Polymarket vs Kalshi
The two largest real-money prediction market venues, compared with live data: what each one is trading right now, how their prices have actually scored against resolved outcomes, and where they currently disagree on the same question. ProbCast is an independent read-only intelligence layer — we aggregate and score both venues, we don't take positions on either.
Live footprint
Open markets ProbCast currently tracks per venue, with 24h traded volume
| Venue | Tracked open markets | 24h volume (tracked) | Markets w/ complete data |
|---|---|---|---|
| Polymarket | 129 | $13.1M | 100% |
| Kalshi | 120 | $2.7M | 0% |
Volume covers the markets ProbCast tracks (sports parlays excluded), not each venue's total exchange volume. Venues report different data fields (order books, liquidity, open interest), so per-market scores are not averaged across venues — a cross-venue average would measure data coverage, not venue quality.
Measured accuracy
Brier score of each venue's price 24 hours before resolution, on markets that have settled (lower = better)
| Venue | Graded forecasts | Brier | Directional accuracy |
|---|---|---|---|
| Polymarket | 509 | 0.023 | 97% |
| Kalshi | 103 | 0.023 | 99% |
Read Brier before directional accuracy: a market priced near 99% earns an easy “correct call” that says little about calibration quality, and per-venue samples are small and differ in composition. Methodology and the full track record — including reliability diagrams and how accuracy decays with lead time — live on the Accuracy page.
Where they disagree right now
Audited same-question pairs, both sides live, gap-ranked
No approved same-question pair has both sides live at this moment. Pairs activate automatically when both venues list the question — the current queue is on the Venue Divergence page.
Full divergence boardFrequently asked questions
ProbCast provides prediction-market data and analytics for informational purposes only. Not financial, trading, betting, investment, legal, or tax advice.