Why Trump 'under-performs' in many US election polls | Nate Silver

do you think that Harris and Waltz can seal the deal and do you think that they will I I look um if I were betting at LAD Brooks or something where it's 5050 I I wouldn't I wouldn't make a bet because they would take a tax on the BET and it wouldn't be worth the tax I think it's about as close as you can get to 50/50 um and also Americans have to get used to the fact that it's going to be close probably on Election Day too you're going to have to get used to the fact that you're going to wake up on November 5th not knowing who's going to win and you might wake up on November 6 still not going who's knowing who's going to win given how slow we are at counting our Vates you don't have to be a a political geek an American political geek to know the name of Nate Silva he is a big name in American political punditry a big number cruncher an analyst of the of the polling doesn't do the website 538 anymore if you know that website but he does put out the the regular silver bulletin on his substack it it looks at the data as it comes in on the 2024 presidential race there's lots of it coming in always always absorbing in the in the background a few years back though Nate Silva wasn't in to politics he was a professional poker player now he's published a new book which looks into risk and those who take risks it's called on the edge and Nate Silva is here with me now welcome to you Nate good afternoon thank you for coming in tell me about The Guiding thought of your your new book then which is about risk and understanding risk and the people who take take risks a key to understanding the economy Politics the world generally yeah if you look at the people who are dominating the economy increasingly it's in Tech and finance and they have the skills that I call the rivian skill set river is a poker term um it means on the one hand they're very analytical they like numbers they like data but they're also very competitive they can be contrarian they're very individualistic um and just seeing the same pattern of a certain type of nerd um over and over again throughout my travels to this metaphorical place and I'm one of these people I guess myself right I feel very at home in the poker World more in the politics world but we have certain things in common right I we just wanted to explore that thought which is I know Central to your book I haven't read your book to be perfectly honest it's it's on my list Nate um but your your thought about the rivian the river folk and The Villages you divide people broadly into those categories just explain that to us yeah the village is more The Establishment right so in the United States it might be Harvard in the New York Times East Coast progressive elite kind of old money um here I mean the UK is probably mostly in that uh in that Village category it's a very literate population but it's a little bit more conservative in terms of of taking risk right um there's a little bit more paternalism in the UK last night I was thinking oh maybe I'll go buy a a beer right at the store and the stores are all closed by 11 right you're in the backseat of your Uber and the alarm dings if your seat belts not buckled up right um it's a different environment in the US where people are very very very individualistic especially on the west coast where people move off from all around the world to Silicon Valley to found companies that might change the world or might wind up making them bankrupt all right let's let's before we get more into your thinking now um as expressed in your book and generally in your in your sub stack to you were a professional poker player yeah were you good at it were you making a lot of money what matters in poker is are you good relative to the other people you're playing at that time there was a poker bubble I would call it so it was good relative to my competition in absolute in an absolute sense I'm I'm pretty good I'm not Elite world class but I I played poker pretty seriously okay and does that mean you got got rich doing it did you make a stack of money I made a little bit of money to Parlay into other things where they made more money sounds like it's it's a hard way the saying is it's a hard way to make an easy living um there's a lot of luck in poker in in the short run all right so so you also done pretty well to be able to invest the money that you were making as a pretty well as a as a 20 something to be able you know play a video game basically uh online poker and make money was was liberating for sure right I know you operating on the online poker poker circuit now look and then you took a turn into Political analysis and analysing analyzing poll data and so on why did that happen um partly because the American Congress passed a law that banned internet poker so I became more interested in politics partly was living in Chicago I'd gone to University of Chicago and Barack Obama was this candidate who was very inspirational and partly because I thought media coverage of politics was lacking it was very based on feelings and Vibes and arbitrary things and and this is an era where the movie Moneyball comes out in the United States it's all about datadriven analysis and so it was kind of applying that that concept of politics okay so applying hard analysis but overlaying that on a sort of instinctual understanding of of people I you you you you believe that polling analysis has to be hard edged and based on good mass and understanding but it also has to be tempered by an understanding of human nature I mean it maybe comes more in poker than in polling I think um but yeah look when you're analyzing political Trends you have to like we exist in the world there's context in the world we have to understand where that's coming from right the fact that in the US now things are so incredibly polarized and partisan that 92% of people have made up their minds already who they're going to vote for um you don't have the big swings like you have