2026 Midterm Election Odds & Trades to Make | Predicting Politics - Wed Oct 21
STACKED extracted the fantasy-relevant player signal from this podcast and kept the full transcript here for context.
43 min
Duration
Estimated from transcript
Oct 21, 2026
Published
Wed, October 21, 2026
0
Player insights
0
Players mentioned
Player analysis
Player insights
No player insights have been extracted for this episode yet.
Full transcript
Episode Transcript
0:00Lisa, tell them about our new show! We have a new podcast, it's called Don't Lie to Me.
0:06I'm Lexi Iwanu, we're two best friends, 22 years apart, but so close at heart.
0:11And I'm Lisa DiGiulio, the old one, which means we have wildly different references, occasionally different opinions, and at this point in my 50s, I've lived it all.
0:21Now in my 30s, being the founder of New Shoes, Good Taste is my business.
0:26Listen on the iHeartRadio app, Apple Podcasts, or wherever you get your podcasts.
0:30I'm Lacey Mosley, author, actor, stand-up comedian, creator, and host of Scam Goddess Podcast.
0:35The true crime comedy podcast, all about robbery, fraud, and those who practice it.
0:40Each week I bring on an amazing guest. Ooh, I scammed them.
0:44What? I could talk to you for an hour.
0:47You're a generational yapper, like me. That's why I had to make yapping a job.
0:52And now you can catch full episodes of my show on YouTube.
0:56Start listening to Scam Goddess on the iHeartRadio app, Apple Podcasts, or wherever you get your podcasts.
1:02Let's go! Every case has a story. Every victim has a voice.
1:06But will it be heard? I'm Nancy Grace. On Crime Stories, we go beyond headlines to examine evidence, challenge the story we're
1:14fed, and pursue justice for victims and their families.
1:17Please join us at Crime Stories with Nancy Grace on the iHeartRadio app, Apple Podcasts, or wherever you find your podcasts.
1:24Hope you join us. I'll see you there, friend.
1:27Hey, everybody. It's me, Debbie Quimau Bell. This season on my podcast, Who's With Me?
1:32I speak with Bruce Lee's daughter, Brandon Lee's sister, and my friend, Shannon Lee.
1:37Back to your initial point of stewarding my father's legacy and giving my energy to that.
1:42You're lending your life to stewarding somebody else's life.
1:45Listen to Who's With Me with Debbie Quimau Bell on the iHeartRadio app, Apple Podcasts, or wherever you get your podcasts.
1:52The Action Network Podcast. No one, no one is offering you action like this.
1:57It's the gambling experience of lifetime. It's my way of saying, I understand, man, Mike.
2:02I know what you want. I know what you need.
2:06This is where you belong. Welcome into Predicting Politics presented by DraftKings.
2:10I'm your host, Mike Calabrese. Each episode, I'll be joined by experts who can demystify the trading experience on DraftKings Predict for the midterm elections.
2:18We'll talk recent news, key events, and the latest polling that are shaking up markets.
2:24And we'll also give it all to you in the hopes of providing the best intel possible for trading on prediction markets.
2:31Today, we're joined by Harry Crane, a professor of statistics at Rutgers, and a member of the CFTC Innovation Advisory Committee.
2:38Harry, welcome to the show. Thanks. Good to be here.
2:42So you're a major figure in this industry. I mean, look no further than your website, harrycrane.com backslash elections.
2:48Your model correctly predicted every state in the 2024 election, for example.
2:52But let's back up for a second. Every superhero has his or her origin story.
2:57How did you get into this business? Okay. Okay.
3:00I appreciate the over-the-top introduction. Thank you. I don't get that much.
3:04But really, the prediction markets and the political prediction markets, actually, I got into that from kind of two different angles.
3:12So, I mean, I've been involved in betting and gambling for a long time.
3:16That's really my main interest or, you know, my interest going back 20, 30 years since I was a young kid.
3:23So, I started playing poker when poker was big and then moved out of poker and into sports betting, moved away from all kinds of betting.
3:32And that's when I went to school for statistics.
3:35But then I came back to this around the 2016 election.
3:39And there I was actually studying political estimation or political forecasting.
3:43I wasn't, at that time, interested in betting on politics.
3:46And I came across the prediction market and I started to investigate that because I was specifically interested in a lot of
3:54the conversation in that election around the polling and, you know, Trump-Hillary.
3:58There were beliefs that there was polling errors. And I was really just looking, is there an – if there was a
4:05polling error, how would you know about it? Or are there other ways to get insight into what's going on?
4:12That's when I got into prediction markets. And now we're talking 10 years later, the worlds have kind of collided and combined
4:20because we have now sports betting on prediction markets.
