All articles
Polymarket·May 16, 2026·5 min read

Is It Worth Being a Polymarket Liquidity Provider? Fee Rebate Analysis

A quantitative look at Polymarket's maker rebate structure, typical fill rates across high- and low-volume markets, and net P&L scenarios that determine whether passive market-making is profitable after gas costs.

Polymarket's CLOB (Central Limit Order Book) pays maker rebates rather than charging maker fees — which sounds immediately attractive. The nuance is that "maker" on Polymarket means something specific, gas is not free, and fill rates on illiquid markets can crater your capital efficiency far below what the rebate rate implies. Here is what the numbers actually look like in production.

How Polymarket's Fee Structure Works

Polymarket runs a hybrid architecture: off-chain order matching via a CLOB operated by a designated market maker, with settlement on Polygon (now migrating to Base for certain product lines). The fee schedule as of mid-2025 is:

  • Takers pay a 2% flat fee on the trade notional, capped at $1.00 per order on orders under $50 and scaling linearly above that.
  • Makers (limit orders that rest in the book and are filled by a taker) receive a 0.1% rebate on the filled notional, paid in USDC at settlement.
  • Orders under approximately $1 notional are not rebated — this is a soft floor that filters out micro-fill spam.

The 0.1% maker rebate is the starting number. Everything else in the P&L is a discount against it.

Fill Rate Reality Across Market Types

Fill rate is the percentage of your resting order that actually gets hit. On Polymarket this is market-category-dependent and often brutal on thin markets:

High-volume markets (major US elections, large macro events, top-20 crypto price markets): order books are tight, bid-ask spreads compress to 1–3 cents on a $0.50/$0.50 market, and aggressive takers clear the book fast. Maker fill rates in these markets can reach 60–85% of posted volume within a 24-hour window. If you post $10,000 notional across both sides of a balanced market at a 1-cent edge, expect $6,000–$8,500 filled. Rebate: ~$6–8.50. Looks small, but at scale and compounded daily, this is real USDC.

Mid-tier markets (state-level elections, mid-cap crypto, sports finals): books are thinner, spreads widen to 5–10 cents. Fill rates drop to 20–40%. The wider spread means your edge is larger per fill, but idle capital drags annualized returns significantly.

Tail markets (niche geopolitical, obscure sports, single-event yes/no): fill rates routinely fall below 5%. Resting an order here for a week, watching it not fill, and withdrawing is a common outcome. The opportunity cost of locked collateral in these markets almost certainly exceeds the rebate.

The practical conclusion: concentrate maker activity on the top 10–15 markets by 24-hour volume. Everything below that threshold requires explicit justification.

Gas and Infrastructure Costs

Polymarket's CLOB matching is off-chain, so placing and cancelling limit orders does not hit the chain. On-chain costs occur at two points: depositing USDC collateral and withdrawing USDC proceeds (both Polygon or Base transactions), and claim settlement when a market resolves.

Polygon gas costs are low — typically $0.001–$0.005 per transaction at normal network load. For a bot cycling capital across 10 active markets with weekly withdrawals, monthly gas is roughly $2–$10. This is not the cost that hurts you.

What hurts is failed transactions during resolution spikes. When a major market settles, Polygon can spike briefly, and your settlement transaction may cost 5–20x normal. Budget conservatively: assume $0.05–$0.20 per market resolution. Across 50 markets per month, that is $2.50–$10. Still manageable at volume.

The more meaningful infrastructure cost is the bot itself. If you are running a custom market-making bot, your relevant cost is latency — stale quotes that get picked off by informed traders represent adverse selection losses, not gas. More on that below.

Adverse Selection: The Real Risk

The largest cost in Polymarket market-making is not fees or gas — it is adverse selection. When news breaks (an election call, a resolution trigger), traders with faster information will hit your resting orders before you cancel them. You are now short a position in a resolved market at the wrong price.

In production, we instrument a "quote age" metric: the elapsed time since our last model update versus the time our order has rested at the current price. If quote age exceeds a threshold — typically 200–500ms on high-speed markets, 2–5 seconds on slower ones — we cancel and reprice. Keeping this discipline tight is what separates flat maker books from ones that slowly bleed.

For a bot posting on a binary market at $0.48/$0.52, the adverse-selection loss on a single bad fill in a resolved market is $0.48 per $1 notional — wiping roughly 480 maker rebates in one event. The math demands that your cancellation latency be low relative to how fast information propagates on your target markets.

Net P&L Scenarios

Let us put this together with round numbers:

Scenario Daily posted notional Fill rate Gross rebate Adverse selection (est.) Net daily P&L
Conservative, top markets only $50,000 60% $30 $5–10 $20–25
Aggressive, broad coverage $200,000 35% $70 $30–50 $20–40
Tail market heavy $50,000 10% $5 $10–20 -$5 to -$15

The numbers do not scale dramatically with capital unless your cancellation infrastructure scales with it. More posted notional on thin markets does not earn more — it just increases adverse selection exposure.

When the Rebate Is Actually Worth It

Passive market-making on Polymarket is profitable under a specific set of conditions: you are concentrated on high-volume markets, your quote latency is under 500ms, your cancellation logic is event-driven (not polling), and you are not trying to squeeze edge out of tail markets. The 0.1% rebate is thin enough that any operational sloppiness — slow cancels, over-broad market coverage, high withdrawal friction — erodes it entirely.

The operators who run this profitably are not passive in the traditional sense. They are running tight feedback loops between their pricing model, the external event stream, and the order management system. That infrastructure overhead is the real barrier to entry, not the rebate math.


If you want a production-ready Polymarket maker bot rather than building the latency and event-wiring from scratch, talk to us — we design, build, and operate these systems at scale.

Need a bot like this built?

We design, build and run trading bots on Solana, Hyperliquid and Polymarket.

Start a project
#Polymarket#liquidity#market-making#fee-rebates#prediction-markets#trading-bots