data
We pull live odds from real sportsbooks
We ingest live odds from multiple sportsbooks on a regular schedule and score every available matchup.
Live now
Every Sunday the takes pile up. We score every game, publish what we're on, and show what we passed on, with the reason.
We detect. You decide.
Where to start
Right now: 12 picks on the board, 0 we passed on. Graded on 2045 settled picks.
01The hold
Knowing what not to trust is the product. Four checks keep weak games off the board, and every pass is logged in public, same as a published pick.
WE PASSED
and we show you why
Sometimes the sharpest pick is no pick.
Most products treat an empty board as a failure to hide. We treat it as the system working. Every pass is a decision with reasons attached: recorded, public, and accountable, exactly like a published pick.
The proof
Picks publish with tamper-evident receipts, frozen before kickoff: hashed, committed, impossible to edit after the game. No fabricated picks, no invented stats, no silent edits. And you don't have to take our word for it. Check a receipt yourself.
Record
Receipts freeze before kickoff. Hits print with period, sample size, model version, and the definition used. Thin samples stay dark with the gap named. Recompute any seal yourself.
Reliability · settled sample
n 2045
Diagonal = perfect calibration · buckets need 30+ settled
Open full record →
Methodology / Trust Surface
Every published pick ties back to live markets, timestamped data, factor scoring, and the gates that keep weak picks off the board.
Settled picks calibrated
the only basis for any number
Cleared today
passed every gate
Gated today
restraint, logged
Live counts · the numbers below explain how each one is produced
Board data as-of
data
We ingest live odds from multiple sportsbooks on a regular schedule and score every available matchup.
data
Each pick is scored against the bookmakers that had a market for the game at the time of scoring. We surface the bookmaker count as a transparency signal.
data
Where available, each pick shows the timestamp of the odds data it was scored against so you can judge freshness for yourself.
model
Confidence is a 0-100 selection score, shown as "NN/100" and never as a percentage. It ranks how much value we think is on the board, and it is not a win probability. The only win probability we publish is the market-implied one, taken from the book prices on the pick.
Until we have enough settled games to calibrate against, confidence is a label, not a number.
model
Each pick carries a risk level reflecting bookmaker consensus, market depth, and known volatility factors.
model
Subscribers with the right entitlement can see a factor-by-factor breakdown of how each pick was scored.
gate
Public performance statistics are only displayed after the platform has accumulated enough settled, canonical picks to compute them honestly.
When you see win-loss numbers, you'll also see the period, sample size, model version, and the exact definition.
This product is research, not certainty. We get better by adding settled rows and clearer uncertainty, not by making louder claims.
Sports wagering is real risk. Only stake what you can afford to lose without changing your week. If you or someone you know has a gambling problem, call 1-800-GAMBLER. Past performance does not guarantee future results.