4:23And, you know, everything has kind of come together and it's kind of blown up, so to speak, in a good way, I believe.
4:31So, that's a short version of it. So, both from a practical side as a user and a participant, but also it
4:38was coming from the academic and the statistical side of just trying to use these markets actually for prediction.
4:45Now, today's a great day to talk to you because you just released a research article on the bias in political prediction
4:52markets, which is a hotly debated topic. To you, what's the difference that you've found in the average prediction market trader and the highest volume bettors?
5:01Because they're not apples to apples. And we see this really, you know, across any market, whether you're playing cards, whether you're predicting sports or politics or entertainment.
5:11There's those that pick things that they kind of want to see happen.
5:15They want to see higher scoring games. They want to see their favorite teams win or their favorite candidates win.
5:22And then there's those, you know, sharps or whales or high volume players that are not necessarily interested in outcomes.
5:29They're just following the numbers. Right. And I think that for people coming from the sports world, this is kind of an
5:36intuitive observation, which is that there's the market has different people in it.
5:41There are mostly recreational bettors, I would call them, just gamblers who are betting on who they think is going to win
5:48or without really having a good way of assessing that or who they want to win.
5:54Now, that's kind of well documented in sports betting, the efficiency and the accuracy of sports betting markets.
6:00You know, it is hard to beat sports betting markets in general, although people do do it.
6:06But in politics, I think that there's a tendency to there's kind of a common sense belief that those results would carry over.
6:14But there really hadn't been any detailed study of it.
6:17And I guess the other thing that I think is different, at least for me, when I think about politics, political betting versus sports.
6:25And obviously, there are diehard sports fans who would never think to bet against their own team.
6:31But I think there's a lot less of those people than there are in politics where everything gets so magnified.
6:38And then, you know, to some degree, as a voter or as somebody who lives in the country where there's an election, there's just emotion.
6:46There's it's just kind of hard not to be emotional about or at least for probably a large percentage of the population.
6:54And so the natural question is, are people biased when they bet on political stuff?
6:59Well, yes, they are. By and large, in the in the study we just released, we looked at granular data.
7:06We had trade level data. We tried to match that up with the the the traders.
7:11Personal politics are people betting on, you know, in alignment with their politics.
7:16And it's not you know, it's there's two different groups you could think of two extremes.
7:21There's people who placed one bet on the election.
7:24And we're talking in 2024 now. Those people predominantly did they well, they were the most politically biased.
7:30If they were Republicans, they bet on Republicans. If they were Democrats, they bet on Democrats.
7:36And 73 percent of those bets were placed on Harris.
7:39And then, you know, as a whole, 28 percent, the volume came from areas that were 28 percent voted for Trump.
7:46So it was really not coming from Trump districts.
7:49I mean, you know, the places that are going to be more active are the metropolitan areas and suburbs and things like that.
7:57But Trump was favored to win and more so than the forecasts and the polls suggested.
8:03And so the real question that this was trying to get at is whether you remember or not around the 2024 election,
8:10there was this discrepancy between the polls and the forecast and the forecast and the markets.
8:16Prediction markets weren't really popular until that time. And there was there were there were a lot of people saying you can't trust these markets.
8:24They're manipulated. It's a bunch of Trump people. It's a bunch of young men who are skewing towards Trump or Republican or whatever.
8:32And therefore, they're biased. And so it first of all, as we show, those two things can be true at the same
8:39time where you have a lot of people in a market who actually are very biased, but the market itself is still
8:47good because you have the whales and the professional syndicates who are making the money by correcting the price.
8:53And then and so it addressed that question directly.
8:57And so that was the goal of it. And I think that there's still a lot more that can be done about it.
9:04But it's really the first, I think, full scale study that was able to be done thanks to having having access to
9:12that data and to be able to dig into it that way.
9:16The conversation about the accuracy polls versus prediction markets continues to heat up.
9:20It's something that is discussed on social media ad nauseum.
9:24And I think just a table set for our audience, polls are obviously a snapshot of time, whereas these prediction markets are not only dealing in real time.
9:33They're also looking to be forward facing as it relates to what is going to happen.
9:39So let's just back up a moment. How would you frame the advantages and disadvantages of each of polling and of prediction markets themselves?
9:47Well, I think that just starting with the. I think if you were to give an advantage to the two, I mean, the prediction market has it.
9:56Well, first of all, it depends what your objective is.
10:00But if your goal is to kind of get insight into what's going to happen, then the prediction market is tailored towards that exact question.
