Every number on this site is generated by an automated pipeline and graded against reality daily. Predictions are logged before matches start; wins and losses both stay on the page.
Predicts the full map veto — bans, picks, decider — for any CS2 matchup, with a public track record graded on every completed match.
Open the predictor →Simulation-only paper ledger: model-flagged edges at flat 1-unit stakes, logged before kickoff, with per-bet closing-line value. No real money.
See the ledger →The prediction model's public scorecard: every pre-match win probability it logged, graded automatically — accuracy and calibration, win or lose.
Check the record →How this works under the hood: the data pipeline, the models, and the honest-measurement rules the site holds itself to.
Read the case study →An automated pipeline ingests match data, odds, and rosters every day — no hand-picked numbers.
Predictions are written down before a match starts. Nothing is added after the fact.
When results land, every prediction is scored by the same script, the same way, every time.
The full record stays up — good runs and bad ones. That is the point of the site.