Russell Bradshaw @Forensic_Stats

That all said, some of the ad-hoc adjustments in MrPs more than cancel out these effects. And in fairness to Survation, they don't herd and are very often rewarded for that. It is worth adding that MrPs aren't just about the mechanics of the model, rather they are more about how much it can predict individual behaviour and how much data we have on each constituency. More political science than data science. What variables are they using to predict things should get far more scrutiny. Does the model just rely on demography and past vote or does it leverage other variables that can add explanatory power. For example, in 2017 MrPs were better than in 2019 because voting behaviour in the 2016 referendum was an excellent predictor of how people switched between parties between 2015 and 2017 and that particular variable was well known for each constituency and relatively easy to build into the model. By contrast, the change in vote between 2017 and 2019 was not as easily predicted by the 2016 referendum and there wasn't any other variable that could just as easily plug the gap. Important advancements here are less about the mechanical, say the application of random forests, great as they are but rather in having really good quality inputs that predict individuals and building synthetic joint distributions of those variables, building high quality seat clusters and it has to be said a lot of MrPs in the UK simply aren't big enough in terms of their sample size if you want to really use the multilevel component. (7)

Horizon Inquiry. Code for how to reverse the sign of a number "Whoever wrote this code clearly has no understanding of elementary mathematics or the most basic rules of programming"

One of the most ominous risks for Europe is that of a major change in Atlantic ocean currents. Recent science suggests it has been greatly underestimated in the past -including by me, having worked on it for over 30 years. Here my half-hour presentation in Vilnius a week ago!

Interestingly, Portugal has seen a rebound of covid since a few weeks Why do I find this interesting? Because it shows that covid (at least still) has the capacity to come through in summer conditions 1/2

🧵The Covid in Schools Scandal While much of the media claims the inquiry is accomplishing nothing, its slowly revealed the gov knew transmission occurs in schools and causes harm to a not insignificant number of children

@Jean__Fisch It is remarkable that much of the heated debate was split by left / right political persuasion (i.e collectivism / individualism), and manifested in entrenched Bayesian priors.

Oh dear. A better answer might be: "the government borrows in order to neutralise the impacts of its spending (money creation) activity on prices and to give the private sector the opportunity to hold government liabilities (money) at higher rates of interest, supporting saving."

Arnaud Bertrand @RnaudBertrand

2 years ago

This is absolutely priceless. And probably the most frightening clip you'll ever watch on the people in charge of the US economy. Jared Bernstein is literally the Chair of the Council of Economic Advisers, the main agency advising Biden on economic policy

4K 15K 37K 15.4M 17K

So by seasonal we really mean three waves a year, not really settling into a predictable pattern then. My back of envelope estimate of attack rate for the November-March wave is ~15% based on ONS data.

alan PW💙🏴‍☠️ 🇺🇦 @penheiro53

2 years ago

@chrischirp Is this info from this UKHSA poster and is the graph telling us something

keyboard_arrow_left Previous keyboard_arrow_right Next

1 4 20 2K 1

@chrischirp Is this info from this UKHSA poster and is the graph telling us something

keyboard_arrow_left Previous keyboard_arrow_right Next

Short thread on the Atlantic Meridional Overturning Circulation #AMOC, which brings a huge amount of heat to the northern Atlantic. If it slows, the region west of Britain cools. And the Gulf Stream moves closer to the US coast, causing warming there. Both are already happening, as satellite data show. 1/6

On Newsnight I was asked what I thought the biggest mistake Johnson's government made during the acute phase of the pandemic. I only got to give one answer, but if I'd had time I would have said: 1. Johnson missed 5 COBRA meetings. Didn’t chair his first one until early March. 🧵

An excellent article from George, which echoes what I have been saying for decades. In this thread, I want to explain the major way, the neoliberal doctrine, has totally corrupted our society, and created a demographic apart from the rest of society, quite deliberately. 1/🧵

@

57 years ago

0 0 0 0 0

The East Asian states acted fast. Within a fortnight they mobilised biotech companies and experts to scale a test. Through shoe leather epidemiology, not mathematical modelling, they mobilised thousands of junior doctors.... (14)

Q. Yes. Well, I mean, you were involved in these discussions. Let me just put this to you: was this decision to reduce the protection for healthcare workers because there simply weren't enough FFP3 respirators? A. Not by me' 'NOT BY ME' - what a fascinating choice of words

keyboard_arrow_left Previous keyboard_arrow_right Next

@orbit_silver Please could you define “Covid centrist” using explicit policy suggestions? There is a very wide gap between GBD and Zero Covid.

@TorstenBell Not a pretty sight. But whisper it quietly, there is also a sizeable fall in household debt to income ratios in 2023. Quite how that will affect the middle majority as we go into 2024 we don’t quite know yet. But it sure won’t feel like 2010-2011.

🚨9th Jan 2020 'Covid, like SARS, was a virus with HIGHER AIRBORNE TRANSMISSIBILITY higher airborne transmissibility...does mean that it could be transmitted via the AIRBORNE route, which includes large droplets & also...fine particles...referred to as AEROSOLS' Who knew? JVT

The single most important chart abt today's Autumn Statement didn't appear in any of the documentation from @hmtreasury or @OBR_UK. But I've put it together myself. It shows household disposable income growth: how much better or worse-off we're getting. And two things leap out...

Just going back through Sir Chris Whitty's testimony at #covidinquiryuk today. It's striking just how critical was of the Great Barrington Declaration. It was "flawed at multiple levels". The idea was sensible "struck me as zero, actually". Then this when pressed by Hallett.

Whitty's excuse about upscaling testing is wrong. He should have set up an advisory group to get this moving from day one. The excuses about lack of infrastructure compared with S Korea is a retrospective excuse and misleading. (1) THREAD