10:08Right. Which is why, you know, referring back to the study, we do see people who are Republican who might have voted,
10:16who might have bet for bet on a Democrat or vice versa, because the market isn't actually asking who are you going to vote for?
10:24Right. It's just asking who do you think is going to win?
10:28And more specifically, you know, do you think the odds are a good price for for what who you think is going to win?
10:36And so it is directly more tailored to that question.
10:40Now, polls, the benefit of the polls is that it actually is collecting data on the people who are actually voting and markets do use the data available.
10:49So polls are a key data point that are getting used and that are being priced into the market.
10:56And so I think that it's it's that they're both kind of well, and also polls are asking the people who are
11:03in the poll, who are you going to vote for?
11:07Which, you know, if it's October 1st, that might be one person.
11:11And then on October 30th, it's a different person.
11:14People change their minds. And as you said, the market is saying who's going to win on this date.
11:20So it kind of doesn't matter. You know, you kind of have to factor in, well, people might change their minds or
11:28there tends to be some kind of shift in one direction or the other as as things get close to the election
11:35or trying to read into those trends. And so they are kind of two different.
11:40They address two different questions. They are related in that, of course, polls are designed to try to give insight into the question of who's going to win.
11:50And it does it by asking a different question.
11:53And then, of course, prediction markets are leveraging all the information available and polls being one of them.
11:59Early voting data, registration data, qualitative news data and other things like that being, you know, additional information that can and is
12:07priced into the prediction markets, which is why over in general they're more accurate because theoretically, I mean, even just thinking theoretically
12:14or logically, they shouldn't be less accurate because at the very least they could just copy the forecast from the polls.
12:21And that's kind of your base case. And so you can only get better if you add more information on top of it.
12:29Kind of a two pronged follow up here. You talked about the idea of, you know, polling someone on October 1st and then polling them on October 30th.
12:39And they may go in two opposite directions at the ballot box.
12:43How late in the game when you look at, you know, historical data for voter preference can really someone be moved?
12:50You know, obviously there's entrenched voters on both sides and there's that meaty middle of independent swing voters that in many cases
12:58wait until, you know, the 11th hour to make their decision.
13:01But can polls capture that for starters? Because, you know, there's polls in the field right now.
13:07They're kind of in that last wave before we get to election week.
13:11So this is kind of as late in the game as they can operate.
13:16And then as kind of a follow up to that, if prediction markets are being fueled by this hard data that's being
13:24collected, once that dries up, what is what are those markets really, you know, motivated by and being moved by?
13:31If there's no new information and it's just, you know, the sizable nature of how much money is flowing in on one
13:38candidate or the other and a little bit of that groupthink starts to filter in.
13:43Right. So, I mean, on the first question is could polls pick up kind of a shift, this shift that you're talking about?
13:51I mean, in principle, they could. In practice, they probably don't or it's extremely hard.
13:56One thing is, I don't know how most polls are conducted or how, you know, any of them really.
14:03But if the poll is sampling different people every time, then, you know, that's going to be hard to pick up.
14:10You can maybe pick up trends, but those trends could also be subject to noise.
14:15For example, one of the big things I was looking at in 2024, and you can easily be fooled by things like
14:22this, which is why I had to get deeper and deeper into it.
14:27But one of the things that was really sticking out to me was looking when they polled, certain polls would ask people,
14:34who did you vote for in 2020? And who are you voting for now?
14:39So, you could kind of see the switchover effect of people who, let's say, voted for Biden and now are voting for
14:47Trump, or people who have voted for Trump and now are voting for Harris.
14:51And there was a fairly consistent, Trump had a fairly consistent like four or 5% advantage in that demographic of, you know,
14:59very small demographic, but that's kind of where the battle is, right?
15:03Somebody who voted for Trump once and is voting for him again, or voted for Biden once, voting for Harris now.
15:10That's kind of the entrenched voter in this case.
15:13The elections won on the margins. Well, that was something that I was looking at.
15:18You're asking, you know, if someone's going to change their mind in a month, very few people do.
15:24And so even theoretically, if you are in principle, if you had a poll where you were polling the same thousand people
15:32in October, October 1st, same thousand people at the end, you would obviously pick up which of those people changed their mind.
15:39The issue from a statistical standpoint is to say, okay, is it going to be a statistically significant observation?
15:46Because out of a thousand people, did 10 people change their mind?
15:50Did 20? You know, it's going to be a very, it's generally probably a very small subsample.
15:56With, there could be exceptions to that though. And so I think that that's why I say in principle, if you had
16:03a large enough sample and you were able to kind of carve out enough to make it meaningful.
