FuturesAI Polymarket Strategies: Which Survive Taker Fees

FuturesAI published a list of five Polymarket strategies on April 18, 2026, with edge sizes in cents. Four days later its Polymarket vs Kalshi comparison listed Polymarket trading fees as 0% maker and 0% taker. Polymarket’s documentation now shows a taker fee on most categories.

So the question for anyone reading those strategy pages today is simple: which of the five still clear the fee? This post does that arithmetic with the official formula, strategy by strategy. Check it on FuturesAI

Educational content, not investment advice. Prediction-market positions can lose the full stake, and leveraged futures can lose the full principal. Fee rates are from Polymarket’s documentation as read on the date shown and can change; the strategy edge figures are FuturesAI’s published estimates, not guarantees.

In This Comparison

• How is the Polymarket taker fee calculated?
• What does the fee cost at different prices?
• Which of FuturesAI’s five strategies survive the fee?
• Why does Yes/No arbitrage fail for a taker?
• How much does resolution arbitrage keep after the fee?
• The Fee-Check Routine

How is the Polymarket taker fee calculated?

Polymarket charges takers a fee of C × feeRate × p × (1 − p) in USDC, where C is the number of shares and p is the share price, and makers are never charged, according to Polymarket’s fee documentation read November 1, 2026.

The rate depends on the market category. The documentation’s table lists crypto at 0.07, sports, economics, culture, weather, combos and “other” at 0.05, and finance, politics, mentions and tech at 0.04. Geopolitics and world-events markets carry no fee. Fees are applied at match time, rounded to five decimals, and the page says part of them funds daily maker rebates.

The p × (1 − p) term is the whole story. It peaks at a 50 cent price and falls toward zero at either extreme. A trade at 30 cents costs the same as one at 70 cents.

Polymarket’s own example: 100 crypto shares at $0.30 cost 100 × 0.07 × 0.30 × 0.70 = $1.47.

What does the fee cost at different prices?

Per share, the Polymarket taker fee peaks at 1.75 cents on crypto markets, 1.25 cents on 0.05-rate categories and 1.00 cent on 0.04-rate categories at a 50 cent price, and falls below 0.15 cents for any category at a 98 cent price, using the documented formula.

The table is my arithmetic from the formula and the documented rates. Values are cents per share for one taker fill. A round trip entered and exited by taking costs twice the figure; a position held to resolution pays it once, since resolution is not a trade.

Notice the shape. The fee is a mid-price tax. Strategies that live near 50 cents pay the most; strategies that live near 98 cents barely notice it.

Which of FuturesAI’s five strategies survive the fee?

Of the five Polymarket strategies FuturesAI published on April 18, 2026, resolution arbitrage, long-tail No baskets and cross-market arbitrage still clear the documented taker fee, while Yes/No arbitrage and weekly volatility carry on crypto markets do not unless the trader posts limit orders and pays no fee as a maker.

The table applies the fee at the prices FuturesAI used in its own examples. Edge figures are FuturesAI’s. Fee figures are mine, from the formula. Where a strategy has two legs, both legs are charged.

Two of five fail outright. Two survive comfortably. One, cross-market arbitrage, survives only at the wider end of its own published range.

The geopolitics exception matters here. Polymarket’s documentation lists geopolitics and world-events markets as fee-free, so strategy 1 and strategy 2 still work as written on those markets, and on any market where you rest limit orders instead of crossing the spread.

Why does Yes/No arbitrage fail for a taker?

Yes/No arbitrage on Polymarket fails for a taker because the two legs sit near the middle of the price range, where the fee peaks: at FuturesAI’s example of Yes at 62 cents and No at 39 cents, the combined fee is 0.4735 × rate per pair, which is 1.9 cents at a 0.04 rate and 3.3 cents at the 0.07 crypto rate, against a published edge of 0.5 to 2 cents.

The arithmetic per pair: 0.62 × 0.38 = 0.2356 for the Yes leg and 0.39 × 0.61 = 0.2379 for the No leg, which sum to 0.4735. Multiply by the category rate and you have the fee in dollars per pair, or in cents per pair if you read the rate as a percentage.

