One thing to note, just for calibrating expectations for Nov, is that the 2016 error made the race go from a predicted ~300 EVs for Clinton on average to 232 in reality. Here, the same error puts Biden at 260. That gives you a rough idea of the implied average EC projection. pic.twitter.com/JwsEkC5UTX
Comments
Yes please! Hopefully Kerry will leave as well
I'd assume the two campaigns have private polling doing just this, but the US system is so crazy I don't know.
I was just gently pointing out there was a small problem in your argument.
If the polls are overstating his position then it depends on by how much:
By ECs
0% 333
1% 333
2% 318
3% 289
4% 289
5% 258 (Biden loses Ariz and Penn)
I reckon Biden will get 289. The Economist is on 334 i.e. that the polls are correct.
The way I look at it, I don't believe that any sacrifice in our own health should be required in order to respect the welfare of other creatures. There is a virtuous cycle that we're part of - it involves eating other animals, but also ensuring that during their lives they are well-nourished and healthy.
Most minerals and vitamins aren't digestible by humans in their raw form. Plants translate them into a more absorbable form. Animals eat the plants and translate them into an even more absorbable form. We eat the animals (and some plants obviously - but often the plants require specific preparation to make their nutrients more bioavailable). If the animal is undernourished, the meat is undernourished. So animal welfare flat out makes selfish sense. But avoiding meat completely does not.
I agree with you about the former, but the latter can’t come back in the way you seem to think without a huge reduction in population and a major economic shift.
I've been droning on for weeks about the margin of error and sample sizes in State polls but up pops some numpty in the New York Times and suddenly it's the most insightful analysis since a brave French Marshal opined to Napoleon at Waterloo that if the Prussians showed up in the evening it might get a little awkward.
If you are going to (and I suspect many do this to support their own personal agenda while I can understand the more astute who do it to bolster their trading position) simply re-tweet headline numbers from some dime-a-call US pollster, at least have the decency to find out the margin of error so we all know when we can stop laughing.
If you want almost any result from a Biden landslide to a Trump landslide, the polls will give it but as a rule of thumb I'd take any poll showing either candidate ahead by five points or less to be in the TCTC category.
I'd also take a long hard look at sampling - what is the split between registered Democrats, Republicans and Independents? What is the ethnic split, the split by income?
Too many pollsters (Rasmussen and Trafalgar in particular) either don't reveal the details of their sampling claiming it is "representative" or hide the details behind a paywall. What do they have to hide? Who are they sampling and in what way does that produce such different results from other pollsters?
If I sampled 1500 voters in East Ham and published it as a national opinion poll, I'd be thrown off the British Polling Council and rightly so but we don't know to what extent the methodology and sampling effectively weaponises these polls to create numbers favourable to one side or the other.
I still don't understand the face
But do not let that inconvenient fact interfere with our desire to paint the UK in a bad light
https://youtu.be/dxm2HZp8c3A
Because I state an opinion does not mean that I still think Boris is fit for purpose which he is not.
And I want this government to succeed on covid and brexit, remember I am a conservative
The Pillar 1 & 2 tests vs capacity graph has been there for months.
Testing has collapsed.
Starmer and the twatterverse say so so it must be true.
Marr reminded him the present need with all the children with seasonal colds is one million per day
So even with hindsight Starmer would be 500,000 daily tests behind the need
And of course we are testing more per day than anyone in Europe
The UK decided to protect the NHS.
If demand for testing (measured as actual Covid prevelance) is going up 10% a day, a 35% increase gets you 3 days extra demand before the system starts collapsing again.
That plateau in test capacity in high summer (Yay! We have enough tests!) might have been unwise.
Interesting, Trump also scores (slightly) better on the crime question. That may help him in some suburbs.
Re the others, obviously they all add up to over 100 but my guess is there is a very strong overlap between the Covid and Healthcare in terms of voters' interests so they are pretty much the same question.