16:09I think what probably is more likely to happen or what tends to happen is it's that, you know, that hypothetical people
16:17who are undecided on whether they're going to vote or not.
16:21Right. So it's kind of like half a vote.
16:24Are they going to vote for Trump or are they going to not vote at all?
16:29Or, you know, the other way around. And that is a case where you might actually get some information because people could
16:37be turned off or motivated and you get a, you get kind of a look into how motivated a case is.
16:44And that could be something that extrapolates beyond just the small sample in your poll.
16:49Midterm elections are just around the corner. And we've partnered with DraftKings where you can follow election markets from the biggest national storylines.
16:57One of the things we've been doing here at the Action Network is following how the markets are reacting as the races change day to day.
17:06And what's been particularly interesting is watching the percentages and price charts move as new polls come out, news breaks, and we
17:13get closer to election day on Tuesday, November 3rd.
17:16With DraftKings, you can see what the market is saying, then make your own real money prediction on how you think things will play out.
17:25DraftKings has all the election excitement in select states, so be sure to check it out, follow the races, and see how the markets are moving.
17:34DraftKings, the crown is yours. Lisa, we should tell them about our new show.
17:38Oh my God, Lexi, we have a new podcast.
17:41It's called Don't Lie to Me. I'm Lexi Iwano.
17:45We're two best friends, 22 years apart, but so close at heart.
17:49And I'm Lisa DiGiulio, the old one, which means we have wildly different references, occasionally different opinions, and at this point in my 50s, I've lived it all.
17:58I'm not going to get anywhere fighting. That is wild.
18:02I learned that about you today. Now in my 30s, being the founder of New Shoes, good taste is my business, and I've always got something to say.
18:11If men were just honest about this, there would be so much more peace in the world.
18:17We're going to talk about dating, relationships, beauty, getting older, and a little wiser.
18:22We're best friends, but we don't always agree. Because you told us how accommodating you are.
18:27Obviously I would do what I think I could do to fight the fight, but there's only so much fight I can fight.
18:35Don't lie to me. Don't lie to me. Don't lie to me.
18:39Don't lie to me. Don't lie to me. Don't lie to me.
18:43Liar. Listen on the iHeartRadio app, Apple Podcasts, or wherever you get your podcasts.
18:48I don't really care. There are thousands of podcasts that can tell you how to better yourself.
18:54Get Worse with Kyle Pru is not one of those podcasts.
18:57So if you're interested in getting thinner, going to therapy, starting peptides, setting boundaries, taking supplements, making your better, getting closure, then I would humbly recommend other podcasts.
19:07But if you're interested in blackmail, love bombing, cloning yourself, or getting revenge, then we've got a little something something for a dastardly individual like you.
19:16Take a listen to some of the incredible advice we've given so far.
19:20Fake a back injury. Get the spray tan. Up your alcohol consumption.
19:24Only women should be allowed to have guns. Kill them.
19:28Kill them. Kill them with a gun. Listen, it's already over.
19:32I'm all about lying if it's something that really happened.
19:35The next time a bear is running at you, you just calmly stop what you're doing, you look the bear in the
19:42eye, and you say, this isn't you. Do not preserve relationships.
19:46You don't need them. You only need one thing, and that's the Get Worse podcast.
19:51Listen to Get Worse with Kyle Pru on the iHeartRadio app, Apple Podcasts, or wherever you get your podcasts.
19:58I'm Lacey Mosley, author, actor, stand-up comedian, creator, and host of Scam Goddess Podcast, the true crime comedy podcast all about robbery, fraud, and those who practice it.
20:07Each week I bring on incredible guests like Conan O'Brien, Iowa Debrey, Lamorne Morris, and Nicole Byer.
20:13Ooh, I scammed them. What? I could talk to you for an hour.
20:18You're a generational yapper, like me. That's why I had to make yapping a job.
20:23And the scams run the gamut. I'm talking about scams from the 1800s all the way up to today.
20:29I also warn you and keep you safe against scammers.
20:33Call your grandparents. Tell your parents to stop clicking on links, please.
20:37And now you can catch full episodes of my show on YouTube.
20:41If there's a scam, I'm here to protect you.
20:44And I'm also here to laugh with you. So if you like fun, if you like comedy, and you like learning and
20:51protecting yourself from criminals, start listening to Scam Goddess on the iHeartRadio app, Apple Podcasts, or wherever you get your podcasts.
20:58Scam Goddess! A decade ago, the ethanol kingpin of Iowa became the king of corn in Brazil.
21:04So we met with a lot of larger farmers, went from Bahia to Tocatines to Mato Grosso.
21:10And he brought a team of executives. They were going to help the country get in on a gold rush.
