Read-only. Uses Kalshi's public history. Places no trades.
Early results from a 1-in-10 sample. Verdict rules unchanged.
Listing: …
Sample pass (1 in 10 events): …
Full pass: …
Category
Role
Price (¢)
Contracts
Distinct events
Events older
Events newer
Avg price paid
Actual win rate
Edge (¢)
Profit / $100 after fees
Older half
Newer half
Verdict
Loading…
How to read this
Every public trade becomes two records: the taker bought their side at that side's price; the maker bought the other side at 1 minus that price. Each record won or lost from how the market actually settled.
Avg price paid and actual win rate are weighted by contracts. Edge = win rate minus price, in cents per contract.
Profit / $100 = (payout − cost − fees) ÷ cost × 100. Taker fee per trade = 0.07 × contracts × price × (1 − price) × the series' fee multiplier, rounded up to the cent. Makers pay 0, except series that charge maker fees (0.0175 × …).
Older / newer half split the 180-day window by the market's close date (older 90 days vs newer 90 days).
GREEN only if: profit after fees is positive in both halves; at least 200 distinct events; and win rate minus its 95% margin of error (counted by events, not contracts) is still above the price paid. RED if profit is negative in both halves. Everything else is GRAY — not proven.
Combo (multivariate) markets are excluded. Voided and scalar markets are skipped. Markets with zero volume have no trades and add nothing.
Past results only. This is a measurement, not advice.