"When Manchester and Yorkshire were put into Tier 3 restrictions, there were plenty of graphs looking at the virus – but none looking at the side-effects."
"Sweden is perhaps the first country in the world to make this case so clearly: isolation kills too"
I'm sticking resolutely to my prediction that Biden will win the electoral college by 335-203 but still see Trump stubbornly between 44-46%.
I think Trump will carry Texas by about 3/4%, Iowa, Ohio and Georgia by 2% each.
I think Florida and North Carolina will be the closest states and Biden will only edge them out by 1% but I would not be surprised for Trump to win either.
Arizona, I think Biden will win by 2/3% and Michigan, Pennsylvania and Wisconsin by 5/6%.
I'm sticking resolutely to my prediction that Biden will win the electoral college by 335-203 but still see Trump stubbornly between 44-46%.
I think Trump will carry Texas by about 3/4%, Iowa, Ohio and Georgia by 2% each.
I think Florida and North Carolina will be the closest states and Biden will only edge them out by 1% but I would not be surprised for Trump to win either.
Arizona, I think Biden will win by 2/3% and Michigan, Pennsylvania and Wisconsin by 5/6%.
He should carry NE2 and will probably take ME2.
I agree with most of that though I think Trump will hold Michigan and/or Wisconsin and Trump will edge Florida by a very narrow margin and North Carolina by a slightly bigger margin
Couple of cutting edge thriller tips from me. Ruth Rendell. Barbara Vine.
Thanks - which particular books from those two authors?
Please try Crooked Letter, Crooked Letter.
Yes, I will.
My 2 are actually the same author. Rendell writes under Vine for her deeper 1st person narrative psychologicals.
What I like is the deceptively skillful prose and story telling. The horror creeps up on you without you noticing and you often end up relating to the deranged protagonist because she makes the weird seem oddly rational.
Also many are set in my part of London and you can read one in a couple of days.
I used to be very keen on Rendell/Vine, read them all I think (and there aint going to be any more obvs, unless some ghastly ghostwriting moneymaking thing happens). I have to say latterly I went off the Vine books a bit; there seemed to me to be a misanthropic, contemptuous tinge, some of it directed at people at the bottom of the pecking order. Also her ornate character names got a bit irritating.
It's been a while - like you, I think I've read most of them - but I don't recall picking up on that. Maybe I would if I re-read them now. Unlikely I will be, given all the stuff I haven't read.
You have changed your logo, I now see. Does this mean the West is no longer doomed?
Och it probably still is, but f.all to do with the statchoos that certain types were hystericising over a few months ago. I wonder if any of them now feel a scintilla of embarrassment about their screeching and wailing? Probably not.
I just watched the clip, and I don't hear anything about 6 tiers. What am I missing here?
Which of the 5 tiers they announced covers Full Lockdown?
Therefore, there exists at least 1 more tier...
Yes, I can invent extra tiers for rhetorical effect, too: -2 = mandatory kissing -1 = mandatory pubgoing ... 5 = full lockdown, restrictions on going out of doors 6 = your apartment will be sealed in plastic 7 = purge everything with fire
...but I won't pretend that's what Sturgeon just said
Does anyone have a view as to medium-term GBP value post Covid-19 assuming a full FTA Brexit deal?
My thinking is GBP/EUR approaching 1.25 and GBP/UDS flirting with 1.40 but others may have better insight.
No idea but my UK pensions sent monthly to Spain would love that 1.25 idea! Truth is I've watched currencies trading for years and predictions are really difficult.
Blimey. That’s a map. Almost exactly traces the linguistic boundary.
By my soul, 'tis a Protestant virus.
Protestants are only abut 3% of the Belgian population. I know this because I made exactly the same point "in another place" earlier this week and was corrected.
@HYUFD - I really can't believe you are posting these posts. Firemen put out fires, train drivers drive trains and pollsters carry out polls.
If you don't drive a train you are not a train driver. Same goes for polls
I do struggle a bit with the 'probablistic' thing when it comes to Presidential elections. Its not like rolling a over and over again because its one roll every four years and the dice is a radically different shape every time. What does it really mean to say Clinton would have won 70% of the time in 2016, that dice will never be rolled again. So any result that turns up the maths guys can say, well we never said that was impossible just unlikely. Silver talks about this stuff himself in The Signal and the Noise, how impossible it is to call the path of hurricanes etc. Having said that I would rather tell the time with a clock that was always wrong by say10 minutes slow or 10 minutes fast rather than a stopped one that was precisely right twice a day.
Ah yes, this has been discussed in many universities over the last 100+ years. You are using the frequentist approach to probability, which doesn't work well in this kind of scenario. Bayesian and De Fenetti probability works much better, but I don't have time to discuss the differences now.
I get the general principles of the Bayesian thing, priors etc, that is Silvers big driver too, but on here everyone hits HYFUD with the dice analogy and I just think that's kind of the wrong metaphor.
I have had all kinds of problems with the Monty Hall Goats thing too btw, so free tutorials always welcome
What would be interesting, and I don't know whether it's been done or not, would be to re-run the 538 predictions in 2016, but with modifications to opinion polls for the undersampling of High School or less voters.
That way you'd have an estimate of how wrong the prediction was for that reason.
One of the problems I have with probabilistic forecasts is that I can't help but reduce them to a deterministic forecast with a certainty attached. But they have a lot more information than that.
They show where that uncertainty is, and what differences in the detail have the most impact on the overall result.
I was wondering something sort of similar - might be a dumb question - but could the pollsters apply their 2016 weighting methodology to the current results they are seeing and show what the old methodology would have been saying now in terms of predicted results? would be interesting to see how much their changes have impacted the results they are showing
Yes, and all of the groups producing ECV predictions have definitely done this.
While it would be interesting to see for a few of us from PB, I totally understand that they want to keep their model assesments on historic data in house. There is too much scope for others to nick new ideas, and for dumb journalists to run dumb stories like "538 now predicing that Trump will win the 2016 election".
