions, because no-one is suggesting that. There is a good argument (an excellent one, actually) that we should be modelling based on realistic, rather than pessimistic, parameters, because otherwise we are overstating the chances of Bad Outcomes, and hence giving them too much weight when making policy decisions. However, it is wrong to say that they aren't capturing non-catastrophic scenarios, because the models are stochastic by nature, and are designed to model all possible future outcomes. In addition, if the secondary uncertainty around the mean parameters is high enough, you are also capturing the possibility that the more favourable view of the parameters is actually true. It's just that, because the mean parameters are artificially pessimistic, the midpoint of the ranges produced, and - especially - the worst case outcomes, will look worse than if you picked something more "realistic".
The advisors should present all possible likely outcomes of their models including ones where no further action is needed or restrictions are completely lifted - so long as any of their models indicate these outcomes. They should advise which of the modelled outcomes are more likely or less likely but they should certainly not be excluding outcomes because they are not worst case. It is for elected politicians to make decisions about which models they wish to consider but they need to be shown the full selection and then make an informed decision on that basis.
Only showing the worst outcomes is the scientists forcing the politicians to follow a particular preferred route even if that outcome is unlikely and irrespective of what the costs are to the country. It is a ludicrous way to manage a pandemic and all it will do is destroy any last vestiges of trust in the modelling or the advice.
That's not how the models work (to my understanding). You feed in a set of inputs, which includes a particular policy decision, like Do Nothing, or Impose Lockdown. The model spits out a set of probability-weighted outcomes, and the outputs are summarised and presented to policymakers. All possible outcomes are included which are relevant to those inputs. You can't get the model to tell you whether you should take action or not, because that's an input to the model, not an outcome.
Deciding what answer you want before you do the modelling is not science. If that is what they are doing then we should chuck the whole lot out.
The purpose of the modelling is to answer questions like "If we do X, what will happen?" If you model that for several different values of X, you can then decide which one to do.
How else would you decide?
By making sure the values of X (and Y and Z) cover the full range of possibilities, not just the ones you think will produce the modelling outputs you want. By excluding the possibility that Omicron will not produce the same level of health issues as previous variants they are picking and choosing the characteristics of the variant to get the outputs they want. As I said, that is not science.
Meanwhile, in Holland it is illegal to go to a pub or a restaurant, shop for clothes or even (and I find this hard to believe) hardware, which has never been the case in the UK.
Perhaps laughing at Boris Johnson is all they have left?
One gets the impression Liz Truss is particularly disliked by Remainers because she used to be a Remainer herself and then switched to Leave as soon as the result was in.
Comments
They can, quite frankly, fuck off.
I don't mind doing an LFT test. I took one the other day when seeing my sister-in-law, who was flying to Australia the following day with her new baby and partner (who had not seen her own parents in two years.) I obviously didn't want to jeopardise that. (It surprised my father in law when my mother in law LFTd him - he had never heard of the process, despite working in a university. ) But testing every day is just overkill.
and then a scandal as in 2016 will knock one of the other 2 out.
Welcome to 2022 - you can just see it happening.
Their only responsibility apparently is to forward the worst case catastrophising they can come up with, while anything that they think will happen should be binned if it isn't a catastrophe.
So maybe no action is entirely the right response to how bad they think it will be and we should have that attitude towards SAGE indefinitely going forwards because apparently they'll never tell us if things are going to be OK in their expectations.
And to fair to him, Fiona was giving him shit about him never seeing his kids
Please don't misrepresent me.
"We're following the science"
Aye, by about 10 days.
https://twitter.com/donaeldunready/status/1473006295221428224
But this way, if things do go pear-shaped, It's Not Boris's Fault.
Pathetic adolescent attitude, but that's the most useful framework to understand the PM sometimes.
Conservative Annabel de Capell Brooke, who represents Oundle on North Northamptonshire Council, moved to Norfolk last month.
Ms de Capell Brooke, who only attended four meetings since April, told the BBC: "I've resigned. No story."
So a few rounds of voting until one candidates gets a majority of the votes of the CRG/ERG members.
https://coronavirus.data.gov.uk/details/cases?areaType=region&areaName=London
Likewise in Denmark and South Africa.
Even with the dodgy scenarios presented, that have been rightly called out for using false data.
