Recent content by oddsfactory

Hi everyone, I'm the founder of OddsFactory AI (oddsfactoryai.com). This is my own project, so treat this as an ad. I'm posting in this section because it allows it, and because the discussions here on calibration and track records are exactly the standard I'm trying to meet. The problem I...

Oli's point about the lineup window raises a problem for anyone publishing picks rather than placing them. A pick posted at noon and one posted after lineups are different bets, but a track record usually counts them as comparable. I'd argue a published record should carry the publication time...

Agree with Klaus and derbyScala that raw form is the least reliable item on the list. A 2-0 at 0.4 xG and a 1-1 at 2.8 are very different teams. Most of the other factors (injuries, motivation, congestion) really feed into the expected XI, and that's the one input that can change after a pick is...

Respect for posting the misses along with the wins. One question on the daily results (e.g. 4/8): are those counted at the odds available when you posted, or at the closing price? A hit rate alone doesn't show profit, since BTTS and draw-no-bet picks pay very differently. Do you track profit per...

The split you describe, model calibrated before your hand touched it and not after, is the cleanest test I've seen. Did you log the adjustments in a separate column? That would let you score "model alone" against "model + me" over the same bets.

Klaus, a 70-75% bin that's 5 points off while the middle bins hold is the shape I'd expect from a model built to disagree with the market: the biggest "edges" are partly model error. Two questions on your 18% shrink. Did you fit it on one season and test it on another, or on all 2,140 bets? And...