When I compared that with the 0.5 to 2 cent edge FuturesAI quotes, I noticed that even the best case, a 2 cent gap on a 0.04-rate market, nets 0.1 cents before gas and before the second or two of latency the strategy depends on. On a crypto market the same trade loses 1.3 cents per pair at the best published edge.

Most people miss this because the fee looks small per share. Two legs, both near 50 cents, is the single most expensive shape the formula allows.

Here’s where it still works: as a maker. Post both sides as limit orders, let takers fill you, and the fee is zero under the documented rules. The strategy becomes a liquidity-provision strategy with the same edge and a very different fill rate. The order-sizing post covers the other half of that trade-off, the cost of crossing a thin book.

How much does resolution arbitrage keep after the fee?

Buying a decided Polymarket market at 98 cents to collect $1.00 two days later returns 2.04% per trade before fees, about 372% simple annualized, and after a 0.10 cent taker fee at a 0.05 rate the return is 1.94% per trade, about 353% annualized, using FuturesAI’s example and Polymarket’s formula.

The fee at 98 cents is tiny because p × (1 − p) is 0.98 × 0.02 = 0.0196. At the 0.05 rate that is 0.098 cents per share, rounded to 0.10. Cost per share becomes 98.10 cents, payout stays 100, gain is 1.90 cents on 98.10, or 1.937%. Multiply by 365 ÷ 2 and you get 353%.

That number still deserves the warning FuturesAI attaches to it: the edge exists because the market is waiting out a dispute window, and the page notes that UMA resolutions have produced controversial outcomes. The fee did not create that risk, and removing the fee would not remove it either.

The same logic flatters the long-tail No basket. At 92 cents the fee is 0.37 cents on a 0.05-rate market, about 0.4% of the purchase price, against a published 5 to 15% IRR. Fees are not what decides that strategy. Calibration is, which the quarter-Kelly post treats in more detail.

If you’ve ever wondered why experienced prediction-market traders drift toward the 90-plus-cent end of the book, the fee table is one quiet reason.

The Fee-Check Routine

1. Read the market’s category on Polymarket and look up its rate in the documentation table. Geopolitics is zero.
2. Compute rate × p × (1 − p) in cents for each leg you will take. Add the legs.
3. Decide whether you exit by trading or by resolution. Trading pays the fee twice; resolution pays it once.
4. Subtract the total from the strategy’s expected edge. Below about half the edge, size normally; above, switch to limit orders.
5. Re-check the documentation before relying on any rate; the page carries no last-updated date.

Before you run any of the five strategies, check the current fee table and the current partner-exchange terms. Check it on FuturesAI

Tips

• Limit orders inside the spread make you the maker and the fee disappears under the documented rules.
• Strategies near 98 cents pay about one twentieth of the fee that strategies near 50 cents pay.
• Count both legs of a paired trade; the fee is per fill, not per strategy.

Warnings

• FuturesAI’s edge figures are estimates from its own backtests and examples, not live quotes.
• Polymarket’s fee page shows no effective date; rates and categories can change without notice.
• This is educational content, not investment advice; prediction-market and leveraged positions can lose the full amount committed.

FAQ

What is the Polymarket taker fee formula?

Polymarket’s documentation gives the taker fee as shares × feeRate × price × (1 − price), charged in USDC at match time, with rates of 0.04 to 0.07 by category and no fee for makers or for geopolitics markets.

Does Polymarket charge makers a fee?

No. Polymarket’s fee documentation states that makers are never charged, and that taker fees fund a daily maker rebate program.

Which Polymarket strategies still work after the taker fee?

Using FuturesAI’s published examples, resolution arbitrage at 98 cents, long-tail No baskets at 92 cents and cross-market arbitrage with gaps above about 2 cents still clear the fee, while Yes/No arbitrage and weekly crypto volatility carry only work as a maker.

Sources

• Polymarket Documentation: Fees (read November 1, 2026)
• FuturesAI Insights: Five Trading Strategies That Actually Work on Polymarket (April 18, 2026)
• FuturesAI Insights: Polymarket vs Kalshi, the Definitive Comparison Guide (April 22, 2026)

Fact-checked against the cited pages as of November 1, 2026. Not investment advice.