Blimey. That’s a map. Almost exactly traces the linguistic boundary.
By my soul, 'tis a Protestant virus.
Protestants are only abut 3% of the Belgian population. I know this because I made exactly the same point "in another place" earlier this week and was corrected.
Germanic language-family speakers are less "touchy-feely" culturally than Romance language-family speakers?
Good to see Mrs Truss earning her air miles. Well done.
Genuine question: how does the UK/Japan deal differ from the EU/Japan deal?
Virtually the same, which is actually a good result for the UK.
Content rules for digital disapplied as only of interest to the French; data protections disapplied - whether that's a good or bad thing is a chlorine chicken type question. The UK doesn't get the tariff quotas that the EU negotiated but can use any leftover EU quota. Opportunity to talk about financial services - I suspect that's a tactful refusal by Japan to include the UK ask in this FTA.
Through gritted teeth, given you clearly want to put the worst spin on it possible, that's actually high-praise from you.
Does anyone have a view as to medium-term GBP value post Covid-19 assuming a full FTA Brexit deal?
My thinking is GBP/EUR approaching 1.25 and GBP/UDS flirting with 1.40 but others may have better insight.
GBPEUR fell to 1.20 in the immediate aftermath of the Brexit vote, when the assumption of most was a very close relationship or even that Brexit might not happen. When May signalled single market and customs union exit in late 2016 it fell to around 1.10. We are now hoping for an FTA even less advantageous than May's proposals. On that basis the current 1.10 level doesn't look absurd. If we assume that builds in some no deal premium then there is some room for upside, so I could see 1.15 or maybe a bit above once a deal is done. I don't think the market believes there will really be no deal. We were at 1.20 in February but that was post-election euphoria, and the govt doubling down on a really minimalist deal combined with the Covid shambles and ballooning borrowing has I think pushed GBP lower and I think some of that is fairly permanent. Meanwhile, market sentiment on EUR has been boosted by Covid since (probably wrongly) people see the recovery plan deal as the start of more effective fiscal risk sharing. Both the UK and EU are expected to more QE so that's probably a wash currency wise. There are certainly some pro-GBP voices in the market, who argue that the Brexit impact is exaggerated and the UK fundamentals are better. They've been saying that since 2016 though, and haven't been right yet.
On "$markets" a No Deal WTO Brexit is now a 33% chance.
That's a massive overstatement imo. I judge it quite close to zero.
I would put it below 33% but above 0%. Maybe 15%. It would be insane for both parties given Covid. But it could happen because both parties overestimate their own negotiating position. The EU in particular is being too stubborn. They are right to think they hold most of the cards, but they don't hold all of them.
I'll stretch to 10. On the politics of it, though, I dislike the fevered speculation about No Deal. Not just because I think it's not much of a possibility but because it plays well for the ghastly Johnson. The more people think No Deal might happen the more relief there will be when it doesn't. And Johnson will surf on this to aid his inevitable "Look, we've done it again when everyone said we couldn't. That's what happens when you hang TOUGH and battle for Britain." Just like last time in other words. "It's a triumph!" And we know what that led to. The big 80. So I'd rather people just assumed there will be a Deal - "easiest one in history" after all - and all we await is the detail of the wretched thing.
Reminds me of a telephone tipster scam that I thought of years ago but never implemented. You get a bunch of people (say 80) to call a number for a racing tip. You pick , say, a 8 runner race where there ideally is no clear favourite. You split the 80 callers into 8 groups of 10 and give each group a different tip so that you have covered the field. You then start charging the 10 who were on the winner for future tips cus they think you are a genius.
I used to work with/for racing tipsters, and there were some of the naughtier ones who did do something along those lines.
Blimey. That’s a map. Almost exactly traces the linguistic boundary.
By my soul, 'tis a Protestant virus.
Protestants are only abut 3% of the Belgian population. I know this because I made exactly the same point "in another place" earlier this week and was corrected.
Germanic language-family speakers are less "touchy-feely" culturally than Romance language-family speakers?
I have lived in Havering almost all my life and in that time my address has always contained "Essex", and never London. @Stuartinromford is right, most people moved here from London and don't want London following them. It is though. We are probably going to move out to Non London Essex in the next few years. Our son was born in Broomfield Hospital in Essex but had to go to Queens in Romford after 4 days -the contrast was striking, enough to make me realise I was happier in Essex than London.
Yes, but the irony is that this process will (and already is) make Havering more like the rest of East London
Yeah but if I don't live there anymore I don't care!
EastEnders demographic is Havering although it pretends to be East London. I happen to think this is kind of racist of the BBC, and my guess is that they dont think the rest of the country would believe it if they accurately represented Stratford/Walthamstow
iirc Eastenders was originally based on Hoxton (which is not really in the East End anyway but never mind).
Can't resist commenting on this, as I've lived in Romford for 25 years and occasionally my path crosses with Mr Rosindell's.
Certainly among the older population there is a feeling that Romford is Essex not London. I'm not sure they'd be quite so keen if their TfL free travel was taken away.
There has been a big change in central Romford over the last few years, it now feels much more like the rest of London. I keep hearing from people that 'Romford's going downhill' and 'it isn't like it used to be' but I've been here a long time and it was always a bit scruffy in places. Scratch the surface and what many people mean is that there are more non-white faces around, and they're off to Clacton.
The rest of Havering is still very much 'East Enders', albeit changing gradually too.
25 years ago having a stall on Romford market was a very good business to be in and there were no empty spaces. Now the market is half the size and what’s left is cheap tat. That’s probably why a lot of lifelong residents think it’s gone downhill - it’s a pretty good indicator. I used to love going to Romford when I was young, it’s 8 mins on the train from me, I’ve not been in 5-6 years and wouldn’t care if I never did again.
Blimey. That’s a map. Almost exactly traces the linguistic boundary.
By my soul, 'tis a Protestant virus.
Protestants are only abut 3% of the Belgian population. I know this because I made exactly the same point "in another place" earlier this week and was corrected.
