Prediction Market Win Rate: Why It Misleads and What to Track Instead
Prediction market win rate is the wrong metric for 91% of traders. 4 worked examples show why a 40% win rate prints money and 70% loses it.
What Win Rate Means on Prediction Markets
Your prediction market win rate is the percentage of contracts that resolve in your favor. Buy 100 Yes contracts across different markets, 55 resolve Yes, your win rate is 55%. Simple division.
But win rate on prediction markets is fundamentally different from win rate on sportsbooks. On a sportsbook, odds are fixed at the time of your bet. A -110 line always pays the same. On prediction markets, you buy contracts at variable prices between $0.01 and $0.99. A contract purchased at $0.20 pays out $0.80 in profit. A contract purchased at $0.85 pays out $0.15. The price you pay determines everything about the math.
This variability makes win rate almost useless as a standalone metric. Two traders with identical 50% win rates can have wildly different P&L depending on the prices they paid. Run your actual trades through the PM EV calculator to see what matters: net expected value per contract, not the ratio of wins to losses.
Why a 40% Win Rate Can Be Highly Profitable
The relationship between win rate and profitability depends entirely on the average price of your winning and losing contracts.
Worked example: Profitable at 40% win rate.
You buy 100 contracts across various markets. Average purchase price: $0.30. This means you risk $0.30 per contract and profit $0.70 on each win.
| Metric | Value |
|---|---|
| Wins | 40 |
| Losses | 60 |
| Profit from wins | 40 x $0.70 = $28.00 |
| Loss from losses | 60 x $0.30 = $18.00 |
| Net P&L | +$10.00 |
A 40% win rate with an average entry at $0.30 produces a 33% return on capital. The math is clear: cheap contracts do not need to win often.
Worked example: Losing at 70% win rate.
Now flip it. You buy 100 contracts at an average price of $0.85. You risk $0.85 for $0.15 profit per win.
| Metric | Value |
|---|---|
| Wins | 70 |
| Losses | 30 |
| Profit from wins | 70 x $0.15 = $10.50 |
| Loss from losses | 30 x $0.85 = $25.50 |
| Net P&L | -$15.00 |
A 70% win rate and you still lost money. Expensive contracts need to win almost every time just to break even. The breakeven calculator shows the exact win rate you need at any contract price.
Contract Price and Required Win Rate
Every contract price has a breakeven win rate: the minimum percentage of wins needed to avoid losing money. Before fees, the formula is straightforward.
Breakeven Win Rate = Contract Price / $1.00 = Contract Price
A contract at $0.60 needs a 60% win rate to break even. At $0.30, you only need 30%. At $0.90, you need 90%.
This is the core insight most traders miss. Buying contracts at $0.80 and above means you need to be right 80%+ of the time. That is an extraordinarily high bar. Professional weather forecasters, the best probabilistic reasoners in the world, struggle to maintain 80% calibration on high-confidence predictions.
| Contract Price | Breakeven Win Rate (Pre-Fee) | Profit per Win | Loss per Loss |
|---|---|---|---|
| $0.20 | 20% | $0.80 | $0.20 |
| $0.40 | 40% | $0.60 | $0.40 |
| $0.50 | 50% | $0.50 | $0.50 |
| $0.70 | 70% | $0.30 | $0.70 |
| $0.85 | 85% | $0.15 | $0.85 |
| $0.95 | 95% | $0.05 | $0.95 |
The table reveals the asymmetry. A $0.20 contract has 4:1 reward-to-risk. A $0.85 contract has roughly 1:5.7 reward-to-risk. This is why expected value, not win rate, determines profitability.
How Fees Push the Breakeven Win Rate Higher
Platform fees make the math worse. Every winning trade loses a percentage to the platform, which means you need to win more often to compensate.
Kalshi example: $0.40 contract with 7% winner fee.
Without fees, breakeven is 40%. With fees, your profit on a win drops from $0.60 to $0.60 x 0.93 = $0.558.
Fee-adjusted breakeven = Cost / (Cost + Net Profit) = $0.40 / ($0.40 + $0.558) = $0.40 / $0.958 = 41.8%
Fees pushed the breakeven from 40.0% to 41.8%. That 1.8 percentage point increase sounds small. Over hundreds of trades, it erodes real money from your account.
The damage is worse on expensive contracts. For a $0.80 contract on Kalshi:
Pre-fee profit per win: $0.20. Post-fee profit: $0.20 x 0.93 = $0.186.
Fee-adjusted breakeven = $0.80 / ($0.80 + $0.186) = $0.80 / $0.986 = 81.1%
You went from needing 80% to needing 81.1%. On cheap contracts, fees are a rounding error. On expensive contracts, fees can be the difference between a positive and negative edge.
On Polymarket, the fee structure is different (2% on net profits), but the effect is the same: every fee dollar raises your required win rate. Use the breakeven calculator with your specific platform to see the exact number.
