Strategy Lab
offline backtests on your real history — every result is leave-one-out + bootstrap CI. Nothing here affects live predictions.
Running experiments…
staking Monte-Carlo — your real settled bets, reshuffled
Loading…
Monte-Carlo match simulator (C-SIM upgrade) — win% distribution from expected totals
Team A exp total
Team B exp total
format
This shows how much the toss-proof win-edge depends on volatility — a small expected-total gap with high variance is far less certain than the point estimate implies. Offline only; not used by live pricing yet.