Germanic language-family speakers are less "touchy-feely" culturally than Romance language-family speakers?
It is probably just money.
Flanders is richer, and so there is less deprivation. The same trend as seen everywhere else.
Unexpected R figures. The UK GOV dashboard shows cases rising in SE/SW and E England. Possibly flat or falling in areas with serious student out breaks, eg NW/NE England. (Which also might be a temporary blip, ie student outbreak dealt with, now increasing organically at a lower rate)
@HYUFD - I really can't believe you are posting these posts. Firemen put out fires, train drivers drive trains and pollsters carry out polls.
If you don't drive a train you are not a train driver. Same goes for polls
I do struggle a bit with the 'probablistic' thing when it comes to Presidential elections. Its not like rolling a over and over again because its one roll every four years and the dice is a radically different shape every time. What does it really mean to say Clinton would have won 70% of the time in 2016, that dice will never be rolled again. So any result that turns up the maths guys can say, well we never said that was impossible just unlikely. Silver talks about this stuff himself in The Signal and the Noise, how impossible it is to call the path of hurricanes etc. Having said that I would rather tell the time with a clock that was always wrong by say10 minutes slow or 10 minutes fast rather than a stopped one that was precisely right twice a day.
Ah yes, this has been discussed in many universities over the last 100+ years. You are using the frequentist approach to probability, which doesn't work well in this kind of scenario. Bayesian and De Fenetti probability works much better, but I don't have time to discuss the differences now.
I get the general principles of the Bayesian thing, priors etc, that is Silvers big driver too, but on here everyone hits HYFUD with the dice analogy and I just think that's kind of the wrong metaphor.
I have had all kinds of problems with the Monty Hall Goats thing too btw, so free tutorials always welcome
What would be interesting, and I don't know whether it's been done or not, would be to re-run the 538 predictions in 2016, but with modifications to opinion polls for the undersampling of High School or less voters.
That way you'd have an estimate of how wrong the prediction was for that reason.
One of the problems I have with probabilistic forecasts is that I can't help but reduce them to a deterministic forecast with a certainty attached. But they have a lot more information than that.
They show where that uncertainty is, and what differences in the detail have the most impact on the overall result.
I was wondering something sort of similar - might be a dumb question - but could the pollsters apply their 2016 weighting methodology to the current results they are seeing and show what the old methodology would have been saying now in terms of predicted results? would be interesting to see how much their changes have impacted the results they are showing
Yes, and all of the groups producing ECV predictions have definitely done this.
While it would be interesting to see for a few of us from PB, I totally understand that they want to keep their model assesments on historic data in house. There is too much scope for others to nick new ideas, and for dumb journalists to run dumb stories like "538 now predicing that Trump will win the 2016 election".
I would be more interested in the 'if we used 2016 Methodology we would have Biden ahead by an additional x points" / or not
US retailer Gap could close all of its own UK stores, putting thousands of jobs at risk, as it mulls shifting its operations to franchise-only in Europe. It may close all of its shops in the UK, France, Ireland and Italy next summer, the retailer said.
Unexpected R figures. The UK GOV dashboard shows cases rising in SE/SW and E England. Possibly flat or falling in areas with serious student out breaks, eg NW/NE England. (Which also might be a temporary blip, ie student outbreak dealt with, now increasing organically at a lower rate)
That's not R, it's incidence, the person on twitter didn't label it at all.
Why? The public is being sold a silly and lazy version of what science is, consisting mainly of debunking the myths and exposing Bad Science, yay! Let's all laugh at the flat earthers!
Of those 5
Causes Covid - jury out, probably not but I'd like to see some RCTs.
Joking here, but the Yay for Ther Science lot are usually primed to squawk ANIKDOTE at any theory not derived from an RCT, so why not here?
Used for spying - unquestionably the case, one way or another.
Trees being cut down for it - seems inherently likely, got to put masts somewhere, no evidence to the contrary presented.
5G is making the human population stupid - it is. It makes them spend more time looking at cat videos on facebook, and having silly arguments about the effects of 5G.
11% say they believe that 5G radiation lowers human immune defences with Bristol residents buying into this theory the most (18%). A further 10% think 5G causes cancer, brain tumours or infertility
This could easily be true, and is taken very seriously by real live mainstream scientists.
@HYUFD - I really can't believe you are posting these posts. Firemen put out fires, train drivers drive trains and pollsters carry out polls.
If you don't drive a train you are not a train driver. Same goes for polls
I do struggle a bit with the 'probablistic' thing when it comes to Presidential elections. Its not like rolling a over and over again because its one roll every four years and the dice is a radically different shape every time. What does it really mean to say Clinton would have won 70% of the time in 2016, that dice will never be rolled again. So any result that turns up the maths guys can say, well we never said that was impossible just unlikely. Silver talks about this stuff himself in The Signal and the Noise, how impossible it is to call the path of hurricanes etc. Having said that I would rather tell the time with a clock that was always wrong by say10 minutes slow or 10 minutes fast rather than a stopped one that was precisely right twice a day.
Ah yes, this has been discussed in many universities over the last 100+ years. You are using the frequentist approach to probability, which doesn't work well in this kind of scenario. Bayesian and De Fenetti probability works much better, but I don't have time to discuss the differences now.
I get the general principles of the Bayesian thing, priors etc, that is Silvers big driver too, but on here everyone hits HYFUD with the dice analogy and I just think that's kind of the wrong metaphor.
I have had all kinds of problems with the Monty Hall Goats thing too btw, so free tutorials always welcome
That is because the Monty Hall thing (as usually expressed) is ambiguous and depends on whether MH has prior knowledge, and tends to have a different answer on either side of the Atlantic.
Agree about the Monty Hall problem. I always need to know the rule about which door Monty opens. You only get the surprising resukt though, if his rule is "never open the door with a car behind it"
Which MUST have been the rule surely? Otherwise what's the point?
