Does Nate Trust His Model? Debates, Prediction Markets and Poker

September eleventh. So last night you and I both were listening to the debate from our respective different parts of the country, and Kamala Harris tried to make her appeal to both of us, right to all Americans all over the country. So what'd you think, Nate.

segments with no mics off, no interruption, and would spend the first minute complaining about whatever Kamala said, and then the next thirty seconds with a coherent point, and then veering off in some other random direction for another thirty seconds. A lot of talk about things like a reference to if you're not really into politics, this will sound bizarre because it is, but Haitian immigrants, oh.

zero evidence of this. But he just didn't know where to go. I think, you know, I was worried, you know, going into this debate. I thought it could go a lot of different ways, and I was just hoping that she wouldn't let him get to her right because as a female also, the risk is so much bigger because if she starts feeling defensive or indignant or anything, you know, because she's female, it's going to come off a lot

worse and it's going to just completely blow up. She did not, and it was actually the other way around where she was able to kind of just needle him, and he every single time squandered most of his airtime in responding to the needles as opposed to saying anything substantive, which I think was the entire point.

shortening of donkey. So this was thought to be a play that a bad player makes it doesn't know to check to the person who made the last bet. It turns out that and the fancy computer simulations you're supposd to this sometimes but can also be a way to exploit somebody, to get under their skin. And it can I'm not gonna talk about the game theory here, but it can throw them off rhythm. Sometimes it is correct actually under game theory, but it's a way to take

charge of the hand. And she did that a couple of times, like kind of premptively saying, Hey, Trump's going to lie a lot, you know, even on issues that would seem like relative weaknesses for her, like, for example, the unpopularity of President Biden in the fact that she's been the vice president for three and a half years. She would say, by the way, I'm not Joe Biden, and like actually lean into that. And I think she

was well prepared for this event. And I don't know to what extent that credit goes to her versus her campaign. The Democrats to have a lot of data. I mean, Donald Trump has had one, two, three, four or five. This is his seventh general election debate, and he doesn't really change form that much, except in the Biden debate, where Biden was so bad that some instinctual part of Trump knew that you have to maybe not go for

the jugular that much. Although Trump was actually kind of bad in that debate too, it's just that when Biden.

to be allowed to continue that aggression. And there have been a few times I even remember when I've made big mistakes and hands because someone has donked it to me, and I'm like, no, you fucker, you don't get to do that, and I raise, which is exactly what I shouldn't be doing in some specific scenarios where they actually, like clearly are donk betting because they have something. And I remember those hands vividly because it's very unlike me

to to start tilting in that particular way. I've gotten over it now, but it's actually, you know, it can be a very effective strategy. And she did that. There was actually one time where at first I was like, wait, what the hell did she just do? And then I realized what she was doing right when she was answering a question and all of a sudden she says, by the way, you know, I invite you to go to Donald Trump's rallies. And I'm like, wait, you know what a non sequort or what did she Why is she

doing that? And then it became very clear was a very good strategic move because she was kind of deflecting away from her record on something that she didn't really want to talk about, and instead then Trump kept, you know, for the next two minutes talking about how he has the biggest rallies in history.

became much stronger as a night war on. I was very nervous when she started giving her first response, because you could tell she was nervous, you know, as someone who has done public speaking, who's kind of watched a lot of debaters, you know, you and I were both debaters all through school. You could tell, like the first few minutes, you know, she was nervous and she was a little stiff. But then she kind of got over it, and I think she kind of became much more fluid

and really kind of picked up in style. And I was very happy to see that. And he had kind of some of the opposite trajectory where kind of he started off thinking that, you know, I'm gonna just completely demolish her, and then he started getting very defensive.

little bit self conscious about it. Right, You're like, I am on terrain that I either rehearsed or that I've spoken about many times, and I'm just gonna say words and lo and behold, they come across coherent most of the time. If you're tired or fatigued, or or you're in a room where the setting is weird, the lighting is, you can be thrown off and that can be hard

to recover from. But you have to like you can't overthink it too much, and you have to trust your experience base that like, yeah, this is actually gonna go fine, and some people are even better in that setting than they would be a meeting or something.

