Polymarket market making bots: earning the spread

A Polymarket market making bot tries to earn money in a very different way from a directional bot. Instead of betting which side wins, it posts orders on both sides of a market, buys a little below fair value and sells a little above, and hopes to pocket the difference, the spread, over many small trades. It sounds almost mechanical, and in calm conditions it can feel that way. But market making is not free money. You are being paid to provide liquidity precisely because it carries real risk, and understanding that risk is the whole game.

Where the spread actually comes from

Imagine a market trading around 60 cents. A market maker might bid 59 and offer 61. If a buyer takes your offer and later a seller hits your bid, you have bought at 59 and sold at 61, keeping two cents while never taking a strong directional view. Repeat that hundreds of times and the small edges add up. The catch is that you only capture the spread if flow is roughly balanced. When it is not, you accumulate inventory on the wrong side, and that is where the losses live.

Adverse selection is the real enemy

The traders most eager to take your quote are often the ones who know something you do not. If news breaks and the true probability jumps, informed traders will lift your stale offer before you can pull it, leaving you short exactly when you should not be. This is adverse selection, and it is the central difficulty of market making. A good bot manages it by quoting wider when uncertainty is high, cancelling fast when the market moves, and never leaving orders sitting through obvious events. A naive bot that quotes a fixed spread all day will be picked off.

Inventory and skew

Every fill leaves you holding a position you did not necessarily want. If you keep buying the yes side, your risk piles up on yes. Serious market making bots track inventory continuously and skew their quotes to unwind it: hold too much yes and the bot lowers both its bid and offer to encourage selling and discourage more buying. The goal is to stay near flat, collecting spread without building a large directional bet by accident. Manage this badly and a market maker quietly turns into a leveraged directional trader.

Why Polymarket is a hard place to do it

Prediction markets add wrinkles a stock market maker does not face. Many markets are thin, so a single trade moves the price and spreads must be wide to be safe. Markets resolve to a fixed outcome, so as resolution approaches, prices gap toward zero or one and holding inventory becomes dangerous. And on very short-term markets, the underlying is close to a coin flip, which means the fair value you are quoting around is genuinely uncertain. If you have not, read Polymarket trading bots explained for the broader context, and prediction market strategies for how market making sits alongside other approaches.

Honest expectations

Market making can produce steady small gains in the right conditions, but it is not passive income and it is not safe. A quiet market can hand you consistent spread for days and then a single event can wipe out a week of it in minutes if your cancellation logic is slow. It rewards low latency, tight risk controls, and constant monitoring. There is no configuration that guarantees profit, and you can lose your entire deposit, especially if a market moves against a position you failed to unwind.

If you want to try it

Start in paper mode, quote wide, and size tiny. Watch how your inventory drifts and how often you get picked off around news. A self-hosted tool like POLBOT lets you run this on your own machine with your own keys, with no custody of your funds and a paper-first workflow, so you can study the behavior before real money is involved. Treat the first weeks as tuition. If the strategy only looks good because you are ignoring adverse selection and the cost of exiting inventory, better to learn that on paper than with your bankroll.

A last word of perspective: market making is a craft, not a switch you flip. The people who make it work spend far more effort on cancelling stale quotes and controlling inventory than on the clever part that captures the spread. If you go in expecting a tidy, hands-off income stream, you will be surprised, and probably not pleasantly. If you go in expecting to babysit a fragile edge and defend it against smarter flow, you have the right mindset, and a much better chance of not blowing up on the first noisy day.

This article is informational, not financial advice. Trading prediction markets carries a risk of total loss. Check that using it is legal in your jurisdiction.