Unexpected R figures. The UK GOV dashboard shows cases rising in SE/SW and E England. Possibly flat or falling in areas with serious student out breaks, eg NW/NE England. (Which also might be a temporary blip, ie student outbreak dealt with, now increasing organically at a lower rate)
That's not R, it's incidence, the person on twitter didn't label it at all.
Blimey. That’s a map. Almost exactly traces the linguistic boundary.
By my soul, 'tis a Protestant virus.
Protestants are only abut 3% of the Belgian population. I know this because I made exactly the same point "in another place" earlier this week and was corrected.
Germanic language-family speakers are less "touchy-feely" culturally than Romance language-family speakers?
Good to see Mrs Truss earning her air miles. Well done.
Genuine question: how does the UK/Japan deal differ from the EU/Japan deal?
UK conceded a little ground on tariffs and Rules of Origin content in eg cars and railway equipment (accepting 50% rather than 55% Japanese) and struck a better deal for Stilton because the EU didn't use all of its equivalent product quota.
Overall estimate is £1.5billion improvement over no deal at all, conveniently forgetting the improvement the recent EU-Japan deal was bringing us since Feb 2020.
Blimey. That’s a map. Almost exactly traces the linguistic boundary.
By my soul, 'tis a Protestant virus.
Protestants are only abut 3% of the Belgian population. I know this because I made exactly the same point "in another place" earlier this week and was corrected.
Indeed. Catholicism is a key reason Flanders is not just part of the Netherlands.
Does anyone have a view as to medium-term GBP value post Covid-19 assuming a full FTA Brexit deal?
My thinking is GBP/EUR approaching 1.25 and GBP/UDS flirting with 1.40 but others may have better insight.
GBPEUR fell to 1.20 in the immediate aftermath of the Brexit vote, when the assumption of most was a very close relationship or even that Brexit might not happen. When May signalled single market and customs union exit in late 2016 it fell to around 1.10. We are now hoping for an FTA even less advantageous than May's proposals. On that basis the current 1.10 level doesn't look absurd. If we assume that builds in some no deal premium then there is some room for upside, so I could see 1.15 or maybe a bit above once a deal is done. I don't think the market believes there will really be no deal. We were at 1.20 in February but that was post-election euphoria, and the govt doubling down on a really minimalist deal combined with the Covid shambles and ballooning borrowing has I think pushed GBP lower and I think some of that is fairly permanent. Meanwhile, market sentiment on EUR has been boosted by Covid since (probably wrongly) people see the recovery plan deal as the start of more effective fiscal risk sharing. Both the UK and EU are expected to more QE so that's probably a wash currency wise. There are certainly some pro-GBP voices in the market, who argue that the Brexit impact is exaggerated and the UK fundamentals are better. They've been saying that since 2016 though, and haven't been right yet.
On "$markets" a No Deal WTO Brexit is now a 33% chance.
That's a massive overstatement imo. I judge it quite close to zero.
I would put it below 33% but above 0%. Maybe 15%. It would be insane for both parties given Covid. But it could happen because both parties overestimate their own negotiating position. The EU in particular is being too stubborn. They are right to think they hold most of the cards, but they don't hold all of them.
I'll stretch to 10. On the politics of it, though, I dislike the fevered speculation about No Deal. Not just because I think it's not much of a possibility but because it plays well for the ghastly Johnson. The more people think No Deal might happen the more relief there will be when it doesn't. And Johnson will surf on this to aid his inevitable "Look, we've done it again when everyone said we couldn't. That's what happens when you hang TOUGH and battle for Britain." Just like last time in other words. "It's a triumph!" And we know what that led to. The big 80. So I'd rather people just assumed there will be a Deal - "easiest one in history" after all - and all we await is the detail of the wretched thing.
Agreed on the optics. And a deal should certainly be the base case - we only disagree whether it is 85% or 90%. Whatever happens Johnson will lie about it and his supporters will believe him.
Blimey. That’s a map. Almost exactly traces the linguistic boundary.
By my soul, 'tis a Protestant virus.
Protestants are only abut 3% of the Belgian population. I know this because I made exactly the same point "in another place" earlier this week and was corrected.
But 68% of Wallonians are Roman Catholic, and the virus appears to be differentially hostile to them.
@HYUFD - I really can't believe you are posting these posts. Firemen put out fires, train drivers drive trains and pollsters carry out polls.
If you don't drive a train you are not a train driver. Same goes for polls
I do struggle a bit with the 'probablistic' thing when it comes to Presidential elections. Its not like rolling a over and over again because its one roll every four years and the dice is a radically different shape every time. What does it really mean to say Clinton would have won 70% of the time in 2016, that dice will never be rolled again. So any result that turns up the maths guys can say, well we never said that was impossible just unlikely. Silver talks about this stuff himself in The Signal and the Noise, how impossible it is to call the path of hurricanes etc. Having said that I would rather tell the time with a clock that was always wrong by say10 minutes slow or 10 minutes fast rather than a stopped one that was precisely right twice a day.
Ah yes, this has been discussed in many universities over the last 100+ years. You are using the frequentist approach to probability, which doesn't work well in this kind of scenario. Bayesian and De Fenetti probability works much better, but I don't have time to discuss the differences now.
I get the general principles of the Bayesian thing, priors etc, that is Silvers big driver too, but on here everyone hits HYFUD with the dice analogy and I just think that's kind of the wrong metaphor.
I have had all kinds of problems with the Monty Hall Goats thing too btw, so free tutorials always welcome
That is because the Monty Hall thing (as usually expressed) is ambiguous and depends on whether MH has prior knowledge, and tends to have a different answer on either side of the Atlantic.
Agree about the Monty Hall problem. I always need to know the rule about which door Monty opens. You only get the surprising resukt though, if his rule is "never open the door with a car behind it"
Which MUST have been the rule surely? Otherwise what's the point?