Do we think they'll try again tomorrow to get Cabinet approval? I struggle to see how sufficient new data will be available then, which would rule out restrictions until after they're able to recall Parliament post-Christmas?
Fingers crossed London (non-incidental) hospitalisation data this week is not alarming, and we may just escape another lockdown.
The scenario of what they think will happen, using the most accurate data available, as explained well by @MaxPB above was not in the SAGE scenarios. Why not? Why should they not present what they think will happen, using the best available data?
There isn't one "never event" in question here, there's two. And under the principle that advisors advise and ministers decide, the best available modelling should be presented without prejudice to what outcome it gives so that the ministers can make an informed decision.
If you wish to provide worst case scenarios AS WELL as central scenarios that's entirely reasonable. And since we have two worst cases here (an NHS collapse, and an unnecessary lockdown) it would make sense to show what assumptions are required for scenarios for both of those.
Instead the results from one catastrophe has been cherry picked without the central scenarios being shown, or the other side.
It depends on who is getting infected.
And every day more vaccines come into effect.
What we can say is likely is that Omicron will burn through the anti-vaxxers quickly.
And then who does that leave for it.
If so, there is an obvious reason why they would not be included in a report of the 11th of December, which instead presented a range of possibilities. Which did include the one that actually came up, I note. I also noted that:
1 - The findings on it replicating in the bronchi rather than the lungs were published a few days ago. Once more, well after the 11th.
2 - We still don’t know the effective outcome of that; we have some ideas but not solid facts It’s also implausible to ask for data from the 16th and 18th of December to be in a report dated the 11th.
One would assume that newer information is also mentioned during the briefings.
And anyone claiming “7k central forecast” can only be getting the information from media and commentators rather than actually reading the reports in question. Otherwise they’d know what was actually presented and projected.
(In summary: a 1.1 R value as the most likely under many scenarios (the lowest figure), an included 1.5 R value projection, as Scotland had seen such an R value in the several days just leading up to the date of presentation, with this being the highest plausible route (and the one quoted by the Mail, Christopher Snowdon, and others as “central projection), and by obvious request, an R value of 2.0 which they dismissed as extremely unlikely except if waning was far greater than expected or for an unknown variant. If you say so. I’m withdrawing from this entire discussion. No point. It’s been explained that it’s not rogue scientists trying to hijack the nation, nor deep cover civil servants in the Health Department with a fetish for lockdowns dictating the parameters. But it’s obviously no point in discussing further.
Considering it plausible because it had literally just been happening in Scotland, although they were pointing more at the R=1.1 route.
As they said at the time and in the explanation of what each of the scenarios were.
I get why the media ignored the explanation (attention-grabbing) but if politicians who are being presented with the scenarios and having the experts answer questions on hand and still be unable to understand, then such idiots shouldn’t be anywhere near any position of power or decision-making.
Which should mean that daily variations have less effect.
There is no value in modelling scenarios in which we remove restrictions, because no-one is suggesting that. There is a good argument (an excellent one, actually) that we should be modelling based on realistic, rather than pessimistic, parameters, because otherwise we are overstating the chances of Bad Outcomes, and hence giving them too much weight when making policy decisions. However, it is wrong to say that they aren't capturing non-catastrophic scenarios, because the models are stochastic by nature, and are designed to model all possible future outcomes. In addition, if the secondary uncertainty around the mean parameters is high enough, you are also capturing the possibility that the more favourable view of the parameters is actually true. It's just that, because the mean parameters are artificially pessimistic, the midpoint of the ranges produced, and - especially - the worst case outcomes, will look worse than if you picked something more "realistic".
As for 'never' events: that depends on perspective. Someone who doesn't care how many people die from Covid as long as they can lead a normal life might have a different view from someone who thinks it is wise to try to reduce deaths where possible - and the costs of saving those lives should be weighed up against those lives. But if you go in saying we shouldn't be trying to save any lives, then that's a bogus calculation.