here like that's a relevant context when you're building a model for example so tell me then Nate why were you and just about every other poster why were you blindsided in 2016 by Donald Trump winning if you had read uh my my blog then 538 you would not have been surprised that Trump won we gave him a 30% chance which is much higher than the market did or the punters did or other people did and so so I think that's kind of a myth I think that was a case where people were ignoring polling showing tight races in States like Michigan and Pennsylvania for example and apply too much of their their group think their Vibes based analysis and then blame the poers for it later if you actually go back or brexit's even better for example there were lots and lots and lots of polls showing leave winning it was a photo finish 50/50 5149 one way the other um but people in London thought no we can imagine this happening and so it's called unthinkability bias right is it still true then that the the the opinion po tend to underprice Donald Trump is that right so they have two times out of two the problem with two times out of two is two is not a large sample size um look in the US people are increasingly outspoken about their support for Donald Trump you know people on what I call the river like Elon Musk now are openly Pro Trump for example they're wearing hats around and having yard signs and things like that um look I think it's a risk the good thing is that you know I'm not a pollster but I I know people in the industry they're very aware that they messed up in 2016 and 2020 um and usually people self-correct or sometimes overcorrect when they make a mistake you've seen in England for a long time Tories were underestimated and people have tried to correct for that with new methods um look the old school method of just dialing somebody in the phone look at random doesn't work very well anymore most people don't pick up their phone call if a stranger calls but they're trying new techniques more modeling more data massaging um it was pretty good the American midterm in 2022 but we'll see about 2024 right and maybe an area where where pure maths intersect with understanding of human nature and character is there such a thing as a shy Trumper people who actually do support him but wouldn't say so and then go down to the polling station fill in the the ballet paper for Trump even though you wouldn't know that they were Trump is that significant how do they exist the Trump voters I know are not very shy maybe the opposite in fact but in the margin I wonder the issue is more that it's hard to get certain types of Voters on the phone people who don't trust the establishment um are more likely to vote Republican and less likely to be pick up a a polster's phone call um so you can do things you can ask people who' you vote for last time to balance your sample that way or which party you registered with to balance your sample that way um but the fact is that you're not equally likely to reach everybody and in the US we're very polarized along social trust and educational lines where if you're a high establishment trust college educated person you're a Democrat if you're low social trust maybe did not complete College then you're probably a republican um that's been difficult for pollsters to deal with right take us into the analysis in your in your new book then I mean your your analogies about the the river dwellers and the and the uh and The Villages and so on do you still you believe do you still have an edge over the rest of the pollers out there um I think edges are harder to find than ever you know For Better or Worse the capitalist economy is very efficient at at giving rewards to people who find edges I think increasing the edges might be in knowing when a model is wrong for example um but look we've put lots and lots of effort into our product over the years actually an election model is a tough thing to do it's a mathematically pretty challenging problem so you might have new entrance in the market who don't have the track record that we do um but for sure I any Gambler knows that any Edge doesn't last for long competition catches up yeah yeah but I'm inviting you really to Bang Your Own Drum here do do you believe as things stand you've maybe got to march on some of the other pollsters making their projections now I think we're the only pollster that has a long track term track record doing this and so I'd say like and we're very transparent one of the things about having this newsletter is I'm going to walk you through every detail of the model I think if you can't explain a model I worry that you don't know what you're doing there are more bad models out there than good models right this is a i t to ask I know you're poll you're a poll analyst but I ask this question over and over again even though it's completely unfair do you think that Harris and Waltz can seal the deal and do you think that they will I I look um if I were betting at LAD Brooks or something where it's 50/50 I I wouldn't I wouldn't make a bet because they would take a tax on the BET and it wouldn't be worth the tax I think it's about as close as you can get to 50/50 um and also Americans have to get used to the fact that it's going to be close probably on Election Day too you're going to have to get used to the fact that you're going to wake up on November 5th not knowing who's going to win and you might wake up on November 6th still not going who's knowing who's going to win given how slow we are at counting our boates it is it is it is tight no so what great pleasure to have you and thank you for sharing your ins absolutely thank you and your book is out now on the edge the thoughts and the writings of a well a renowned analyst of the numbers coming up with just a minute more here on times radio I'll get you up to date with the news in the House of Commons that arms ban on Israel on dab Digital radio on the times radio app and on your smart speaker news from every angle times radio