James May fixed this for me with toy cars, they literally did the stick twist with dozens and dozens of cars and goats and the guy twisting ended up winning way more cars - sadly cant find the video. Still does my head in though
Unexpected R figures. The UK GOV dashboard shows cases rising in SE/SW and E England. Possibly flat or falling in areas with serious student out breaks, eg NW/NE England. (Which also might be a temporary blip, ie student outbreak dealt with, now increasing organically at a lower rate)
That's not R, it's incidence, the person on twitter didn't label it at all.
Quite so - an appalling tweet with an unlabeled chart. If that was really the R rate, we could all relax quite a lot. But it isn't.
Blimey. That’s a map. Almost exactly traces the linguistic boundary.
By my soul, 'tis a Protestant virus.
Protestants are only abut 3% of the Belgian population. I know this because I made exactly the same point "in another place" earlier this week and was corrected.
Germanic language-family speakers are less "touchy-feely" culturally than Romance language-family speakers?
Ah, but English is a Germanic language..
Expatiate on the rationale for that dictum?
What counts in classifying languages into families is grammar and core vocabulary. I forget the exact figures but something like 48 of the 50 commonest English words are Germanic in origin. Of course, flip through a dictionary and you'll find Latin words aplenty, perhaps many times more than Germanic ones. But those are the leaves on the tree, whereas the trunk and roots are emphatically Germanic.
Could be to conform to the Whitehall system, sufficiently so to get qualification for employers' support un der Mr Sunak's tiers.
I see she has in any case added a lower tier. And so on.
It makes complete sense to treat different parts of a country differently according to the level of threat. It's why Sir Keir's Circuit Break nonsense was nonsense - he probably only did it to play politics anyway I suppose.
Do you think she did 5 tiers rather than 3 to differentiate Scotland from England?
James May fixed this for me with toy cars, they literally did the stick twist with dozens and dozens of cars and goats and the guy twisting ended up winning way more cars - sadly cant find the video. Still does my head in though
Blimey. That’s a map. Almost exactly traces the linguistic boundary.
By my soul, 'tis a Protestant virus.
Protestants are only abut 3% of the Belgian population. I know this because I made exactly the same point "in another place" earlier this week and was corrected.
Germanic language-family speakers are less "touchy-feely" culturally than Romance language-family speakers?
Wallonia is the French AND German (and Walloon) speaking region.
Blimey. That’s a map. Almost exactly traces the linguistic boundary.
By my soul, 'tis a Protestant virus.
Protestants are only abut 3% of the Belgian population. I know this because I made exactly the same point "in another place" earlier this week and was corrected.
Germanic language-family speakers are less "touchy-feely" culturally than Romance language-family speakers?
Ah, but English is a Germanic language..
Expatiate on the rationale for that dictum?
What counts in classifying languages into families is grammar and core vocabulary. I forget the exact figures but something like 48 of the 50 commonest English words are Germanic in origin. Of course, flip through a dictionary and you'll find Latin words aplenty, perhaps many times more than Germanic ones. But those are the leaves on the tree, whereas the trunk and roots are emphatically Germanic.
@HYUFD - I really can't believe you are posting these posts. Firemen put out fires, train drivers drive trains and pollsters carry out polls.
If you don't drive a train you are not a train driver. Same goes for polls
I do struggle a bit with the 'probablistic' thing when it comes to Presidential elections. Its not like rolling a over and over again because its one roll every four years and the dice is a radically different shape every time. What does it really mean to say Clinton would have won 70% of the time in 2016, that dice will never be rolled again. So any result that turns up the maths guys can say, well we never said that was impossible just unlikely. Silver talks about this stuff himself in The Signal and the Noise, how impossible it is to call the path of hurricanes etc. Having said that I would rather tell the time with a clock that was always wrong by say10 minutes slow or 10 minutes fast rather than a stopped one that was precisely right twice a day.
Ah yes, this has been discussed in many universities over the last 100+ years. You are using the frequentist approach to probability, which doesn't work well in this kind of scenario. Bayesian and De Fenetti probability works much better, but I don't have time to discuss the differences now.
I get the general principles of the Bayesian thing, priors etc, that is Silvers big driver too, but on here everyone hits HYFUD with the dice analogy and I just think that's kind of the wrong metaphor.
I have had all kinds of problems with the Monty Hall Goats thing too btw, so free tutorials always welcome
What would be interesting, and I don't know whether it's been done or not, would be to re-run the 538 predictions in 2016, but with modifications to opinion polls for the undersampling of High School or less voters.
That way you'd have an estimate of how wrong the prediction was for that reason.
One of the problems I have with probabilistic forecasts is that I can't help but reduce them to a deterministic forecast with a certainty attached. But they have a lot more information than that.
They show where that uncertainty is, and what differences in the detail have the most impact on the overall result.
I was wondering something sort of similar - might be a dumb question - but could the pollsters apply their 2016 weighting methodology to the current results they are seeing and show what the old methodology would have been saying now in terms of predicted results? would be interesting to see how much their changes have impacted the results they are showing
Yes, and all of the groups producing ECV predictions have definitely done this.
While it would be interesting to see for a few of us from PB, I totally understand that they want to keep their model assesments on historic data in house. There is too much scope for others to nick new ideas, and for dumb journalists to run dumb stories like "538 now predicing that Trump will win the 2016 election".
You can see approximately how SurveyUSA has changed between 2016 and 2020 as they have comprehensive cross tabs.
They key bit is their HS and less education component has gone from approx 12—16% to 30-38% in their polls.
Has a big effect on the results if you reweight their 2016 to 2020 and vis-versa.
@HYUFD - I really can't believe you are posting these posts. Firemen put out fires, train drivers drive trains and pollsters carry out polls.
If you don't drive a train you are not a train driver. Same goes for polls
I do struggle a bit with the 'probablistic' thing when it comes to Presidential elections. Its not like rolling a over and over again because its one roll every four years and the dice is a radically different shape every time. What does it really mean to say Clinton would have won 70% of the time in 2016, that dice will never be rolled again. So any result that turns up the maths guys can say, well we never said that was impossible just unlikely. Silver talks about this stuff himself in The Signal and the Noise, how impossible it is to call the path of hurricanes etc. Having said that I would rather tell the time with a clock that was always wrong by say10 minutes slow or 10 minutes fast rather than a stopped one that was precisely right twice a day.
