What Is Polymarket Reward Farming?

Polymarket reward farming is a market-making strategy in which liquidity providers place qualifying orders within Polymarket's designated reward spread and receive a share of the market's daily liquidity incentive pool. The income comes from the exchange's incentive programme rather than from predicting outcomes.

The mechanics of the payouts are covered in how Polymarket liquidity rewards work. This page is about the strategy built on top of them.

What it is not

• Not betting. A reward farmer quotes both sides of a market at once. Whether it resolves YES or NO is not the source of income and, when the two sides stay balanced, barely matters.
• Not classical market making. A traditional maker earns the spread: buy at the bid, sell at the ask, repeat. A reward farmer largely gives that up, quoting at minimum size where the reward score is best rather than where the spread is widest. The spread becomes a rounding error next to the incentive.
• Not arbitrage. Nothing here depends on a mispricing. The edge is a published subsidy, available to anyone who meets the conditions.
• Not risk-free, and not passive. Orders get filled, inventory accumulates, and unwinding it costs money.

The economics

A reward farmer's profit is a two-line equation:

net = liquidity rewards earned − cost of unwinding filled inventory

The first term is fairly predictable. It depends on how many markets you quote, how tight, how consistently, and how many other makers are splitting the same pools. It is largely an engineering and market selection problem.

The second term is where reward farmers actually fail. Every qualifying order is genuinely tradeable, so fills are a feature of the strategy rather than an accident. The question is what those fills cost, and that is governed by adverse selection: the trader most eager to take your resting order is disproportionately likely to know something you don't. Fills cluster around news, which is exactly when the resulting position is most expensive to exit.

There is one structural feature that makes the strategy viable. Reward scoring does not consider queue position, so an order resting behind a deep queue scores the same as one at the front while being far less likely to be filled. Choosing markets where you can rest safely behind depth is the difference between farming rewards and paying for the privilege of quoting.

Why it is nearly always automated

Three properties of the scoring make this a software problem:

1. Scoring is time-weighted. Income is roughly proportional to minutes spent resting in the band, so uptime beats cleverness. A bot that never sleeps outperforms a sharper human who does.
2. Reward pools are per market. Income scales with the number of markets quoted simultaneously. Dozens or hundreds of markets, each needing its own orders, is not manual work.
3. Quotes go stale. When the midpoint moves, orders fall out of the rewarded band and silently stop earning. Staying in-band means repricing continuously.

The publicly visible top of the programme reflects this. The nine highest-earning reward-farming wallets we measured on-chain collected $742,677 in thirty days, and their common signature is breadth and uptime rather than clever pricing: the wallets earning six figures a month providing liquidity, with their addresses published so the figures can be checked.

How automated systems approach it

Most reward-farming bots converge on the same shape, because the scoring rules push them there: rank markets by reward density, quote both sides at minimum size, reprice as the midpoint moves, hedge whatever gets filled, and rotate out of markets that get crowded. The component-level walkthrough is in how a Polymarket reward farming bot works.

ClydeScan Reward Farmer is our implementation, running on our own capital with its results published as measured. Over the most recent fifteen-day window it earned real reward income and still finished net negative once inventory costs were counted. That is not a footnote; it is the honest state of the strategy at our size, and the whole reason we publish both halves of the number.

Common questions