Why Smart Money on Polymarket Buys What No One Wants

Prediction Markets and the Price of Boring

Prediction markets are supposed to tell us what’s going to happen, though arguably they do a better job at telling us what the consensus expectation is. In practice, they mostly tell us what people want to bet on. We’ve spent a lot of time trading and studying prediction markets at Starkiller, in fact our Chief Investment Officer, Leigh Drogen, was the founder of Estimize, the largest corporate earnings prediction market in the world where the founder of Polymarket was a participant while still in high school.

What stands out in these markets isn’t that they are bad at forecasting, they tend to be pretty good, It’s that different participants are playing completely different games. Some people want action. Some want narrative exposure. Some want a miracle. Very few want to wait. That mismatch creates a very specific kind of opportunity and alpha.

Binary Markets, Real Endpoints

Every prediction market contract resolves to either $1 or $0. Some resolve on a specific date, others resolve when an event happens, but all of them include a backstop resolution date if the event never occurs. The important part is that the terminal payoff is always known. As time passes and information accumulates, many markets become effectively decided well before they formally resolve. When that happens, prices stop being about uncertainty and start being about preferences, especially preferences around excitement, optionality, and capital usage.

The Gambler’s Premium

The gambler’s premium in prediction markets isn’t about liquidity, it’s about convexity and vibes.

Long shot outcomes are more fun. They come with upside, drama, and the possibility of being spectacularly right. Near certain outcomes are boring. They just sit there, slowly grinding toward a dollar. Because of that, long tails are consistently overbid. People pay too much for small probabilities because that’s where the action is. This isn’t unique to prediction markets. You see it in sports betting, lotteries, and options markets all the time.

The tendency to overpay for long shot outcomes and underpay for near certainty is a documented feature of betting and prediction markets, often attributed to probability misperception and a preference for lottery like skewness. In one often cited paper, Longshots, Overconfidence and Efficiency on the Iowa Electronic Market, Berg and Rietz study binary political and economic contracts and find persistent overpricing of low probability outcomes relative to realized frequencies. Even in markets explicitly designed to aggregate information efficiently, participants systematically pay up for tails, leaving near certain outcomes trading at a discount to their eventual payoff.

What’s striking is how stubborn this premium is. Just one example we can point to is Polymarket’s “Interstellar object is an alien” market. While several of our team members would be the first to chew your ear off about how it's likely we've already been visited by aliens or alien technology, there was no serious scientific dispute here, no new information coming, and no realistic path to a surprise resolution for this specific instance of a comet passing through our solar system.

And yet, the market still assigned meaningful value to the alien tail. As late as the end of November you could still buy the "NO" outcome for 97.5c in pretty decent size, giving you an effective 30% APR to an end of the year expiration. You could have executed the same trade for the "US recession in 2025" market as late as early December at an even higher effective APR to expiration.

The Liquidity Discount Near the Finish Line

As markets approach resolution, a second force starts to matter more. Once an outcome is basically decided, holding a contract becomes a patience test. You can wait for a deterministic payout, or you can sell now and free up capital for something more exciting. Most participants choose excitement. The result is a liquidity discount. Near certain contracts trade below terminal value not because the probability is wrong, but because people don’t want capital sitting idle.

This has nothing to do with forecasting errors. It’s about how people actually use prediction markets.

Terminal Value Arbitrage

Terminal value arbitrage is what happens when you’re willing to buy certainty from people who don’t want it. In prediction markets, two things consistently push prices in the same direction.

The gambler’s premium keeps long shot alternatives inflated, mechanically depressing the price of the dominant outcome. And liquidity preferences push near certain contracts below terminal value as resolution approaches.

Neither of these is irrational. They’re features of who participates in these markets and why. Together, they create a persistent gap between price and payoff for outcomes that are very likely to resolve one way.

This Isn’t About Being Smarter

Individually, these positions look unexciting. A few cents over a few weeks doesn’t appeal to most traders. But spread across a large number of independent markets, something changes. You’re not betting on headlines. You’re not trying to predict surprises. You’re letting time do the work while outcomes resolve. The portfolio starts to behave less like a speculative book and more like a carry strategy. We like this because the edge doesn’t depend on being clever in public. It depends on being comfortable owning positions that other people don’t want to hold.

Why This Exists Now

Prediction markets are finally big enough to matter and still weird enough to be inefficient.

There’s real liquidity, constant turnover, and a steady stream of participants optimizing for excitement rather than risk adjusted return. As long as that remains true, tails will stay expensive and certainty will stay cheap. It could be said that behavioral alpha is the only true persistent alpha, everything else is just a function of being able to compute better or faster.

A Very Old Trade in a New Wrapper

If this all sounds familiar, it should. Markets have always overpaid for convexity and underpaid for certainty. You see it in options, in credit, in merger arbitrage, anywhere patience is scarce. Prediction markets just make it unusually clean, and throw in the ability for anyone to indulge their specific beliefs, whether it be aliens or a policy some politician promised to enact.