New to these reports? Start here
- Dotted-underlined words have a plain-English definition — hover or tap them. Every term is also on the glossary page.
- "Null" means we found nothing, not that something broke. Most reports here are negative results, on purpose — knowing an idea doesn't work is the point.
- Two questions get asked separately. First, is the effect real? Second, is it already priced into the betting odds? An effect can be completely real and still useless to bet on.
- A "calibration" row is a self-check. It runs the same method on something already known to be true. If that fails, the whole report is unreliable — so it's reported alongside the findings.
- If a confidence interval includes zero, the real effect might be nothing at all, so no claim gets made.
Do shift-derived ice time minutes improve shots-on-goal predictions? (2026-10-05)
Data: NHL shift charts 2021-22..2025-26 (6,936 games, ~5.3M shifts; 57 2024-25 games missing from the NHL feed) → player_games (TOI by EV/PP/PK + SOG). Regular season only. Scored 2022-23..2025-26: 169,298 player-games with ≥ 5 prior games. Every prediction is point-in-time.
Two predictors: - A: season-to-date SOG/game, shrunk toward last season / position. - B: SOG rate per situation-weighted minute (PP minutes weighted ×1.88, measured) × the player's last-5-games minutes by situation.
Lower deviance and Brier are better.
| n | deviance A | deviance B | B − A (95% CI95% confidence intervalThe range the true value is plausibly in. If this range includes zero, we cannot rule out that the real effect is nothing at all.) | Brier SOG≥2 A/B | |
|---|---|---|---|---|---|
| All | 169,298 | 1.2460 | 1.2450 | −0.0010 (−0.0023, +0.0002) | 0.2144 / 0.2142 |
| Recent TOI moved ≥ 2 min | 21,866 | 1.2689 | 1.2557 | −0.0131 (−0.0199, −0.0063) | 0.2182 / 0.2164 |
| Stable TOI | 147,432 | 1.2427 | 1.2434 | +0.0008 (−0.0002, +0.0017) | 0.2138 / 0.2139 |
Read: - Overall: ice time adds essentially nothing. A season average already captures who shoots. - Deployment-change games (13% of games: a player moved up or down the lineup, or on or off the power play): a real but small gain. That's about a 1% deviance improvement and 0.0018 in Brier for 2+ shots. As a probability shift it is roughly 1-2 points. - Against a prop line: this is well inside a typical prop hold (~5-8% two-way). It could only matter if books are slow to move lines after deployment changes. That is plausible but untested: there is no historical NHL prop data (OddsPapi has none).
Next: - NHL SOG and points props are captured from 2026-10-05 (capture-only). - After ~4-6 weeks, test ONLY deployment-change games: does B's probability beat the captured line enough to clear the hold? If not, close the angle. - Faceoff data was not pursued (too small; inside xG; priced).