We’re just under a month away from the election, which means that gambling season is upon us. I’ve decided to create some markets, with actual money, for people to bet on. I’m doing this through a partnership with Manifold, which is sponsoring this post. As most of you know, it is the site that lets you create your own markets and bet mana, their own play currency.
Manifold is now also doing Sweepstakes markets, which allow you to win real money by cashing out when you win. See the rules here. Sign up and use promo code FOP9O and you’ll get $10 in both mana and sweepcash.
This post introduces the four Sweepstake markets that I have been involved in creating and will be betting on. If you want to come bet with or against my positions, click on the embedded links to the markets below.
Note that to bet on the Sweepstakes version of the market, you need to click on the yellow S tab in the top right corner. See the two screenshots below.
If the purple M is highlighted you’re in the mana market. If it’s the yellow S, you’re in the Sweepstakes market, and playing for real money
The four Sweepstakes markets I’ve created can be grouped into three broad categories: vibes based, long-shot based on uncertainty the market is not taking into account, and a bet where the market is overestimating a tail end event. Let’s consider these in turn.
Ted Cruz Is No NFL Star
Republicans have a good shot at taking the Senate. They are sure to flip West Virginia, which would put them at 50 seats. That alone will give them control if Trump wins the presidency. If not, Montana looks overwhelmingly likely to flip too. Ohio, Pennsylvania, and Michigan are also up for grabs.
All of this assumes Republicans don’t lose any seats. I think if there’s one place it might happen, it’s Texas, where Ted Cruz is running against Colin Allred, a UC Berkeley School of Law graduate, member of the House, and former linebacker for the Tennessee Titans. Right now, the market gives Cruz an 81% chance of retaining his seat. Note again that if you want to bet on the Sweepstakes market below, click on the link and then the yellow tab in the top right once you get to the site.
I say Allred is underpriced, but my view is mostly vibes based. With every election, one needs to start with the polls. As I write this, Cruz is leading by an average of 2.7 across the last six that have been conducted. Since Texas overwhelmingly votes Republican, most analysts would say that the results are likely to revert to the mean of the state by Election Day.
But I just feel like Allred is a great candidate, and Ted Cruz is, well, Ted Cruz. Polls that show a tight race report that Allred’s favorability is positive, while Cruz is underwater. The challenger is of course less well known, but this gap will narrow by Election Day. Cruz’ unpopularity in his state gives me confidence that my dislike of the man is shared by others and not based in some quirk in my personality.
This recent Texas Monthly profile of Ted Cruz reminded me of why I find him so distasteful. There’s something off about the guy. And the fact that he’s smart makes him engaging in MAGA theatrics that much more disturbing. When the whistleblower nonsense was making the rounds last month, the only politicians I saw fall for it were MTG and Ted Cruz. Now, I expect this from Greene. She’s perhaps doing the best she can with the gifts God gave her. But Cruz is either lying or suffering from one of the worst cases of brain rot we’ve seen in politics. A US senator chasing clicks is a depressing sight. I remember one time seeing him advocating the need for fewer doors in order to prevent school shootings, and found the combination of the goofy ad hoc nature of the argument along with the passion with which he delivered it to be unsettling.
It’s one thing to say that stuff happens, dead schoolchildren is the cost of freedom. Or one can call for gun control, or just remain silent. To say that door control is the solution, however, plays to a voter who must be an outlier in terms of both motivated reasoning and stupidity. What kind of man does this? Perhaps one who listens to Trump call his wife ugly and then spends the next nine years on his knees treating him like Jesus in lockstep with the rest of the party. What worms these people are.
What were we talking about again? Oh yeah, prediction markets. Anyway, maybe I’m too emotional to judge Ted Cruz’ prospects objectively. But based on vibes and perhaps my own wishful thinking, I’m going to give Allred a 30% chance of pulling off the upset.
Trump Wins the Popular Vote and Loses the Electoral College
In 2016, Hillary won more votes but Trump became president. The next time, the same thing almost happened again but Biden barely squeaked by in the key states. Republicans have a natural advantage in the electoral college, and recent history tells us that there’s a distinct possibility we see another election where the GOP candidate gets fewer votes than his opponent but nonetheless is inaugurated on January 20.
For a while now, prediction markets across platforms have given about a 25% chance that Trump wins the electoral college but loses the popular vote and a 25% chance he wins both. Kamala winning both is given about 50% and rounds off the odds. The possibility that Trump actually wins the popular vote and loses the electoral college is considered extremely remote. Right now it’s at a 2% chance in the Sweepstakes market.
Nate Silver might say even that is too high. He recently found that in 40,000 simulations, a split outcome where Democrats win the presidency happened 102 times, which translates to a 1 in 400 chance.
My view is that this should be closer to 3-4% instead of 2% or even 0.25%. In a world where there are four possible outcomes, you start with an assumption that each has a 25% chance of happening. Of course you then adjust your forecast based on what we know about the real world. Here, we think we know the election is close, and these days Democrats tend to run up the numbers in population-rich blue states while struggling more in much of the rest of the country. So a split outcome favoring the Democrats seems unlikely. But by how much?