The other thing that I thought was very interesting was, you know, Trump won the coin toss and decided that he was going to give the closing, you know, the final closing, and I actually think that that worked against him, because he spent most of his closing remarks yelling at what she said, as opposed to kind of giving any summary of anything that he had said. So so I think that she came across much better at the end of

the debate. And the way that you know, impressions and memory formation works, I think that that really works in her favor.

Jadie Vance is one. Not preparing for the transition from Biden to Harris when they had been telegraphed at least for a few weeks as a possibility, a strong possibility. His convention speech where he rambled off topic after a more subdued thirty minutes on the assassination attempt, you know,

frankly accepting that first debate with Biden. I mean, I guess you may not have played out all the tactics, but if he hadn't done that, then Biden's on stage tonight having that debate from June tonight, and Democrats are appropriately panicking because they're probably totally fucked at that point. So thanks to the White House for moving that forward. If you're a Democrat, I guess yeah, for sure. By the way, did you notice that he did not call her by her name a single time during the night.

He kept saying she and pointing at her, and he never said vice President Harris. He never said Kamala Harris. He used her name exactly zero times. I've actually haven't seen anyone mention that, But that's something that really struck me and kind of shows how, you know, it shows you so much about his attitude about kind of his conception of women, his conception of her. You know, she's

not even worthy of calling by the name. Because I was I was actually paying attention to it because I was wondering if he was going to mispronounce her name on purpose the way he's always done, and what her reaction was going to be.

I'm worried about climate change, but believe me, I'm not going to hurt your jobs. So I think that she tried her best to appeal to Pennsylvania, which is obviously crucial, and.

Biden's mistakes. You can do a fair amount of that. I do think that's an issue for Harris with some voters. But look, they got another data point. So the convention, we can talk about the trajectory of the horse race. Maybe in a moment. I thought she gave a very good convention speech. Also that was also kind of agro and aimed at male voters. And she 's some of these same themes tonight, talking about competitiveness with China, talking

about how aggressive and lethal our military is. For instance, sometimes when you hear something for the second time, it closes a deal, right, you know, So obviously there's some chance that there's a bounce in the polls Harris, by the way, so we're not just totally giving you our subjective opinion. In the CNN post debate poll, she won sixty three thirty seven or something, and Market.

the you know, forty five percent range. So fairly close to fifty to fifty. The way the model works, the model was very disappointed. Shall we say, not that it has emotions. It's not some large language model that the model is disappointed with Harris's post convention polling, where typically after a convention is one of the best periods for

a candidate throughout the life of the election cycle. And Harris was not only not getting a sustained bounce, she may have gotten one for a couple of days and then RFK drops out the new cycle changes, but actually been declining. She had lost a point or two in polls of Michigan, Wisconsin, you know, Pennsylvania, most of those rust belt blue wall swing states. So it's like this should be a really good period and it's a really

bad period. Therefore, it had gotten really grumpy about her polling, and therefore a lot of Democrats, some of whom are completely insane and should go, I don't know, moved to Fiji or some nice country where they don't have to blow up my Twitter mentions. We're upset with me as a result, But we have to talk about that.

funding Nate, which is bizarre. I mean, Founder's Fund has invested in in Lift and Airbnb and Facebook and and you know all these all these startups eventually get invested in by most of the big VC firms at some stage of fundraising. I mean, if you work for open Ai. Elon Musk co founded open Ai, are you on Elon Musk's payroll? You know substack Mark Andreeson and recent Horror,

which is a big investor in substack. I think he isn't as triggering for people as Elon or Peter Teal, but like, but yeah, you know, you have to understand how business works and how vet your capital works. But people people cannot fathom because their brains are so poisoned by politicscept Like, you know, it's not like I want Trump. I mean, you know, I'm disclosed here. I'm trying to

be nonpartisan to our audience. You know, I'm gonna vote for Harris if if I'm being super duper honest, I mean, most of the audience for kind of high brow news content are Democratic leaning voters. The audience at the newsletter is very is very colle educated. You know, when when Harris is doing well in the polls, then we probably convert a little better from we get traffic either way, we convert a little better to paid subscriptions when there's good news for Harris.