Ah yes, this has been discussed in many universities over the last 100+ years. You are using the frequentist approach to probability, which doesn't work well in this kind of scenario. Bayesian and De Fenetti probability works much better, but I don't have time to discuss the differences now.
I get the general principles of the Bayesian thing, priors etc, that is Silvers big driver too, but on here everyone hits HYFUD with the dice analogy and I just think that's kind of the wrong metaphor.
I have had all kinds of problems with the Monty Hall Goats thing too btw, so free tutorials always welcome
That is because the Monty Hall thing (as usually expressed) is ambiguous and depends on whether MH has prior knowledge, and tends to have a different answer on either side of the Atlantic.
Agree about the Monty Hall problem. I always need to know the rule about which door Monty opens. You only get the surprising resukt though, if his rule is "never open the door with a car behind it"
Which MUST have been the rule surely? Otherwise what's the point?
The Monty Hall problem is one of my favourite things. It is a perfect example of how common sense can mislead you in areas like probability.
Just checking my own derived R against the ONS data and it's a pretty close match. Hopefully that has held true for the last week where there isn't any ONS coverage.
Is "Nippy" an affectionate name for Sturgeon or a putdown?
How about Wee Jimmy ?
Ah well that one is obviously nasty.
I find that one is also used by people who don't really care much about politics or Scottish politics. It's obviously not nice, but I don't think it marks one out as an opponent. Nippy does, to the extent I tend to discount what people have to say when they use that word; you probably aren't going to learn anything from someone who calls her Nippy. See also maomentum, selfservatives, libtards, and so on.
@HYUFD - I really can't believe you are posting these posts. Firemen put out fires, train drivers drive trains and pollsters carry out polls.
If you don't drive a train you are not a train driver. Same goes for polls
I do struggle a bit with the 'probablistic' thing when it comes to Presidential elections. Its not like rolling a over and over again because its one roll every four years and the dice is a radically different shape every time. What does it really mean to say Clinton would have won 70% of the time in 2016, that dice will never be rolled again. So any result that turns up the maths guys can say, well we never said that was impossible just unlikely. Silver talks about this stuff himself in The Signal and the Noise, how impossible it is to call the path of hurricanes etc. Having said that I would rather tell the time with a clock that was always wrong by say10 minutes slow or 10 minutes fast rather than a stopped one that was precisely right twice a day.
Ah yes, this has been discussed in many universities over the last 100+ years. You are using the frequentist approach to probability, which doesn't work well in this kind of scenario. Bayesian and De Fenetti probability works much better, but I don't have time to discuss the differences now.
I get the general principles of the Bayesian thing, priors etc, that is Silvers big driver too, but on here everyone hits HYFUD with the dice analogy and I just think that's kind of the wrong metaphor.
I have had all kinds of problems with the Monty Hall Goats thing too btw, so free tutorials always welcome
That is because the Monty Hall thing (as usually expressed) is ambiguous and depends on whether MH has prior knowledge, and tends to have a different answer on either side of the Atlantic.
Agree about the Monty Hall problem. I always need to know the rule about which door Monty opens. You only get the surprising resukt though, if his rule is "never open the door with a car behind it"
Which MUST have been the rule surely? Otherwise what's the point?
The Monty Hall problem is one of my favourite things. It is a perfect example of how common sense can mislead you in areas like probability.
Which was nicely explained in Mark Haddon`s excellent book: "The Curious Incident of the Dog in the Night-Time"
Both sides need to be seen to have pushed as far as they can.
Indeed. If, fish were the sole (sorry) impediment and everything else were agreed, I suspect the Commission (cough Germany) and to an extent the U.K. would get cheque books out and buy out French boats lock stock and barrel. The salty tars of Boulogne could then bugger off to the Riviera with the loot leaving the fish to the Brits.
Macron has to be seen making a song and dance about this, but in the end I couldn’t see the likes of Germany, let alone Italy, Poland etc (who have no axe to grind on fish whatsoever) letting France bugger it all up, if fish was the last thing to deal with.
Blimey. That’s a map. Almost exactly traces the linguistic boundary.
By my soul, 'tis a Protestant virus.
Protestants are only abut 3% of the Belgian population. I know this because I made exactly the same point "in another place" earlier this week and was corrected.
Germanic language-family speakers are less "touchy-feely" culturally than Romance language-family speakers?
Wallonia is the French AND German (and Walloon) speaking region.
Nippy does, to the extent I tend to discount what people have to say when they use that word; you probably aren't going to learn anything from someone who calls her Nippy.
Both sides need to be seen to have pushed as far as they can.
Indeed. If, fish were the sole (sorry) impediment and everything else were agreed, I suspect the Commission (cough Germany) and to an extent the U.K. would get cheque books out and buy out French boats lock stock and barrel. The salty tars of Boulogne could then bugger off to the Riviera with the loot leaving the fish to the Brits.
Macron has to be seen making a song and dance about this, but in the end I couldn’t see the likes of Germany, let alone Italy, Poland etc (who have no axe to grind on fish whatsoever) letting France bugger it all up, if fish was the last thing to deal with.
Extreme pedant alert! You couldn't grind an axe on fish. They're far too squishy.
This is very hard to watch. Not very illuminating. Jumbled sentences from both of them. I don't think it will change many votes. I don't think Trump will actually lose votes this time, but nor will he gain them. It is not the game changer that Trump needs.
If you believe Trafalgar or Rasmussen and IBID/TIPP he does not need a game changer, just not to lose any support tonight
If you believe the poll modellers like The Economist or Nate Silver, he most certainly does need a game changer.
