Polymarket Fees Explained: Maker, Taker, and the CLOB Fee Schedule in 2026

The honest answer to "what are Polymarket's fees?" in 2026 is: it depends on the market, and the schedule has changed multiple times.

Short version: most Polymarket sports markets currently charge approximately 0% maker fee and 0-2% taker fee, with one important exception — NCAA men's basketball, which since February 18, 2026 charges 2% on payout per winning share rather than on volume. Polygon gas is paid by the protocol relayer, not by you per fill. The rest of this post is the detail behind those numbers and what they mean for your real breakeven edge.

Maker vs Taker on a CLOB, Briefly

Polymarket runs a Central Limit Order Book. If your order sits on the book and someone hits it, you are the maker. If your order crosses the spread and immediately executes against existing orders, you are the taker. Most exchanges charge takers more than makers because makers add depth to the book. Polymarket follows this pattern, but the absolute numbers have trended toward zero on most sports markets as the platform has competed for volume.

Polymarket's Actual Fee Schedule in 2026

This is where it gets specific. The "Polymarket fee" people think of as a single number is actually several different numbers depending on the market type.

Standard sports moneylines (NBA, NHL, MLB, NFL, soccer, tennis, esports): 0% maker fee, ~0% taker fee on most markets. This is the headline number. Polymarket has aggressively reduced fees on liquid sports markets to grow volume. Our internal backtest scripts for esports markets like CS2 and LoL set TAKER_FEE_C = 0.0 because that has been the reality on those markets for the relevant period.

NCAA men's basketball: A 2% fee on winnings was introduced on February 18, 2026. This is structured differently from a per-volume fee — it applies to the profit on a winning position rather than to every fill. If you buy at 60c and the contract resolves to $1, your gross profit is 40c per share and the fee is 0.02 × 100c = 2c per share, leaving net 38c. Losses are not fee-charged because there are no winnings to take a percentage of.

Neg-risk markets (multi-outcome events): Polymarket's neg-risk model is used for markets where multiple outcomes are mutually exclusive — for example, "which team wins the championship?" listed as separate YES/NO contracts per team. Fees on these markets vary and have been subject to repeated adjustment. The neg-risk adapter contract handles the cross-outcome math, and the fee structure has differed from standard markets. Check the current fee on the specific market before trading.

Politics and entertainment markets: Generally follow the standard model, but volume-tiered fee promotions have appeared and disappeared. Recent quarters have seen extended zero-fee periods on certain political markets to drive volume.

Per-side or per-round trip: Fees apply per fill, not per round trip. If you buy a contract at 60c and sell it at 65c, you pay the relevant fee twice — once on entry, once on exit. Hold-to-settlement strategies pay it once. This is one structural reason hold-to-settlement is preferred for low-edge strategies — you save half the fee load.

What About Polygon Gas?

Polymarket settles on Polygon, but you do not pay gas per fill — the protocol's relayer infrastructure batches and pays the on-chain cost. You sign trade authorizations off-chain (via your wallet) and the order matching and settlement happen on Polymarket's side without you submitting a transaction per trade.

What this means practically: when you calculate your real per-trade cost, you can ignore gas. It is a real cost in the system, but it is not a cost you bear at the trade level the way a Uniswap user does.

How Fees Compare to Sportsbook Vig

This is the comparison that matters for anyone migrating from DraftKings or FanDuel.

A typical sportsbook moneyline is priced with 4-8% vig — both sides of the market sum to 104-108% implied probability, which is the bookmaker's edge baked into every price. Win or lose, you have paid the vig at the moment you bet.

Polymarket is peer-to-peer, so the two sides of a market sum to ~100% before any fees. On markets where the fee schedule is zero and the spread is tight, your structural cost is genuinely much smaller than a sportsbook's. The catch is that Polymarket books are thinner than sportsbook books on many markets, and the effective cost you pay through wider spreads can match or exceed sportsbook vig. A market priced 53c bid / 58c ask has a 5c spread, half of which you eat crossing — that is functionally a 2.5% taker cost even if the fee schedule says zero.

The honest comparison: on liquid Polymarket sports markets, you pay materially less than on a sportsbook. On thin Polymarket markets, you can pay as much or more once spread is included. Always check the book depth before trusting the headline fee number.

