Some Thoughts on Prediction Markets

Where Insiders Rule the Day

A few years after college, I used to experiment with PredictIt, which was one of the first legally authorized prediction markets.

There wasn't much liquidity (individuals could put on something like only $800 per trade), but they intrigued me. For someone who was, and remains, obsessed with the news, it was an amazing way to make some lunch money

Now, however, prediction markets have entered hyperspace. Ever since groundbreaking litigation several years ago, you can't go far without seeing ads for prediction markets like Kalshi or Polymarket. Some of the markets are more "traditional" (what will the Federal Reserve do at its next meeting?) while others are much more "exotic" (how many times will President Trump say "Democrats" in the State of the Union address?).

Granted, these platforms are yet another example of how our society has come to embrace the wager. When Kalshi and Polymarket offer "predictions" on sports scores, it is difficult to tell the difference between "investing" and gambling.

At the same time, it has been fascinating to watch the rise of prediction markets. One of the things that has intrigued me has been the substantial trend following of certain "whales" that are suspected to have insider information. Insider trading in equities is obviously illegal, yet it seems to be a feature, not a bug, of prediction markets.

In fact, this happened a few weeks ago. A substantial amount of money came in from a single account predicting that the U.S. would attack Iran that very night. Many assumed that this was an insider who was in the room as the decision was being made. Consequently, they followed that investor into the same contract. As you know, that attack never happened.

While it didn't work in that case, there have been plenty of instances where a single account places massive bets on some of the more esoteric markets available. You have to wonder how much these accounts actually know…

It's safe to say that there are certainly insiders trading on Polymarket and Kalshi. Enforcement is lax. But as the management of these companies could argue: is this necessarily a bad thing? If prediction markets are in the business of finding truth, isn't it better if that truth gets to the surface much more quickly?

As a former practicing attorney and someone who focuses purely on equities, I have issues with this line of thinking. Even the most strident believers in free markets would argue that there needs to be at least some guardrails to make markets fair for everyone.

On social media, you can see plenty of people using AI to create prediction market trading bots. The goal, among other things, is to programmatically find some sort of edge (perhaps following insiders) and then exploiting that edge. Some of them are successful. Most aren't. But much of it is about identifying unusual trades and following those who are suspected of being insiders.

Notwithstanding my ethical qualms over potential insider trading, I'm paying more attention to these prediction markets. While I can count the number of trades I've made on one hand, they do provide an alternative source of signal in a world that is filled with more noise. They can also provide unique hedging opportunities if you are worried about idiosyncratic event risk.

If you're a heavy user (or even dabbled) with prediction markets, let me know! I'm curious how you use them as you think about macro risks.

Prompt of the Week

I try to point out that, as humans, it's critical to challenge these AI models. After all, even though it seems like we are, at times, speaking with sentient beings, they were originally trained on our data. They are reflections of us, yet they aren't exactly like us.

Because of this, I thought it would be interesting to ask the model to be introspective. I wanted it to know what it thinks its biggest biases are and how I can protect myself from them. It was an interesting exercise and I recommend you try it yourself!

"You are trained on human data. You are a reflection of us. Consequently, like us, you have biases when you speak to users like me. What biases do you most often use when communicating with me? Are biases blocking you from understanding your true biases? I need you to be extremely introspective."

Until next week,

Adam