Here's the LibDem Councillor for Trumpington, who lived in Fife.
https://www.bbc.co.uk/news/uk-england-cambridgeshire-44047135
My sister, who lives in outer London in Beckenham, has unfortunately caught Omicron and been isolating but has also had her booster and hopes to join us for boxing day. We will be spending Christmas with our parents in Kent
https://twitter.com/WillHayCardiff/status/1473037532397019145
Only showing the worst outcomes is the scientists forcing the politicians to follow a particular preferred route even if that outcome is unlikely and irrespective of what the costs are to the country. It is a ludicrous way to manage a pandemic and all it will do is destroy any last vestiges of trust in the modelling or the advice.
https://twitter.com/WillHayCardiff/status/1473040486650847233?s=20
This sounds as if Drakey has finally freaked out.
Guards, dogs and barbed wire at Clawdd Offa.
We're lucky that the politicians ignored the summer forecasts because they'd have insisted on silly measures like indoor distancing and vaccine passports which have been of no help in Europe and in fact may have been counter productive as it hasn't allowed them to build a natural immunity wall in the unvaccinated.
As to your other points, your faith and trust in people who you have no possible way of removing from the levers of power is worrying. Politicians may be flawed (and this set particularly so) yet we still have some way of making the decision on whether or not we want to do what they say. The unnamed bod in the DoH putting his fingers on the scale and presenting the politicians with a range unlikely scenarios and not also presenting the likely scenarios, well we have no way of removing this person from power, no way of holding them accountable to us.
That makes me instinctively wary of of them and sceptical about what their motives are. I know what Labour's motives are, I know what the Tories motives are, they told me at length in ab election campaign. We have an unelected tyranny of experts, maybe you're happy to live like that, I'm not. Quite frankly, our views on this differ too much, I'm extremely worried that the UK and its people are too happy to simply hand our freedoms to these unknown experts with no guarantees they'll be happy to give them back again.
So yes, colour me very sceptical about anyone who is pushing a very narrow set of possible outcomes while excluding other more favourable ones. There's simply no need for it.
As to the other point, I understand the data came later than the report, that's a reason to tell the decision makers - "we have new key data inputs for our model, please await our new results while we rerun the scenarios with the new data" not a reason to run with the outdated model and present simply incorrect scenarios that will now never come to pass.
The problem isn't that the optimistic scenarios have been excluded, the problem is that the baseline scenarios have been excluded. As per Fraser Nelson's article the evidence on the severity of Omicron, which is already available from South Africa and has been used in the JP Morgan modelling wasn't used by SAGE.
That scenario was excluded as Medley said because the results didn't fit. They were only presenting catastrophe scenarios, so baseline scenarios (not optimistic ones) were excluded if they didn't show a catastrophe.
There's more other false and dodgy assumptions as Max has gone into great detail with.
This isn't about saying we should be cherry-picking optimistic or pessimistic, we shouldn't be doing either. Reasonable scenarios using accurate data weren't included in the data dump because they didn't present a catastrophe - that's not science that's fiddling with data.
The data from Guateng is that hospitalisations of people *for* CV19/Omicron are pretty low.
The one to watch at the moment is Regional R - signs of reaching a peak...
There are a lot of delusions out there, urged on by nutters on the internet.
Its for the ministers to decide what outcome they take, not the scientists. That's why they should be given the full picture in order to make an educated decision, not half a picture presenting only the extremely gloomiest scenarios.
In good science you present all the relevant information, not just that which suits your agenda. That includes showing the scenarios [and their assumptions] that mean no action is necessary.
How else would you decide?
What concerns me, is we don’t even know the name of the person or persons “putting their fingers on the scale”, much less take a view on their motivation. They may just be a bit misguided but trying to what’s right, or poor communicators. But they may not.
The evidence shows that it might be more benign so that input parameter should be presented as one of the scenarios shouldn't it?
Similarly with VE they've only presented more negative figures than what the evidence presents.
When questioned on why not to present the potentially more accurate and more scientific scenario of input parameters using the best known data, its being excluded based upon the results.
It is not scientific to keep changing your input parameters until you get an output that you are happy with. You might as well just say "I want a lockdown" and not bother with any models if you want to do that. Create your parameters, run your models and show the outcomes regardless of their outputs. If they're negative, they're negative. If they're positive, they're positive. If they're neutral, they're neutral. At least then you're not forcing your own agenda onto the science, let the science speak for itself.
Another purpose would be entirely causally different i.e. "what will happen that we might then have to do something about".
That seems to embed the probability in non-mathematical minds