21:17Carbon and its derivatives are going to be really the next great commodity that the globe's going to trade.
21:23But back home in Iowa, trouble was brewing. If you live in Iowa, your land, your water, and your voice could all
21:31be at risk thanks to a man named Bruce Rastetter.
21:34Now people are questioning if his climate solutions have anything to do with climate at all.
21:40You gotta give Bruce and the guys credit. They're Republicans.
21:43They don't give a s*** about any of this now.
21:47On this season of Drilled Carbon Cowboys, the story of how the ethanol kingpin of Iowa became the king of corn in
21:54Brazil, and what it tells us about the limits of technology and markets to solve the climate crisis.
22:00Listen on the iHeartRadio app, Apple Podcasts, or wherever you get your podcasts.
22:05I'm former U.S. Women's National Soccer Team player Carly Lloyd.
22:08I'm Fox Sports reporter Jenny Taft. And on our new podcast, Whatever It Takes, we're talking motherhood, marriages, careers, and figuring out how to balance it all.
22:17And we'll discuss the game that brought us together, because nothing is off limits.
22:22Do you remember your initial reaction when Fox called you to be a part of the World Cup?
22:28I was like, no, I don't want to do it.
22:32You said no? I played with players that didn't, I don't think that they really, truly love the game.
22:38I found the love with the ball up my foot.
22:41But in my head, it was the long game.
22:44It was like, I don't care what I'm being asked to do at Fox.
22:49I'm going to go get some reps, and I'm going to learn it and love it and embrace it.
22:56He said, this group of people will never be in the same room as part of the same team ever again in your life.
23:04And it is your job to make sure that you enjoy the moment, you're proud of what you accomplish together.
23:11Listen to Whatever It Takes with Carly Lloyd and Jenny Taft on the iHeartRadio app, Apple Podcasts, or wherever you get your podcasts.
23:19Prediction Pro is now live. Head to predictionpro.com today to check out our brand new subscription offering.
23:24You can compare prices across prediction markets to make sure you get the best price, and use Prediction Pro's free tools to
23:32help you find the best edges possible, including access to our model that will help show you where we see trading edges in the market.
23:40Plus, daily expert analysis and analyst trading tracking. Now, Nate Silver and FiveThirtyEight, they model out their probability on races, but it's almost entirely built on polling.
23:50And it brings up a big question, which is if the polling is off, then their inputs are fundamentally not accounting for
23:57something that is going to basically skew their output to a point that it's no longer useful.
24:03And we've heard this from Republican strategists and pollsters saying that this particular cycle, it's been increasingly difficult to get the classic
24:10MAGA voter to participate in online panels, in polling at a large enough sample size for it to be meaningful.
24:17How much do you believe, and you talked about getting into the weeds in these last few cycles and looking at polling
24:25and kind of trying to fill in the gaps, so to speak, how much do you think on the right this is
24:32a talking point to try to make sure that there's not disenfranchisement across their base?
24:37Or do you think it's true that there's just this mystery box as it relates to what the classic MAGA voters are going to do this time around?
24:47I don't have a direct answer to your question, because I think that is kind of a fundamental question that I feel
24:54like I've been hearing it now for four, five, six elections or every election, which is that one is that there's a democratic bias in the polls.
25:04Although if you look at the polling results, that kind of has proven to be the case.
25:09But it's also still, and that's one of the things that makes, I think, political prediction and betting so interesting is that
25:17although we've been talking about this for 10, 20 years, you could even go back 20 years, there's only been how many
25:24elections in the last 20 years, right? And so it's like things are very much moving in slow motion.
25:31And then there's always the point of, well, there was a bias observed last time.
25:36And so there's an expectation that a good pollster would adjust for that.
25:40You, of course, don't know if they did. And so there's kind of that question, which I think does make modeling of polling difficult.
25:49But all of these poll, all of these forecasters, Silver or others, they're all making assumptions.
25:54They all have assumptions in their model. Every model has assumptions in it.
25:58That's the way it works. There's nothing wrong with that.
26:02The trouble is we, if we're consumers of this information, don't necessarily see what those assumptions are.
26:08It's not obvious necessarily when the assumptions are wrong.
26:11And there was a good example in 2024 where if you look at all of the market prices and you look at
26:18all of the forecasters from, let's say, July, August, September, October, into the presidential election, they're kind of going in the same direction.
26:26Now, the polls were below the markets, but they were kind of parallel, let's just say.
26:32When one went up, the other tended to go up and so on.
26:36Except there was a period between mid-August and mid-September where everybody was going in the same direction.
26:42And Nate Silver went up to Trump 70% and then came down.