Anyway time to go. All the Johnson enthusiasts are out this afternoon.
If they do confirm someone before the election then they can create a lot of rage among the base and concern about their healthcare among swing voters, and reasonably argue that they'll need a Dem presidency and Senate to salvage the situation.
Oh wait ...
If Trump and McConnell install a judge then it morphs into "they've be rushed on a judge to steal your healthcare, vote for me to fix that"
This is why the SC pick is actually a distraction for Trump. His entire focus should be the economy.
Now you could argue that there is insufficient capacity but you then need to be practical and discuss how you would increase that capacity.
And that is harder to achieve than it is to posture about.
Healthcare is hugely strong for them (as shown by 2018) and it avoids directly talking about court packing ornothet things that makes moderates nervous.
https://www.theatlantic.com/ideas/archive/2020/09/experiment-wisconsin-might-reveal-key-defeating-trump/616367/
It is actually possible to change minds on the economy.
Regarding the SC, the motivations of Republican senators and Trump are very different. And Trump has a better feel for what creates applause than he does for strategy.
Re Minnesota, I know the immediate pushback will be "look at the polls" but, if you look at both sides activities, it is clear they both think Minnesota will be close this time.
Michigan is perhaps the more interesting one. John James is running again for then Republicans. The latest poll has the Democrat incumbent Peters at +4. It is worth noting that last time the polls at this stage were showing James running 15%-20% behind Debbie Stabenow, who was arguably a stronger candidate but the final result was a 6,5% difference with James coming up strongly in the final weeks.
The obvious pushback is Robert's point that ticket-splitting is far less common nowadays. That is true. However, there is a feature in this race that puts a spanner in the race namely that James is Black and Peters is White, and there is an inherent feeling amongst many Black Americans (especially older) that you do not go against "your people" (as my in-laws say). So I can see a situation where many Black voters may vote for Biden but, at the least, abstain when it comes to voting for Peters.
https://twitter.com/wabbey/status/1307624322102681600
Sadiq Khan, Andy Burnham and Mark Drakeford haven't even considered getting the equivalent work done.
Now consider how much more effective Labour's message would be if they could point to something Labour had achieved at regional level.
But they can't because they haven't.
Rasmussen and Trafalgar have made an assumption that there ARE lots of "shy Trumpsters" and they adjust their raw data accordingly, using sleight of hand such as "who do you think your neighbours are voting for?" type baloney. Their objective is to stand out from the crowd and cross their fingers that their core assumption - Trump will overperform in key states like last time - is correct. In which case they make themselves a big rep.
The fact of the matter is that the comprehensive polling average (with them included) is a Biden lead of almost 7 points. The latest poll (today) is from NBC, an A rated pollster. Biden by 8. Same lead as a month ago with that pollster. Biden also has a solid lead in most of the battleground states and is within touching distance in several states which were clear wins for Trump in 2016 such as Texas, Ohio, and Georgia.
Are the polls wrong? Yes, obviously. But it's unlikely they are all wrong and biased against Trump by several points unless their methodologies really are missing a chunk of Trump voters. And to the extent they are wrong it is just as likely they are wrong the other way. That they are understating Biden's lead. Indeed given that pollsters tend to over-correct for their latest high profile screw up, it is more likely that the error, if there is one and it's material, is in this direction.
So, Biden leads by 7, his lead looks stable, his lead is at least as likely to be understated as overstated, there are very few undecideds, and the election is only 6 weeks away. He should be a 1/4 favourite and yet is only a shade of odds on. It's the betting opportunity of a lifetime and I have not hesitated in having the biggest spread bet of my life.
Don't overthink it people. It's great fun to do that, but in this case don't. Commonsense and the evidence coincide, and the consequential and obvious conclusion is the right one. America has had enough of Donald Trump as their president, one term was quite sufficient, and they will be voting him out decisively on 3/11.
https://twitter.com/KevinMKruse/status/1307679955933855744?s=19
If Biden is hoping for a polling error in his favour as per 2012 then he will need to hope for black turnout as high as Obama got then with the over 90% of the black vote Obama also got then.
