How to read a sports forecast record
One winning pick proves little. To evaluate a forecast, start with what was shown before the event, then inspect the result and the wider record.
1. Check when the call appeared
Compare the event start time with the first captured publication time. A later screenshot cannot prove a pregame call. Pontis's older record was reconstructed from saved picks files; those rows do not have a verified first-seen timestamp. Read the missing-time label instead of assuming those calls were locked before kickoff.
2. Separate estimate from price
A model probability is an estimate, not a guarantee. A listed sportsbook price is a captured observation that can change. Check its source and time. Even a model that picks many winners can fail to show betting value when the available prices are poor. A probability or win rate alone cannot establish profit.
3. Read every kind of outcome
Look at the date range, sample size, wins, losses, pushes, and voids. A public record should show unfavorable periods as readily as winning ones. Historical tests and calls published as time passed answer different questions, so keep those groups separate.
4. Compare the same games
When usable betting prices exist, Pontis's reconstructed record shows its win rate and the betting favourite's win rate on the same priced games. The all-games Pontis record can have a larger denominator. The betting-favourite pair is not a closing-line-value or same-price return measure. Check the matched count and method before interpreting the difference.
Questions to ask
- Can I find the misses and voids?
- Which entries have a verified first-publication time?
- Do both comparison rates use the same games?
- Is the result a live publication, reconstruction, or historical test?
Pontis Picks publishes forecasts and research; it does not accept or place wagers. A reader can choose to do nothing with a forecast.
Inspect the reconstructed record · Open the Pontis Picks board · Responsible sports-betting resources from NCPG