If Nate Silver was right in 2016 Hillary Clinton would have won over 300 EC votes and now be President
How many times.....? Nate Silver's group was 28% right, which was more pro Trump than any of the other groups giving probability predictions for the 2016 election.
Or to put it another way, Biden is in a much better position in the polling than Clinton was 4 years ago and has been for months. If HYUFD really believes Trump should be favorite it's not based on any evidence.
Experience. Few predicted 2016 result.
Experience should teach you not to fall into the normalcy bias mistake. It happened with me in 2019 when I assumed Corbyn would repeat 2015.
The circumstances this time around are vastly, viscerally, different from 2016. The polls don't lie. They didn't in 2019 and they're not in 2020. But it's much more than that. The GOP have distanced themselves from their President. They know they've lost.
And lost big.
I'm going to use 538 and compare Trump's popular vote forecast lead with that of the Republican candidate in the closest Senate races. IA Trump +0.3 GOP -1.0 GA Trump 0.0 GOP +2.6 NC Trump -2.3 GOP -2.9 MT Trump +10.2 GOP +3.0 ME Trump -12.1 GOP -3.0 KS Trump +10.6 GOP +4.9 SC Trump +7.9 GOP +5.2 AZ Trump -2.9 GOP -6.0 AL Trump +19.2 GOP +6.2 AK Trump +7.4 GOP +6.3 MI Trump -7.9 GOP -6.4 CO Trump -11.7 GOP -7.7 TX Trump +2.1 GOP +8.1 MS Trump +13.2 GOP +9.1 MN Trump -7.7 GOP -12.0 KY Trump +19.1 GOP +13.4 ---- Mean Trump +2.8 GOP +1.2
I don't want this to be true, but Trump still looks like he is an electoral asset for the Republicans. I want American voters to judge him so badly that he drags the whole GOP down with him. But the numbers don't show that story.
Otherwise Republican Senate candidates should be outperforming Trump.
Trump won the biggest electoral college victory for the Republicans since Bush Snr in 1988 and he won Wisconsin for the first time for a Republican presidential candidate since Reagan in 1984, in the rustbelt he is an electoral asset for the GOP compared to a generic Republican, in California he certainly is not (Hillary won California by an even bigger percentage than Obama did in 2008 or 2012 for example) but it is the former he needs for EC victory, he can afford to lose California heavily and the popular vote
There are only three Midwest states in the 16 that I looked at and Trump's advantage over the GOP was smaller in those (+1.0) than in the set as a whole (+1.6).
In Iowa Trump did better than the GOP candidate, in Michigan fractionally worse (though Trafalgar suggest the GOP will win both the Michigan presidential and Senate race) you provided no data for Wisconsin or Pennsylvania in the rustbelt both of which Trump won for the first time for a Republican since the 1980s and Trump won Ohio by more than even George W Bush did too
HYUFD - please don't reply to this post with Trafalgar did best last time because that is not going to be the question.
Do you believe Trafalgar is a pollster or pundit? I genuinely don't know, but @rcs1000 and @Alistair have both posted evidence that Trafalgar is not a pollster but a pundit. They may well be a very good pundit.
If Trafalgar is a pundit it should not be posting spoof polls. If it is a pollster how do you respond to @rcs1000 and @Alistair?
If Trump wins the EC again Trafalgar will definitely be a pollster and a goldstandard one at that, if Trump loses the EC it will just be a pundit, simple really!
Why would Trump winning or losing make them a pollster or not?
They are either engaging in legitimate polling or they are not.
If they get the correct result obviously a legitimate pollster, the whole point of polls is to forecast, correctly, election results and the views of the public otherwise they are worthless
There is a bet on the throw of an unbiased six-sided die. If it doesn't come with a six, you get evens. A good value bet - yes? You take it.
Someone in the crowd, - let's call them Trafalgar - says "It's going to be a six!" The die is thrown. It comes up six. You lose.
Four years later, a similar bet is offered. Evens that it won't be a six. Analysis shows that that the chances that is won't be a six are 5/6 so evens is still a value bet.
But someone on a blog points out that Trafalgar got it right last time and is still saying it will be a six. Do you listen to them or rely on the analysis?
Barnesian nails it
I'd suggest the difference is that you don't know if the die is weighted. You couldn't work it out last time and you are not so sure this time.
So Evens vs 5/6 might not be so great
You are stretching massively - Barnesian made it clear that it is an unbiased six-sided die.
Plus last time Trafalgar shouted "it will be a six" repeatedly and were generally wrong. They hit a six in a state and then are perceived as a sage by ignoring all the times it wasn't a six elsewhere.
Reminds me of a telephone tipster scam that I thought of years ago but never implemented. You get a bunch of people (say 80) to call a number for a racing tip. You pick , say, a 8 runner race where there ideally is no clear favourite. You split the 80 callers into 8 groups of 10 and give each group a different tip so that you have covered the field. You then start charging the 10 who were on the winner for future tips cus they think you are a genius.
That’s similar to what Derren Brown did for a show “The System” a few years back. He gave thousands of people random racing tips, and a handful of them ended up with half a dozen winners in a row and thought he was a genius tipster.
Probably Brown's best show. It has the 10 heads in a row sequence as well.
His decision to make the final race to convince his marks that he was a genius a Novice Hurdles race was utterly brilliant. Of course there were fallers, so of course the person who got given the winning "tip" was even more thoroughly convinced.
Yes, one of his best works. I really loved the “10 Heads” sequence, amazing that no-one had got that on camera previously.
Both sides need to be seen to have pushed as far as they can.
Indeed. If, fish were the sole (sorry) impediment and everything else were agreed, I suspect the Commission (cough Germany) and to an extent the U.K. would get cheque books out and buy out French boats lock stock and barrel. The salty tars of Boulogne could then bugger off to the Riviera with the loot leaving the fish to the Brits.
Macron has to be seen making a song and dance about this, but in the end I couldn’t see the likes of Germany, let alone Italy, Poland etc (who have no axe to grind on fish whatsoever) letting France bugger it all up, if fish was the last thing to deal with.
