Odds & probability at this price
Kalshi fee across the price curve
Fee = 0.07 × contracts × price × (1−price). It's a parabola — highest at 50¢, near-zero at the extremes. This table holds your contract count fixed and walks the price.
| Price | Fee (100 contracts) | Fee % |
|---|
Contract prices are probabilities with a spread
Every prediction-market contract pays exactly $1.00 if the event happens, $0.00 if not. A price of 62¢ isn't a bet size — it's the market saying "62% likely," and it's what you pay per contract to be on that side. Your edge only exists if your own probability estimate diverges meaningfully from the market's, after fees are subtracted. Compare your read to the market's on the breakeven win-rate calculator for the same fee-adjusted-edge logic applied to leveraged trading.
Kalshi vs Polymarket: the fee gap actually matters
Kalshi is a CFTC-regulated exchange and charges a curved trading fee on most contracts — roughly 7% of price×(1−price), which sounds small but bites hardest exactly where markets are most uncertain (near 50¢), and light near the edges (5¢ or 95¢). A $1,000 stake at 50¢ loses roughly $17.50 to fees round-trip on Kalshi; the same stake at 5¢ loses under $4. Polymarket, an offshore on-chain exchange settled in USDC on Polygon, currently charges 0% trading fee on most markets — the cost there is gas (cents) and the bid-ask spread you cross, not a published formula.
That fee asymmetry changes strategy. On Kalshi, near-coin-flip markets (40-60¢) are the expensive ones to trade in and out of — better suited to holding to resolution than active scalping. On Polymarket the fee curve doesn't exist, so spread and liquidity depth matter more than price level. Either way, the number that actually predicts your long-run result isn't the contract price, it's your breakeven probability after costs — and that's always worse than the sticker price, never better.
The other trap: contract price answers "what does the market believe," not "what should I believe." A 90¢ Yes contract implies 90% — reasonable-sounding, but if your own model puts it at 80%, buying at 90¢ is negative expected value even though it "usually wins." Prediction markets punish confident agreement with consensus and reward correctly-priced disagreement, same as any options or futures market.