Why 91% of Polymarket Traders Lose Money
Layerr's analysis of Polymarket wallets found that roughly 91% of traders lost money. This statistic is consistent with broader retail trading data. The reasons map directly to win rate math.
Reason 1: Bias toward high-probability contracts. Retail traders overwhelmingly buy contracts priced above $0.70. These feel "safe" because they win most of the time. But as the table above shows, the reward-to-risk ratio is terrible. A few losses at $0.85 wipe out dozens of small wins at $0.15 profit each.
Reason 2: Ignoring fees on thin margins. A contract at $0.85 with an 87% true probability has a raw edge of 2%. After Kalshi's 7% winner fee, that edge shrinks to about 0.7%. After Polymarket's spread costs, similar story. Most retail traders do not run this calculation. They see 87% and think "that is probably going to happen" without checking whether the price already reflects that probability.
Reason 3: No framework for when NOT to trade. Profitable traders skip most contracts. A sharp prediction market trader might evaluate 50 contracts per week and trade 5. The other 45 did not have enough edge after fees. Retail traders trade on every opinion they hold. This is the difference between prediction market strategy and gambling.
Reason 4: Win rate tunnel vision. Tracking win rate instead of EV per contract leads to the exact mistakes above: buying expensive contracts for high win rates while hemorrhaging money on the P&L.
Tracking Your Actual Edge vs Your Win Rate
Win rate tells you how often you are right. It says nothing about whether you are making money. The metric that matters is edge per dollar risked: your average net profit divided by your average cost per contract.
Worked example: Measuring real edge.
Over 200 resolved trades, your record shows:
| Stat | Value |
|---|---|
| Total trades | 200 |
| Wins | 96 (48% win rate) |
| Losses | 104 |
| Average buy price on wins | $0.38 |
| Average buy price on losses | $0.42 |
| Average profit per win (after fees) | $0.56 |
| Average loss per loss | $0.42 |
| Total profit | 96 x $0.56 = $53.76 |
| Total loss | 104 x $0.42 = $43.68 |
| Net P&L | +$10.08 |
A 48% win rate. Solidly below 50%. And you made money. Because your wins were at cheaper contracts (higher payout per win) and your losses were contained.
The real edge metric: $10.08 / (200 x average cost) = $10.08 / (200 x $0.40) = $10.08 / $80.00 = 12.6% return on capital.
Run your trade history through the edge calculator to test whether your results reflect genuine skill or variance. At 200 trades, you start to get meaningful signal. Below 100 trades, even a 60% win rate could be noise.
The Metrics That Actually Matter
Stop tracking win rate as your primary metric. Track these instead.
1. Expected value per contract. This is the single number that determines long-term profitability. Positive EV across your portfolio means you are making money over time. The PM EV calculator computes this for any individual trade.
2. Calibration. When you estimate 60% probability, do events happen 60% of the time? Plot your estimates against actual outcomes. If you consistently overestimate (saying 70% when the true rate is 55%), your edge is illusory. Calibration is what separates informed traders from confident ones.
3. Edge per dollar risked. Total net P&L divided by total capital deployed. This accounts for contract price variation, fee drag, and position sizing all at once. Target 5%+ per cycle of capital to justify the time and risk.
4. Closing price accuracy. Compare your purchase price to the final settlement. If you buy at $0.40 and the contract settles at $1.00, the closing market agreed with you. If you buy at $0.40 and it closes at $0.38 before resolving Yes, the market did not think much of your edge. This is the prediction market equivalent of closing line value.
The pipeline connects: estimate probability, calculate EV, size with Kelly, execute, then track calibration and edge per dollar over 200+ trades. Win rate is a byproduct of this system. It is never the objective.
Frequently asked questions
- What is a good win rate for prediction markets?
- There is no universal good win rate. A 35% win rate on cheap contracts ($0.20-$0.30) can be highly profitable. A 75% win rate on expensive contracts ($0.80+) can lose money. Track expected value per contract and edge per dollar risked instead.
- Why do most prediction market traders lose money?
- Roughly 91% of Polymarket traders lose money because they buy high-probability contracts with thin margins, ignore fees on those thin margins, and trade on opinion without calculating expected value. The reward-to-risk ratio on contracts above $0.80 is brutal.
- How do fees affect breakeven win rate?
- Fees raise the breakeven win rate on every contract. A $0.40 contract needs 40% pre-fee but about 41.8% after Kalshi's 7% winner fee. A $0.80 contract goes from 80% to 81.1%. Use the breakeven calculator with your platform's fee structure for exact numbers.
- How many trades do I need to evaluate my prediction market performance?
- At least 200 resolved trades. Below that sample size, variance dominates and even a 60% win rate could be random. After 200 trades, you can meaningfully assess calibration, edge per dollar, and whether your results reflect skill.
- Is prediction market win rate the same as accuracy?
- No. Win rate measures how often contracts resolve in your favor. Accuracy (calibration) measures how well your probability estimates match reality. You can have a low win rate and high accuracy if you correctly identified 30% events and bought them at $0.20. Both matter, but accuracy is the deeper skill.