26:46It was predictable. And I was actually predicting it on my Twitter at the time saying, in five days, Silver is going to be here.
26:54In five days, he's going to be here. You could kind of tell what was happening.
27:00But that was a case where it's an obvious model error.
27:04It's now because he has some assumption in his model about the convention effect and how polling kind of tends to favor the convention candidate after the convention.
27:13And since Harris didn't get that favorability in the polls, Trump's chance went way up.
27:18But then it kind of settled back down. I think he's since corrected the fact that that was just an anomaly in 2024.
27:26So I bring that up to say there's all kinds of stuff like that in polls, in models.
27:32That's why the betting markets, I think, are just the better alternative because, yeah, betting markets make mistakes as well.
27:39There's anomalies there. They depend on polls. And the polls we're finding might all be fake or a lot of them are fake.
27:47The other ones are biased. But that is kind of the job of a bettor is to try to figure that out
27:54and to really do the extra investigative work into the weeds, as you said, to try to see, you know, is there
28:02extra edge that I can get out of, you know, looking into this further?
28:07Now, I mentioned earlier, you have a great website, harrycrane.com backslash elections.
28:11Phenomenal resource to compare prediction markets, the consensus of all prediction markets to the models that are forecasting what's going to happen on November 3rd.
28:19Let's talk specifically, though. We've talked a lot about the last cycle in 2024.
28:24There's a big difference right now in terms of some of these models, these prediction markets and polling as it relates to the Senate.
28:32Prediction markets have this just a touch above 60 percent that the Democrats are going to win control of the Senate.
28:39A huge delta in the models forecast, though. Vote predictor forecasting only about a 50 percent chance, while silver bulletin, you know,
28:47from Nate Silver giving Democrats a 76 percent chance to take the Senate.
28:51At least in your opinion, can you explain how these models can be so far removed from one another, despite the fact
28:59that most of their inputs should be the same?
29:02At least theoretically, they should be the same. Yeah.
29:05I mean, I think that that's – it's interesting because you look at this and the numbers are what you said.
29:12So you've got the market, which is at 65 percent for the Senate for Democrats, and then you've got a couple that
29:20are up in the 70s or close to 80, and then you've got a couple down near 50.
29:26So they're on either side. And I think that that is – again, that comes down to the assumptions that are built into the model.
29:34They're all presumably using the same data or essentially the same data.
29:38But someone's – you know, in a case like this, it's just a given that some but one of these parties is
29:46model – well, I guess you'd have to look – one of two things is happening.
29:51Either at least one of them is just bad compared to the other ones, or they're all equally bad and they're all
29:59kind of noise – they're all just kind of noise on top of the main signal.
30:04Right? So if you've got a prediction market at 65 percent and then you've got plus or minus 10 percent is where
30:12some of the forecasters are up here, some are over here, then they're essentially – there's an argument there to say, well,
30:19the forecasters are really just noise on top of the prediction markets.
30:23I don't know which one it is, but I think that both are plausible.
30:28And I think that they definitely – I think it definitely is going to come down to some kind of an assumption.
30:36There's probably a modeler here who is either making a kind of an ad hoc adjustment or has an estimate for what
30:43the democratic bias is and or has an assumption of the uncertainty in the polling data.
30:49So they are either trusting polls too much or not trusting them enough, and that's kind of what's – those are kind
30:56of the two obvious candidates for what would cause deviation in the way that we're seeing here.
31:02Going from the battle for the Senate holistically to some of the races specifically, you're also seeing big polling differences from these prediction markets.
31:10South Carolina, for example, that Senate race between a Republican appointee, Darlene Graham and Democratic pediatrician Annie Andrews, the governor race in Oregon as well.
31:19My question for you is when you look at some of the polling, when it comes to state-by-state races, do you start
31:26to get into the historical preference of their voters to potentially vote for both parties, having those tradeoffs and those swaps and
31:34their willingness as opposed to historically just voting a straight party line across the board?
31:39Or do you kind of wipe that slate clean and say there's so many variables and elements in the environment right now
31:46that are different and specific to 2026? I'm going to throw the baby out with the bathwater.
31:52Or do you kind of mix and match? I think it's a little bit of both.
31:57I think that if you look in a place like South Carolina, that's a strong Republican spot.
32:03Trump won very convincingly. So even if there's people unhappy with Trump or something, yeah, they might lose voters, but it's pretty
32:11hard to overcome a 15% to 20% advantage or whatever it may have been.
32:15Also, since these are such strongholds for one party or another, they tend not to get high-quality polling.
32:22The polling that I've seen, a lot of it is getting voted.