It should also be noted that some pollsters are now including more non college educated whites in 2016 however the polling evidence is there has if anything been a slight swing to Biden amongst that group anyway relative to Hillary but a swing to Trump with richer voters in households earning over $100k a year, yet the latter comprise the same amount of the sample as 2016
However with all due respect the Testing regime is essential a national enterprise run by Dido. I know this because last week people from Bristol and Weston Super Mare were being sent for tests in Abercynon. This was particularly peculiar as Abercynon is within the County Borough of Rhonnda Cynon Taff, which is in lockdown.
As Starmer has picked up so many 2019 LDs and a few 2019 Tories that change has been missed by most
Getting things done is part of getting on in life.
Anyone can say that someone else is responsible while doing nothing.
But if you want to impress the world then show the world what you can do.
What have Sadiq Khan, Andy Burnham and Mark Drakeford shown the world they can do ?
... thinks back ...
Sadiq Khan shut down some underground stations thereby increasing congestion at the others.
Is there anything else ?
The problem is you don't need all the polls to be wrong, just some of them and you can get a monumental f*ck up and lose a huge amount, especially if you are spread betting.
So, in Florida, if Trafalgar are right, Trump will win by +2 but, if Monmouth is correct, it will be +5 to Biden. In PA, Rasmussen has a tie, NBC has +9 Biden. In MI, you can either take your +2 for Trump with Trafalgar or your +8 for Biden with EPIC.
And this isn't taking into account states where polling is poor and a Biden win SHOULD mean they are uncompetitive but where both campaigns' actions clearly suggest they see it will be tight. Minnesota and Nevada spring to mind.
I am telling you, if it had been one of those cheap Chinese ones from AliExpress, at that frequency of irony it would have been curtains for you!
Doing a bit of reading up on this as far as I can see the UK is examining almost every possible option for COVID-19 testing, the biggest problem is that on the whole most of the alternatives to what we are currently doing are not clinically good enough. I'm absolutely certain that if there was a quick fix we would be rolling it out, money certainly doesn't seem to be an issue.
* It's small scale only in terms of the tests per site, but across the country it should add tens of thousands of fast and accurate tests each day.
I've used Google and dug through Wikipedia to look at the national polls ahead of each US Presidential Election going back to 1976.
In aggregate national polls have been pretty accurate. Out of the 11 elections, 7 have seen an a polling miss of less than 1.5%. 3 of 1.5-3.0%. And then one - 2012 - where the miss was more than 3%.
(I'm defining miss as a simple average of eve of poll predictions for the lead between the candidates.)
Polling errors have been roughly equally distributed between the incumbent and the challenger, and between Democrat and Republican. There is, however, a tendency for polling errors to correct: in 8 out of 11 cases the polling error went in the opposite direction between elections.
I therefore think that we have to take the national polling average pretty seriously.
It's possible, of course, that it's four points wrong in President Trump's favour. But I think it would be wrong to regard that as more than a 20% chance. Of course, it's also possible that Trump's voter efficiency is better than in 2016, and that he sweeps the Midwest again, despite a worse relative national polling position.
More likely is that President Trump simply gains on Joe Biden before the election.
But. (And it's a big but.) If the national polls don't change, then President Trump is facing an uphill battle.
https://coronavirus.data.gov.uk/
I believe yesterday was 4322. Today is 4422
Germany 362
Look at Daily cases by date reported, Data tab, on this page,
https://coronavirus.data.gov.uk/cases
19-09-2020 4,422 390,358
18-09-2020 4,322 385,936
17-09-2020 3,395 381,614
16-09-2020 3,991 378,219
Astonishing that the Government ever thought that using figures like this were a basis for judgements on international travel.