Calculating Your Real Breakeven Edge

Here is the formula that matters:

breakeven_edge_c = taker_fee_c + slippage_c + latency_penalty_c

For a typical liquid Polymarket sports market in 2026: - taker_fee_c ≈ 0c (most markets) to 2c (some markets) - slippage_c ≈ 1-3c (the half-spread you pay crossing the book) - latency_penalty_c ≈ 0.5-2c (price drift between your decision and your fill)

Realistic breakeven edge for a hold-to-settlement strategy: 2-5 cents. Below this and the costs eat your edge in expectation.

For a strategy that sells before settlement (round-trip), double the fee portion: you pay it on entry and again on exit. Breakeven jumps to 4-8 cents, which is why we maintain a hold-to-settlement preference on most strategies.

Our own bots use these constants directly in their backtests. For example, the CS2 backtest assumes TAKER_FEE_C = 0.0 because that is the current reality on the LoL/CS2 markets, while the moneyline backtest assumes TAKER_FEE_C = 2.0 for sports markets where the fee has historically applied. When fees change, we update the constant and rerun the backtest — that is the only way to know whether a previously profitable strategy is still profitable.

The NCAAMB Fee on Winnings: What It Actually Costs Per Trade

The 2% NCAAMB winnings fee deserves attention because it changes the math in a non-obvious way.

The mechanics: when a contract you hold resolves to $1, you pay 2% of the payout — 0.02 × 100c = 2c per winning share. Losing shares pay no fee because there are no winnings to take a percentage of. The fee scales with payout, not with your stake.

That means the fee as a percentage of profit depends entirely on your entry price. Worked examples on a contract that resolves to $1:

The lower the entry price (the bigger underdog you bought), the smaller the fee bite as a fraction of your profit. The closer to even or to favorite, the bigger.

But the number that actually matters for strategy design is expected fee per trade, not fee per winning trade. If your strategy wins 70% of the time, you only pay the 2c fee on those 70% of trades — expected fee per trade is 0.70 × 2c = 1.4c. If your gross edge is 3 cents per trade, the NCAAMB fee alone consumes nearly half of it before slippage even enters the math. That is why our moneyline bot carries a wider edge threshold on NCAAMB markets than on NBA. The February 18, 2026 fee change is a real strategy input, not a footnote.

Maker Rebates

Some crypto exchanges pay makers a small rebate for posting orders that get filled. We are not aware of a current, durable maker rebate program on Polymarket sports markets. If one exists when you read this, treat it as a tailwind rather than core edge — rebate programs have started and ended multiple times.

Selling vs Holding to Settlement

Fees are charged per fill, not per round trip. A buy-and-hold strategy pays the fee once on entry; the contract resolves to $0 or $1 mechanically without you submitting an exit. A buy-low-sell-high strategy pays twice — entry and exit — which can double the fee load on the same gross P&L. This is one of the main reasons we hold to settlement instead of selling on most strategies.

What to Watch

Polymarket fees have moved repeatedly. The NCAAMB winnings fee was added Feb 18, 2026. Neg-risk markets (multi-outcome events with mutually exclusive contracts per outcome) have seen multiple fee adjustments. Sports markets have trended toward zero. Treat any fee number in a backtest as something to revisit before you redeploy — pull the current schedule from Polymarket's official fee documentation and update your constants.

The Practical Takeaway

There is no single "Polymarket fee." There is a fee schedule that varies by market, plus slippage that is often larger than the explicit fee, plus latency cost that is usually smaller but real. The number that should drive your strategy decisions is taker_fee + slippage + latency, calculated for the specific market you trade, with the current fee schedule, then translated into expected fee per trade given your win rate.

On most liquid sports markets, that total still comes in materially below sportsbook vig. On thin markets, it does not. Knowing which side of the line your specific market sits on is the actual edge.

Related deeper reads: - The Complete Guide to Prediction Market APIs — fees in the context of the broader API landscape. - Hold to Settlement, Never Sell — why fee math pushes most strategies toward holding. - How We Profit on Polymarket — edge-after-costs framing on our own trade history.