i'm hugely bullish on prediction markets (the first co I covered @polychain in 2020 was @Polymarket) but as they get more popular the question remains: how does liquidity scale? compared to other markets, books on Kalshi/PM are still pretty shallow it's hard to MM a prediction market because (1) it's pretty hard to hedge most markets. At the time of writing this, the front page market on Kalshi is about what Trump will say in his meeting with South Korea --> if you're looking to make the book here, it's really hard to do that without taking on a directional view on the market. (2) prediction markets have their own versions of theta decay. as more information comes out closer to the resolution of the event, prediction markets have really weird tail behavior. Let's take a market like "Will the President be impeached by EOY" Everyday that the president is not impeached, absent news, the value of "Yes" shares decays, and that rate of decay only hyperaccelerates closer to EOY unless a shock announcement comes. in assets like options, you can at least hedge with other instruments, but with PMs there's no spot/other instrument to hedge on, which means that making liquidity for someone who wants to buy "No" shares in this example means you have to take on the "Yes" position. of course, in practice most people will run some sort of mean reversion strat to make sure they're not holding inventory too long, but you still get left with weird time dynamics especially closer to market resolution arguably solving the liquidity issue will be almost as valuable as building the winning PM. the kicker is I don't think the OB model will be suitable for all types of PMs. flash markets for shorter term markets like what @kash_bot is doing won't have time to spin up an OB and are probably better off relying on market structure mechanisms built for the long-tail (like AMMs) either way, if you're looking into prediction markets, you should spend time tracking who's trying to solve the liquidity scale problem --> it's not as sexy, but it is needed for these to actually hit scale