AoA
Join
← Intel
The rake · 4 min read

The fee math nobody tells you

Maker vs taker, and why it's 4× more expensive to trade a coin-flip than a longshot.

Kalshi's taker fee is 7% × p × (1−p) per contract, where p is the price. It's a parabola: it peaks at 50/50 and drops to almost nothing at the extremes. Settlement is free — winners and losers pay nothing at resolution.

PriceTaker fee% of priceMaker fee
$0.10 / $0.90$0.0063~0.7–6.3%~$0.0016
$0.30 / $0.70$0.0147~2.1–4.9%~$0.0037
$0.50$0.01753.5%~$0.0044 (0.9%)

A resting limit order (maker) pays ~1.75%·p·(1−p) — about 75% cheaper than a taker order. That single choice is the biggest controllable lever you have.

Why the 50/50 zone is a trap

Worked example: a true-54%-chance bet at 50¢ is +8% EV gross. The taker fee eats 44% of it → +4.5% net. As a maker, only 11% is eaten. And if your read is off by just 2 points, the taker version flips negative. Most retail has no idea.

The platform profits when you churn

The market is zero-sum before fees and negative-sum after — this affects every participant, regardless of skill. The house wins on volume (~$1.5B annualized in early 2026), not on who's right. Churn is the product they sell.

The platform's incentive (churn) is exactly what kills the retail trader: round-trip fees, 50/50 trading, over-trading. An honest solver's advice is anti-rake by construction — be a maker, trade the extremes, hold to settlement, abstain by default.

Source · Adapted from our Prediction-Market Edges research · Kalshi fee schedule, Market Math, GWU exchange-economics paper

Informational decision-math education, not betting advice. 18+. SIM-safe · read-only odds. We show our calibration and our losers.