Edgebook Research · October 2026

Most Polymarket traders lose to the price they pay

We looked at 119,708 resolved positions from 551 active traders. Only about one in three made money, and very few showed skill you can tell apart from luck.

Data collected 2026-10-26 from Polymarket's public data. Methodology at the bottom.

36%

of traders with 50+ resolved positions were profitable

4.7%

showed edge strong enough to rule out luck

19%

less often than priced: how much contracts bought under 10¢ underperformed

50%

of traders with a 60%+ win rate still lost money

1. Long shots are overpriced. Favorites are slightly underpriced.

If the market price were a perfect probability, contracts bought at 30¢ would win 30% of the time. Across our sample, cheap contracts won less often than their price implied, and expensive ones won slightly more often.

Contracts bought under 10¢ were priced to win 4.3% of the time. They won 3.4%. That gap is the classic favorite–longshot bias, also seen in horse racing and sports betting: people overpay for a small chance at a big payout.

Each bar is a price range: how often those contracts actually won, minus how often their price said they would, in percentage points. Lines show the 95% range. Hover for details.

Show the numbers

Bought at	Positions	Price implied	Actually won
0–5¢	6,542	2.4%	1.7%
5–10¢	4,058	7.3%	6.2%
10–20¢	7,333	14.8%	12.2%
20–30¢	9,196	25.0%	20.7%
30–40¢	11,155	34.7%	31.0%
40–50¢	15,201	45.0%	42.4%
50–60¢	19,782	53.9%	53.1%
60–70¢	11,675	64.8%	66.0%
70–80¢	10,067	74.7%	76.4%
80–90¢	9,800	85.0%	86.0%
90–95¢	5,586	92.5%	93.5%
95–100¢	9,313	97.7%	98.0%

2. Very few traders can show real edge

For each trader with at least 50 resolved positions (384 traders, median 299 positions each), we measured how much more often their picks won than the prices they paid implied.

39% came out ahead of the price. But only 4.7% were far enough ahead that luck is an unlikely explanation. By chance alone you'd expect about 2.5% to clear that bar. Meanwhile 10.7% were reliably behind the price.

Number of traders by edge, in percentage points. Edge = how much more often their picks won than their entry prices implied, before fees.

3. Win rate lies

A high win rate feels like skill. It isn't. Buying expensive contracts wins most of the time, but one 90¢ loss wipes out the profit from nine 90¢ wins. Of the 103 traders with a win rate of 60% or more, 50% lost money. And 27% of traders who won under 40% of the time were profitable.

The only test that matters is whether your picks win more often than the price you paid says they should.

4. Where traders overpay

Grouped by market type, most categories are too noisy to call. Two stand out with ranges that don't include zero: soccer (−5.1 pts) and crypto (−1.6 pts). Most crypto positions in the sample were in short "Up or Down" markets.

Average edge by category, weighted by shares, with 95% ranges from resampling traders. Bars crossing zero mean we can't tell the direction.

Are you good, or lucky?

Edgebook runs this same analysis on your own wallet: your edge over the price, by category, with a chart like the one above. Free, read-only, no sign-in.

Check your wallet

Methodology

Sample. 600 wallets drawn at random from accounts trading on Polymarket on 2026-10-26: half from the live trade feed, half from trades in the 80 highest-volume open events. That's a sample of active traders, not all Polymarket users, and it leans toward frequent traders and bots.
Positions. We used each wallet's finished positions: sold, redeemed, or resolved but never redeemed. Polymarket's "closed positions" feed mostly leaves out losses nobody bothered to redeem, so we added resolved positions from the open-positions feed. Skipping that step makes traders look far better than they are.
Limits. For heavy traders we used the most recent 300 finished positions, plus unredeemed ones from the same period. 30 wallets with very large unredeemed lists were left out because we couldn't see them completely. Results restricted to wallets with full histories point the same way.
Edge. A resolved position counts as a win if its side paid out. Edge compares that against the average entry price. Positions sold before resolution count in profit figures but not in edge. Fees are not included, so real results are slightly worse.
Statistics. Price-range ranges are Wilson 95% intervals. Category ranges come from resampling traders 400 times. A trader "rules out luck" when their average edge is more than 1.96 standard errors above zero.
Categories are assigned by keywords in the market title and URL, so a few markets will be misfiled.
We publish only aggregate results. No individual wallet is named.