a mistake poker players make. Actually they assume that everybody else plays the same way that they do, right that if they'd be bluffing in a particular spot, and maybe they have the wrong bluffing frequency, we call it that. Therefore the other player must be also and they're mapping themselves to the rest of the world.

time capturing a representative sample electorate. And also the news story itself. Part of the benefit of the debate is people watching debate night, but also then these clips circulate of Trump talking about immigrants eating dogs and cats, and the abortion stuff circulates, and you get buzz and favorable headlines for Harris. By the way, Taylor Swift, yeah, I.

tell Taylor, Hey, Taylor, just hold off. Let's save us for a couple of weeks from now, when if there is a debate bounce that it might fade in the new cycle changes. But anyway, so first you get these kind of either online polls that have these large panels, that have very large samples, which can be fine poles, but tend not to show a lot of movement for reasons that aren't getting worth into or really crappy one day IVR automated polls, So I wouldn't pay too mu

attention to anything until maybe this weekend. This weekend, the first traditional polls that survey people Wednesday, Thursday, Wednesday, Thursday, Friday will come out probably Saturday or Sunday. Those might show more movement. If there is movement, those are a bigger deal. And then in the following week you'll start to see probably more state polling data. So by the end of next week we'll have a pretty good idea

of where the race stands. Now, if I had to guess you look at that number in the CNN poll typically that does actually correlate with movement in the head to have polls. You know, one of the biggest numbers ever because gaps ever in the c and M pole was in the first debate in twenty twelve when ntt Romney, you know, kind of clean Barack Obama's clock and Rodney gained about temporarily it faded, but gained about three points

in polls as a result. After the first debate with Trump and and Biden, Trump only gained three points or so. Right as disastrous that the bit was for Biden. Now, maybe that Trump is like near his theoretical ceiling anyway, Right, it's hard for Trump to get much more than fifty percent of the electorate. He's a pretty unpopular guy. But you know, so Beasonville, the guest might be that Harris gets half that a point or a point and a half. The question is kind of what the baseline is exactly.

If you look at the Silver bulletin pulling average, she had been ahead by two point two points, I think or something. So that gets her to three and a half or so. On top of that, our model assumes that we're no longer in this convention period, so it's no longer penalizing Harris's polling. If she were to get up to three and a half point lead, depending on what happens in the swing stage, she would emerge out of that as frankly a slight favorite in the model. Now.

I don't want to guarantee that, though, because there's a little bit of ambiguity about where the race stood beforehand. In some of the very recent national polling, The New York Times has this very good poll to do with Santa College, very large sample size, very rigorous methodology. Trump had been ahead by a point in the popular vote, and Trump's a head by a point nationally. He wins electoral college like for sure. So there have been some

bad data for Harris recently. If you can get back though to the three and a half point range three four that you know, that would count as a as a big tournament events.

are going to be looking at it. But I think that you do need to have some of these things that you can sprinkle in along the way because it's a long road ahead, and as we've talked about before, like for Harris, any little slip up is going to really really matter. And she doesn't even have to mess up, she just has to, you know, do something that's perceived the wrong way. She's the one who has to prove, I think, and to be kind of on she has to be perfect, right, she has to be one hundred percent.

of that will stay with Jill Stein or whatever else. Right, but like, you know, maybe you have five percent of the actual electorate in play or four percent or something like that, right, you need to get probably half that. It's non trivial problem, actually, And like so really reaching into like the marginal constituencies, maybe not doing the mainstream media hits, but doing some of the weird podcasts and find,

you know, profiling who are the undecided voters. Maybe it's maybe it's men who just are a little uncomfortable with voting for a woman. Don't like Trump, but maybe a little uncomfortable voting for a woman or things like that. I don't know, you know, I think they tend to over index on like the Nikki Hayley, Liz Cheney voters.