According to Nate Silver’s model, right now Kamala is leading in the national vote by 3%. She’s also winning in enough states to pull off a slight victory in the electoral college. Simply imagine that the swing state polls are right, but national polling is underestimating Trump by 3%-4%. That’s all it would take to get the split outcome favoring Democrats. Now, this is unlikely because of the tendency of errors to be correlated; polls underestimating Trump nationally means they are likely to do so in swing states too. But I can imagine a scenario where he does much better than expected in California and New York but lands about where we thought he was in Pennsylvania, Michigan, and Wisconsin. Not the most likely outcome, but I think it’s higher than 2%.
A recent article by Nate Cohn discusses how pollsters are just sort of winging it. Since they underestimated Trump twice already, some of them are bending over backwards not to do so again. Them over- or under-shooting the mark seem about equally plausible. The lesson I take from this piece is that we know it’ll probably be close but have no idea whether the pollsters are doing a good job capturing current dynamics in swing states. When there’s a great deal of uncertainty, bet on the event that is given a very low probability based on models that assume we know a lot more than we do.
Democracy Will Survive
Right now, the Sweepstakes market gives a 10% chance that the election will not be certified on January 6. The mana market is higher, at 18%. I think that’s crazy, and the number should be closer to 4%. Note that even in 2021, when a mob shut down Congress for six hours, the election was certified by 3:32 AM on the seventh. In part, this was because Congress after having scrammed for its life was determined not to let the insurrectionists win. This January, the Secret Service is going to beef up security, and authorities know to expect trouble. In 2021, the surprise nature of the event made sure that they were caught flat footed. You don’t expect a speech by the president to end up with his supporters marching towards the Capitol to try and hang his vice president, but the Trump era has been full of surprises.
Another storming that delays the results for several hours is probably not in the cards. One could imagine other shenanigans though. Could Democrats or Republicans mount a serious challenge to the election results that end up delaying certification? We should consider each scenario in turn. If Trump wins, I think Democrats obviously accept defeat. They have made acknowledging the outcomes of elections core to their identity over the last four years, and being the Elite Human Capital party, they tend not to turn on their beliefs that quickly.
If Trump loses, he’ll of course demand every Republican in any position of power do whatever it takes to get him back into office. But he couldn’t pull off a coup even when in the White House. He’s much less likely to be able to do so while holed up in Mar-a-Lago. In Georgia, Trump has apparently gotten control of the election board and they’ve implemented some MAGA preferred policies. I’m unsure if this means that Trump supporters are in a position to steal Georgia, but the other swing states don’t see anything similar going on. For the certification of the election to be delayed, the entire outcome must depend on Georgia, the results in that state must be close enough to be controversial, MAGAs have to try to steal it, and this has to remain unsettled in the courts by January 6. This is a highly unlikely series of events. I guess it’s also possible that the certification gets delayed based on controversy in one state even if it doesn’t matter for the election outcome, but I expect Congress to put its foot down here.
Perhaps there’s a 4% chance something goes wrong. MAGAs might plan a major terrorist attack. Someone could get assassinated between the election and January 20. Yet these scenarios are very far fetched.
For similar reasons, I think we’re getting an inauguration on January 20. The market gives it a 98% chance of happening as of now. I say it’s over 99%.
Three Types of Bets
These three categories of bets I’m making here are recurring types in prediction markets. To make money, you have to outsmart the conventional wisdom. Sometimes, the prevailing view overestimates tail end events, in which case you bet on the sure thing and treat your position as a low-risk bond. I put a huge portion of my net worth into Michelle Obama not being the Democratic nominee back when the market was giving that a 5-10% chance of happening due to conservative stupidity. I never for a moment felt my money was at any risk.
Other times, an event is given a very small chance of happening, and you think going against the crowd will return a positive expected value, even if you are likely to lose. If Trump wins the popular vote but loses the electoral college, I’ll be a very happy man, as I put a lot of money on this happening, even though I’m pretty confident it won’t. Think about this. If you are good at picking out events that the market puts at 2% probability but should be 4%, you’ll lose 96% of the time but end up doubling your money! My experience is that these markets are rare so when you see one you should jump on it.
Finally, there’s stuff that’s in the middle of the distribution, where you think the market is wrong by maybe 5 to 20 percentage points in one direction or another. This is the Colin Allred bet and it might just be based on vibes. Allowing oneself some emotional indulgence once in a while out of a hatred for Ted Cruz is fine and adds some fun to the whole process.
If you have decent judgment on these things, you too can make good money through a combination of safe bets, positive expected value lotto tickets, and vibes.
Remember when Ted Cruz personally blocked U.S. visas for the anti-CCP Hong Kong protestors? I'm right there with you on the wishful thinking Allred bandwagon.
https://reason.com/volokh/2020/12/21/ted-cruzs-terrible-case-for-keeping-out-hong-kong-refugees/
Regarding the odds of certification on Jan 6th, one factor not in your analysis is the various county boards of elections - key counties, in swing states - that are now controlled by MAGAts, or where changes in state law now give them the ability to hold up the certification process. Georgia is the obvious outlier that you mention, but there are many others (see: https://www.rollingstone.com/politics/politics-features/trump-swing-state-officials-election-deniers-1235069692/ , or the database at https://publicwise.org/election-threat-index/ ). Many only hold sway in very red counties, but the list most notably includes the chair and vice-chair of the WI state election commission. And it wouldn't take many of them gumming up the works in litigation for 6 weeks to delay a swing state past the Dec 14th electoral college deadline, giving rise to a wider array of potential objections on Jan 6th.
The 10-18% odds of a delay past Jan 6th that were given at time of writing seem fair to me - I'm certainly not very high-confidence in any number meaningfully below that.