In Defense of Play Money Markets

markets for experimentation

Every now and then there is a debate between generally the forecasting community (Metaculus, Manifold, etc) and the prediction market community (Polymarket, Kalshi, etc.) that play money markets are not valuable in that forecasters are not willing to put real money on the line.

Yes comparing two identical, high volume, non “out of bounds markets” real money markets are better than play money markets. However play money markets have some really good (underrated) benefits.

• Being a discovery mechanism for new real money markets

• Listing “out if bounds” markets like assassination markets / death markets, pope markets (listed on Polymarket but not Kalshi)

• Provide more value information in longer tail markets (50 unique forecasters is better “wisdom of crowds” than 3 traders betting a total of $2k)

• Benefit from not being able to fall down the spectrum into gambling

• Can approach significantly longer term markets 10, 20, 30 years which real money markets cannot reasonably

They Can Go Where Real Money Can't

Probably the biggest advantage? Play money markets can touch all the forbidden topics that real money platforms absolutely cannot deal with:

• "Will anyone attempt to assassinate Elon Musk in 2025?"

• “Will Milei survive to the end of his first term as president?”

• “From which government will the next assassinated head of state be from?”

It's not that real money platforms don't want to offer these markets - they usually can’t! Each platform has these weird regulatory blindspots, but play money markets can cover basically everything. There are of course moral grounds for potentially not trading or creating these markets, but you cannot deny that they do not produce at least some valuable information.

Money != Better Predictions

Here's a situation that happens all the time:

• On a play money platform, you get like 50 different people making predictions on some niche topic

• On a real money platform, there might only be 3 traders who put down a total of $2,000

Which one do you think actually gives better information / predictions? Often it's the play money one! Diverse viewpoints beats having a few people with money on the line.

Which market do you think is more informative

• Will there be any reported human-to-human transmission of highly pathogenic avian influenza H5N1 globally before 2026? Via Metaculus, play money but with >125 unique participants, >30 thoughtful / insightful comments that are well researched

• Will there be human-to-human transmission of the bird flu? Via Kalshi, real money but only $25k in all time volume, ~$150k in open interest, 3 meaningless comments, and I would estimate less than 20 unique traders

Falling down the slippery slope of gambling

Real money prediction markets are constantly fighting not to just become gambling sites. This creates all these weird pressures:

1. They need questions that get people excited to trade, not necessarily the most impactful / informative ones (note the limited markets and activity on AI risk and Health)

2. They have to focus on short-term stuff so people don't have to wait forever to get paid. Kalshi has now started paying interest on all cash and open positions and its likely that Polymarket will (hopefully) soon have a yield bearing stable. But getting paid only ~3%/yr on a 10 year market isn’t to attractive.

3. They can't afford to keep running markets that nobody's trading on. It does not look good to look at a Polymarket for information, see there is only $750 in volume this both makes you not want to trade it and not trust the information it provides.

Play money markets don't have these problems. They can just focus on what predictions would be most useful, not what will make money. They can keep a market open even if only a few super-knowledgeable people are participating.

The Super Long-Term Stuff

This is the advantage nobody talks about enough. Play money markets can handle questions that won't resolve for decades:

• "Before 2025, will an AI system be shut down due to exhibiting power-seeking behavior?"

• “What will the population living on Mars be in 2050?”

• “What will the world population be in the following years?”

Try running a real-money market on a 30-year question! It's basically impossible. The platform probably won't even exist that long. Play money platforms don't have to worry about any of that. You can get decent estimates for long term markets today.

They Work Together

The smartest way to think about this isn't "real money vs. play money" but how they actually complement each other

• Play money markets come up with new ideas and cover weird topics

• Real money markets give super strong incentives for accuracy when the conditions are right

• Both types give us different kinds of useful information

Play money markets can generally act as a decent discovery mechanism for real money markets.

Polymarket’s and Kalshi’s market listing teams have been working overtime lately and as their respective platforms grow will reach a point where they will need to sacrifice the quality of their markets. Both Kalshi and Polymarket have been thinking about this, Kalshi now has a market builder as part of their platform and Polymarket is likely working through something similar.

I wrote previously that it could be interesting for Polymarket to offer something like a pump.fun derivative in which anyone can create any market, a AMM bonding curve is seeded, and if the market reaches a certain volume/open interest/unique number of traders it “graduates” onto the main Polymarket platform (and moved into an orderbook, which aligns with AMMs and Orderbooks; The future is hybrid).

Manifold for instance started is a play money market and then tried its hand at adding in real money markets via sweepstakes (note, this is now being shut down likely do to limited traction and also crackdown on sweepstakes law). While this was live it was pretty cool there would be various markets that had both play and real money markets, play money markets sometimes would get good enough traction and then transition into a real money market. I think this was the first time that a real money and play money market existed together on the same platform. More people should try this (if you are please send me a note)!

---

Overall I really like play money markets as they allow for differentiated markets that are very hard to offer on real money platforms and serve as a great discovery / scaling mechanism.

Kalshi, Polymarket, and others, please consider pairing your platform with a play money market!