They do fine with those. But maybe the more marginalized it might be all you know, it might be younger black voters and Hispanic voters who you know, do not grow up with memories of the Civil rights era and require a little bit more persuasion.

were to Biden. I mean, it's a tiny constituency, but like the crypto people are a group that, like Trump, has gone out after bitcoin declined after debate last night because they saw Trump as losing. For instance, areas like that, perhaps she.

some of which you are less good. I'm aware that prediction markets are generally pretty smart, and those also Harris, you know, Polly market her whatever, forty five is, forty four is percent range instead of thirty eight. That's like not a huge gap. But look, one reason to believe the model is that like is that, you know, if I were doing this subjectively, then I'd be rooting for Harris,

probably right. I am not a Trump fan, not a huge fan of Harris's politics, frankly either, but I consider after January sixth, there are lots of reasons why I would never vote for Trump, and she's the alternative, you know,

for various reasons. I think. Probably also, if you were somebody who was really agitating, as I was, for somebody, not necessarily Harris, frankly, but somebody to replace Biden, then you're gonna look really smart if you're one of the people who said you gotta get rid of Biden, and then that person wins and Harris is probably good for

site traffic and newsletter subscriptions. Frankly. So the fact that I have all these incentives to root for Harris is the reason to trust the process I set up ahead of time, before I knew how this would affect anything, and the same process yielded reliable model results since two thousand and eight over many elections, And so I'm not involving my emotions. So yeah, if I were, you know, look, as a Bayesian, you'd kind of combine your forecast with

other information, other data. So probably realisticly would be like forty two percent, taking the mid point of like polymarkets forty six and our models thirty eight, and you get to, like Harris forty two or forty three is probably my subjective belief that seems not to go out on too far of a limb.

Wester go ahead. So the mystery bounty is that if you knock out a player, usually after some threshold, if you get to day two of a tournament, then you knock out a player, you get like a little card that says you get to draw from this like big vat literally of envelopes, and some of the envelopes contain very large amounts of money, such as what was top prize here, like something one hundred thousand, one hundred thousand euro Yes, one hundred thousand euros. It's one hundred and

ten whatever thousand American dollars. So yeah, it's a way for a cheap entry to have. Like usually you can only make these six figure sums if you like win the tournament or finish in the top like two or three. This is a way to have another opportunity without having to win, without having to beat everybody, just knocking one player out and you win one hundred thousand euros or a million dollars and some of the worldsairs of poker ones. So did you win a big bounty then.

a min bounty. So a min bounty is a minimum bounty. In this particular tournament, that was one thousand euros, and this was by far the most common bounty. So we are going to now pivot to a very non risky business way of thinking. So here I am with ten bounties to draw. I draw the first one. It's a men. I wait a while to draw another one. It's another men. So at some point I've drawn five men bounties and I have five left. Now the one hundred k's are gone.

And there is a player who everyone in the poker world knows and anyone who was following the main event this year knows as well. Nicholas Sausted, otherwise known as Lena by his online screen name. He had a very deep run in the main event, and a lot of poker players, including me, were rooting for him to win because he's a phenomenal player and a really really nice guy. Lena is also someone who's quite lucky and runs incredibly well.

He plays well. But see this is why I said, We're about to pivot into a non risky business way of thinking. By this point, he had been knocked out of the bounty event, and he had drawn two bounties. He had drawn a twenty five K and a ten K. So what I decided to do was send a text message to Nicholas and ask him if he would draw my remaining bounties. This is an incredibly non scientific, irrational

way of thinking. I was like, I am going to get this very lucky player who's very good at drawing bounties. By the way, we know, right, you cannot be good or bad at drawing bounties. This is random. But I'm going to get him to draw my remaining bounties. So he said, sure, he would do it.

are the top bounties left? And I said, well, there's a twenty five K left and a ten K left. And he said, okay, we got this. So he draws and he draws, and you know, he draws my five bounties and he said, do you want to open them? Do I open them? I said, let's take turns. So I opened the first one. It's a twenty five K. He opens the second one it's a ten K, and

he's like, okay, I did it for you. He's like, the only other ones left are the one case right, And I was like, yep, so then the other three were one case.

And if we're ever playing a mystery bounty together, I hope that you will draw my envelopes for me again, because I'm a very rational person who does not aspire to be superstitious ever.