32:26A lot of it is you've got 42% for one and 42% for the other.
32:31So where's the other 16%? And so that's supposedly undecided or something like that.
32:36I don't know if 16% of voters in South Carolina are undecided.
32:40Maybe they legitimately are or they won't show up or they'll do something.
32:44But in all likelihood, there's just kind of low quality in those polls.
32:49And the historic – but also – and I don't have this right in front of me right now.
32:55If I'm just looking at what the Senate race in South Carolina is, for example, it's saying about 82%, 80% in the prediction markets for the Republican.
33:04I mean that – I don't know what it was, let's say, six years ago or whenever the last time Graham ran.
33:12I know this is a different Graham. But 80% for a Republican in South Carolina kind of seems low to me.
33:19So that kind of – and I don't know.
33:22Maybe I'm misremembering. So, you know, in some sense you could argue that is kind of maybe influenced by the fact that the polls are what they are.
33:31I get in this argument all the time as someone who covers college basketball, college football, college baseball.
33:38And, you know, the casual listener may say, if you get something wrong and you chalk it up to a bad beat
33:45or you have some kind of complaints, you know, that's something that should be removed.
33:50Either get it right or you get it wrong.
33:53But the reality is there's always going to be moments – Texas State against San Diego State game last month where they more than doubled their yardage.
34:02But somehow found a way to lose the game.
34:06I view that as my process was right, the outcome was wrong.
34:10When you look at what you were able to accomplish in 2024, predicting every single state correctly in that election, when you
34:17go back over your process, is it all sunshine and rainbows?
34:21This was perfectly calibrated? Or do you still view it as there's learning lessons about something that you could, you know, refine
34:28the next go around and make sure that your process is always evolving?
34:33Yeah. I mean, I think that it's easy in a case like that to feel – I mean, anytime that happens, of course, you feel good.
34:42You feel like a genius. But, you know, the opposite is when the opposite happens and you feel like an idiot, right?
34:49And that's kind of the experience of doing this and of betting on these things.
34:54I think, you know, of course, I didn't get any evidence to the contrary that my method was wrong.
35:01But I admit – I've admitted all – you know, there was a lot of, you know, guesswork in that process because,
35:08I mean, it's the nature of the game that it's guesswork.
35:12And I think that an election like that or an election in general, like I said, there's very few repetitions to really evaluate your model.
35:21You bet on basketball, baseball, sports, you can refine a model statistically and, you know, you can – you get that data.
35:28You don't really get it in politics. And there were ad hoc elements to my approach, very ad hoc, you know, not
35:36ones that you would publish on a website because everyone would say what's literally just made up.
35:41But what's not made up is, you know, you're putting a bet on one side or the other.
35:47You know, one thing that I like to do or I have done and this is because it's worked over several elections
35:55and, of course, it's a small sample and it's anecdotal.
35:58But it gets to that question of the marginal voter who may have voted for, you know, Trump.
36:05Yeah, I know people who voted for – who were in 2016 voting for Bernie and then when Bernie lost, they voted
36:12for Trump and then in 2020 voted for Biden and then in 2024 were voting for Trump.
36:18You know, I know several people like that and I take those people's opinions pretty seriously because they seem to track some
36:25segment that gets – you know, the Bernie to Trump was a legitimate thing that people thought was crazy.
36:32But, you know, when you know somebody like that, you're like, well, they're tapped into something that I'm not.
36:38One final question before we get out of here.
36:41There's a narrative in sports betting that every year markets get a little bit sharper.
36:46Now, this is different as it relates to prediction markets because you don't have a house necessarily setting odds and refining things
36:54and getting ahead and, you know, taking into account home field or home court advantage or historical travel spots or letdown spots.
37:01All of that's off the table. It's really just on the crowd.
37:06Do you think that the crowd is getting sharper year over year or do you think that's an overblown talking point?
37:13Well, again, not to beat a dead horse here, so to speak, but it's hard to say because there have been –
37:20now, this is our – this is 2026. There really aren't that many samples to go on.
37:26And logically, you would think they are going to get sharper because, well, you always learn from the past.
37:32And also, as they become more popular, as more money gets into them, of course, there's incentives for these big syndicates to devote resources to them.
37:41And, you know, the French whale was a good example, right, in 2024.
37:46The guy who bet – was it 30 to 50 million, something like that on Trump?
37:51To me, that was what it – that's what I intuitively thought when I saw the price – when I saw these stories.
37:59But that's only because I'm familiar with that happening in other domains.
38:03And politics is almost one of the best domains for it where he was doing his own polling because, you know, my
38:11– one of my – this thing I can't get out of my head and with political prediction or betting specifically, which is different than sports.