Extreme pedant alert! You couldn't grind an axe on fish. They're far too squishy.
Quite so. My comment was off the scale in its inaccuracy.
James May fixed this for me with toy cars, they literally did the stick twist with dozens and dozens of cars and goats and the guy twisting ended up winning way more cars - sadly cant find the video. Still does my head in though
I can explain it to you...
I understand the logic, it just seems counter-intuitive. its easier if there are a million doors and he opens 999,997 of them excluding yours and then asks if you want to swap. suspect thats a very long show and a large investment in goats
I think it was Alistair who posted the African American support for any republican president peaked at 14% (Reagan).
Could Trump beat that?
Quite possibly, it would be ironic if the white vote swung slightly away from Trump compared to 2016 but he was re elected thanks to improving his shares of the Black and Hispanic vote
I understand the logic, it just seems counter-intuitive. its easier if there are a million doors and he opens 999,997 of them excluding yours and then asks if you want to swap. suspect thats a very long show and a large investment in goats
It was also a standing joke. When the problem was discussed on this site some years ago, I failed to grasp it for a loooooooong time.
So now cases are still out of control due to Government failure, we've now resorted to lying about the cases and making faulty graphs. Goodness me.
There getting rather out of control everywhere. France, the Netherlands, Eastern Europe, and my very local area has an incidence, I learnt this morning, of over 1000 (one thousand) per 100,000.
Given Drakeford’s in charge here, who’s Government is failing?
Blimey. That’s a map. Almost exactly traces the linguistic boundary.
By my soul, 'tis a Protestant virus.
Protestants are only abut 3% of the Belgian population. I know this because I made exactly the same point "in another place" earlier this week and was corrected.
Germanic language-family speakers are less "touchy-feely" culturally than Romance language-family speakers?
Wallonia is the French AND German (and Walloon) speaking region.
Yes, and if you look closely at the eastern end of Walloonia, you can see the German-speaking bit along the border seems to have a smaller incidence, at least in its northern parts, than the French-speaking bits next to it.
Comments
I think Trump will carry Texas by about 3/4%, Iowa, Ohio and Georgia by 2% each.
I think Florida and North Carolina will be the closest states and Biden will only edge them out by 1% but I would not be surprised for Trump to win either.
Arizona, I think Biden will win by 2/3% and Michigan, Pennsylvania and Wisconsin by 5/6%.
He should carry NE2 and will probably take ME2.
https://www.standard.co.uk/news/politics/uk-and-eu-very-far-apart-fish-deal-a4572744.html
Therefore, there exists at least 1 more tier...
https://www.google.com/amp/s/www.irishtimes.com/news/ireland/irish-news/north-covid-row-after-dup-minister-claims-virus-rates-higher-in-nationalist-than-unionist-areas-1.4384116?mode=amp
Jesus Christ.
-2 = mandatory kissing
-1 = mandatory pubgoing
...
5 = full lockdown, restrictions on going out of doors
6 = your apartment will be sealed in plastic
7 = purge everything with fire
...but I won't pretend that's what Sturgeon just said
That's what she said
My conclusion is that this means they aren't going to go there.
While it would be interesting to see for a few of us from PB, I totally understand that they want to keep their model assesments on historic data in house. There is too much scope for others to nick new ideas, and for dumb journalists to run dumb stories like "538 now predicing that Trump will win the 2016 election".
It gives further advantages to British agriculture and services than the EU version:https://www.politico.eu/article/five-things-from-the-uks-trade-deal-with-japan/
There's the ONS chart.
Flanders is richer, and so there is less deprivation. The same trend as seen everywhere else.
Of those 5
Causes Covid - jury out, probably not but I'd like to see some RCTs.
Joking here, but the Yay for Ther Science lot are usually primed to squawk ANIKDOTE at any theory not derived from an RCT, so why not here?
Used for spying - unquestionably the case, one way or another.
Trees being cut down for it - seems inherently likely, got to put masts somewhere, no evidence to the contrary presented.
5G is making the human population stupid - it is. It makes them spend more time looking at cat videos on facebook, and having silly arguments about the effects of 5G.
11% say they believe that 5G radiation lowers human immune defences with Bristol residents buying into this theory the most (18%). A further 10% think 5G causes cancer, brain tumours or infertility
This could easily be true, and is taken very seriously by real live mainstream scientists.
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6701402/
WTF is it doing in this list?
There is no better example of bad science than an illiterate twit like this setting out to expose Bad Science.
Overall estimate is £1.5billion improvement over no deal at all, conveniently forgetting the improvement the recent EU-Japan deal was bringing us since Feb 2020.
Of course, flip through a dictionary and you'll find Latin words aplenty, perhaps many times more than Germanic ones. But those are the leaves on the tree, whereas the trunk and roots are emphatically Germanic.
Do you think she did 5 tiers rather than 3 to differentiate Scotland from England?
Biden +2 in Florida BUT
Massively good internals for him
Big lead in already voted and has hoovered up almost all the 3rd party vote within those that have voted compared to their 2016 vote.
They key bit is their HS and less education component has gone from approx 12—16% to 30-38% in their polls.
Has a big effect on the results if you reweight their 2016 to 2020 and vis-versa.
https://twitter.com/Rasmussen_Poll/status/1319605958037950470?s=20
Macron has to be seen making a song and dance about this, but in the end I couldn’t see the likes of Germany, let alone Italy, Poland etc (who have no axe to grind on fish whatsoever) letting France bugger it all up, if fish was the last thing to deal with.
You couldn't grind an axe on fish. They're far too squishy.
Could Trump beat that?
https://www.youtube.com/watch?v=9R5OWh7luL4
https://www.politico.com/news/2020/10/23/early-voting-numbers-swing-states-431363
Some interesting figures in this article.
Given Drakeford’s in charge here, who’s Government is failing?
In fairness it’s a shit show everywhere.
The difference in tactility between those two ethnic groups is huge.