Prediction markets are certainly in the news. Polymarket is probably 60% of the way to making it to the common American lexicon (as it has already with TPOT). Kalshi seems to be breaking new boundaries for how they're aggressively pushing novel markets into states that have traditionally been very resistant to this (i.e. New York, California).

Opponents have called it a sleight of hand - calling it sports betting but relabeled as financial markets - often bringing up stories of financial ruin, where ordinary people get addicted to betting and end up mired in debt.

Proponents argue that this is just an evolution of markets, pointing out that there was moral panic over futures markets, with similar accusations of gambling - but futures are now a hedging instrument.

As with everything, there's no binary answer here, and there's nuance - but I feel the take rate is a strong proxy for whether something should be considered gambling.

The take rate (otherwise known as the rake) is the amount that the venue will receive in exchange for doing the transaction. Sometimes this is explicit - Robinhood will take a fee for options trades, or implicit - I can expect to lose roughly ~0.01 of every dollar I put into Blackjack for instance (search up RTP for casino games).

It's easy to point to the expected value of the games and say that this is a good proxy for whether something is gambling - but I don't feel that's true.

With sports betting it's simple - the EV of playing the platform is the rake - there are no explicit fees. They want users to engage in highly -EV bets like parlays or encouraging problem gamblers to gamble more.

However with stocks, if a person buys an index fund or a meme stock on Robinhood, the expected value of both of these assets are wildly different, but the rake is the same (basically 0). In other words, there's no incentive for the platform for you to invest in the index fund or the meme stock.

This is what separates a 'poor investment' from gambling in my view - Robinhood doesn't care what you do in this case.

Let's go back to why this is relevant:

Polymarket doesn't charge any fees for using their platform whereas Kalshi does (1% settlement fee).

This results in very different outcomes - despite them being nominally the same concept.

Since Kalshi takes fees, it creates incentives

• the high take rate enables them to things like VIP programs, where they offer whale spenders free credits, perks - whatever to keep them on the platform, since the fees they generate from this are so good.
• If you look at stories where this happens - this is where problem gambling comes from, the egging on of behavior from VIP managers who dangle rewards for continuing a ruinous habit.

That's not to say it's all bad - the benefit of being P2P is that your book isn't focused on maximizing revenue from these money losing whales - so it's still a net win over existing sports book. They in theory shouldn't care whether consumers are winning or losing a lot of money - as long as they're generating fees from it.

Parlays is a telling example of how this is a benefit.

A traditional sportsbook would hand over first prefer users to do parlays over normal bets - since parlays are more -EV for users. In the case of prediction markets, parlays are normal markets but with different resolution criteria - the fee is the same. While they might push parlays since they're so popular, it ultimately doesn't really matter to Kalshi that users pick one option or the other.

In the case of Polymarket - they don't take fees, so it's economically inefficient for them to.

If they set a very low fee - they simply wouldn't have the economics to support a VIP program where they could encourage problem gambling, or mainstream ads that would promote gambling.

Instead, the incentives would point to them pushing more volume via fees and onboarding more mature counterparties - in other words maturing as a market, alongside options or futures.

A wise man once said show me the incentives and I'll show you the outcomes - and I think take rate is a good way of understanding these incentives for prediction markets.