Short version: market making on prediction markets is genuinely available to retail, because thin books mean wide spreads. It is also the fastest way to lose money if you quote without a fair value, because every fill you get is disproportionately likely to be the one you didn't want.
This is strategy four of ten in Algorithmic Trading Strategies, Translated for Prediction Markets, and it is the one where the gap between "sounds sophisticated" and "is profitable" is widest.
The equities version
A market maker quotes both sides of an instrument, earns the bid-ask spread when both sides fill, and manages the inventory that accumulates when they don't. Their edge is not a view on direction. It is being paid for providing immediacy to people who want it now.
Why it transfers unusually well
Two structural features of event contracts favour a patient retail maker.
The spreads are wide. Outside the flagship series, you are frequently one of a handful of resting orders. A spread that would be unthinkable on a liquid equity is normal here, and that spread is the raw material.
The instrument is bounded. Your worst case on any single contract is knowable before you quote. There is no gap-down-eighty-percent scenario. That makes position sizing arithmetic rather than guesswork — run it through Kelly and cap it per market.
The two things you need before you quote
One: a fair value you trust. This is not optional and it is not negotiable. Our Edge models publish a derived model probability for the commodity and crypto series — Silver Edge, Gold Edge, Oil Edge, Bitcoin Edge — each built by comparing an options-implied probability for the underlying against the contract's own price. Where you have that, you have something to quote around. Where you don't, you need to build a fair value from base rates and maintain it, or stay out of that market.
Two: your true cost. The Kalshi Fee Calculator tells you what a round trip costs. A maker quoting a three-cent spread with a two-cent round-trip cost is working very hard for one cent, and one cent does not survive a single adverse fill.
Then point all of this at the markets that have drifted far enough from a fair value you trust to be worth quoting into. Our Mispricing Scanner did that triage automatically and is paused while it is rebuilt for the 2026 midterms; the method is in how the mispricing scanner worked, and the Ladder Arb Scanner flags Kalshi spread and total ladders with books wide enough to make a market in.
Adverse selection is the whole game
Here is the lesson that costs the most to learn firsthand.
Your resting order does not fill randomly. It fills when someone wants to trade against it, and disproportionately often that someone knows something you don't — a starter is out, a filing dropped, a forecast was revised. They see it, they hit your stale quote, and you are now long the wrong side of a market that has already moved.
In equities this is diluted by enormous continuous flow, much of it uninformed. In event contracts the information is discrete and lumpy, so the ratio of informed to uninformed flow hitting a stale quote is far worse. A maker without a live fair value is not earning a spread. They are writing free options to better-informed people. That sentence is the entire risk of this strategy.
How to actually run it
- Pick markets where you have a model, not markets that look wide. Width without a fair value is a trap, not an opportunity.
- Quote inside the spread, not at the touch. You want fills from people paying for immediacy, not from people picking you off.
- Cap inventory per market before you start. Write the number down. Binary resolution concentrates a one-sided position into a single outcome.
- Pull quotes around known catalysts. Scheduled releases, injury reports, filings, kickoff. This is when informed flow arrives and your stale quote is worth the most to somebody else.
- Widen as expiry approaches. Time sensitivity accelerates near settlement — check the Theta Edge Calculator for the shape of it.
- Track fill quality, not fill count. The metric that matters is what the market did in the ten minutes after each fill. If it consistently moved against you, your quotes are stale and no amount of volume fixes that.
Where it goes wrong
Quoting around the market's own mid. If your fair value is derived from the price you are quoting against, you are not making a market — you are amplifying the consensus and paying fees for the privilege. When the consensus is wrong you will be maximally long the wrong side, with no independent signal to tell you.
Next
If you want the short-cycle version of this, read scalping the 15-minute series. If the fair-value question is what interests you, read volatility trading with binary event contracts — that is where the Edge models come from. And TWAP execution is the taker-side discipline that makes every other strategy cheaper.
Trade responsibly, and never quote a market you cannot price.