38:19You know, before the game starts in sports, you don't know the outcome and it does have to play out, you know,
38:27and there are things that will happen. Theoretically, you know, in a country of 300 million people, the election is – it's
38:34guaranteed that one of those two people is going to win because everyone – enough of those people know exactly who they're going to vote for.
38:43And that's a deterministic outcome. Of course, the reason why there's probabilities is we don't know what people's opinions are.
38:50But theoretically, if you polled enough of them and you had enough insight into them, you would have a pretty good idea of what's going to happen.
38:59And so it seems almost very – you know, a great candidate for private research because the polls that we see online are 1,000 people.
39:08So there's a margin of error of 3%. Well, if you – it takes a lot to bring that margin of error down.
39:16But if you go up to 10,000 or 100,000 people in a poll, you now do narrow that down quite a bit
39:23and you could get a much bigger edge and a much sharper estimate of what's going on.
39:29If you'd like to follow Harry over on X, Harry D.
39:33Crane, also head over to his website, harrycrane.com. Has lots of fun visualizations for everything midterm elections.
39:39Harry, thank you so much for joining the program.
39:42Thanks a lot. Great stuff from Harry. We're going to be back on Friday with Eli McCown Dawson.
39:48He's a senior elections analyst for Nate Silver's Silver Bulletin.
39:51Lots of early voting data is pouring in. We're going to get to all of that, his thoughts on it, and more
39:59as we inch closer to Election Day. See you then.
40:02Action Network reminds you, please gamble responsibly. If you or someone you care about has a gambling problem, help is available 24-7 at 1-800-GAMBLER.
40:10Someone once said everybody's got a plan until they get punched in the mouth.
40:15I did. Mike Tyson. Oh, well, Mike Tyson and Boost are here to tell you that your mobile plan is punching you
40:23in the mouth with prices that can knock you out.
40:26That's a low blow. Totally. But you can do something about it.
40:30You have to punch back. No, Mike. You got to switch to Boost with a plan that's $25 a month forever.
40:37Oh, that's a great plan. Maybe you can do both.
40:41Boost Mobile. What's your plan? Lisa, tell them about our new show.
40:45We have a new podcast. It's called Don't Lie to Me.
40:48I'm Lexi Iwanu. We're two best friends, 22 years apart, but so close at heart.
40:54And I'm Lisa DiGiulio, the old one, which means we have wildly different references, occasionally different opinions.
40:59And at this point in my 50s, I've lived it all.
41:03Now in my 30s, being the founder of New Shoes, Good Taste is my business.
41:08Listen on the iHeartRadio app, Apple Podcasts, or wherever you get your podcasts.
41:13I'm Lacey Mosley, author, actor, stand-up comedian, creator, and host of Scam Goddess Podcast.
41:17The true crime comedy podcast, all about robbery, fraud, and those who practice it.
41:22Each week I bring on an amazing guest. Ooh, I scammed them.
41:26What? I could talk to you for an hour.
41:29You're a generational yapper, like me. That's why I had to make yapping a job.
41:34And now you can catch full episodes of my show on YouTube.
41:39Start listening to Scam Goddess on the iHeartRadio app, Apple Podcasts, or wherever you get your podcasts.
41:44Scam Goddess! There are thousands of podcasts that can tell you how to better yourself.
41:49Get Worse with Kyle Pru is not one of those podcasts.
41:53So if you're interested in getting thinner, going to therapy, starting peptides, setting boundaries, taking supplements, making your bed, or getting closure,
42:01then I would humbly recommend other podcasts. If you're interested in blackmail, love bombing, cloning yourself, or getting revenge, then we've got
42:08a little something something for a dastardly individual like you.
42:12Listen to Get Worse with Kyle Pru on the iHeartRadio app, Apple Podcasts, or wherever you get your podcasts.
42:18Every case has a story. Every victim has a voice.
42:22But will it be heard? I'm Nancy Grace. On Crime Stories, we go beyond headlines to examine evidence, challenge the story we're
42:29fed, and pursue justice for victims and their families.
42:32Please join us at Crime Stories with Nancy Grace on the iHeartRadio app, Apple Podcasts, or wherever you find your podcasts.
42:39Hope you join us. I'll see you there, friend.
42:42I'll see you there, friend.
More episodes from The Action Network Sports Betting Podcast
NFL Touchdown Show | Week 7
Oct 22, 2026
NFL Week 7 Betting Preview
Oct 21, 2026
NFL Futures Watch | Week 7
Oct 20, 2026
NFL Betting Playground | Week 6
Oct 16, 2026