How good are these forecasts?
Every forecast scored on seasons the model never saw while its settings were chosen.
Held-back seasons
| Forecast | Forecasts | Log loss | Favourite won |
|---|---|---|---|
| The model | 1253 | 0.673 | 59% |
| Base rate | 1253 | 0.687 | 55% |
| Recent form | 1253 | 0.681 | 56% |
Against the betting markets
Each market's price before the start, scored on the same finished games as the model. A negative figure means the market was more accurate.
| Market | Since | Priced | Model log loss | Market log loss | Model vs market |
|---|---|---|---|---|---|
| Bookmakers (Pinnacle, DraftKings, bet365) | 21 Jan 2026 | 384 | 0.656 | 0.646 | −1.5% |
| Polymarket | 1 Sep 2026 | 164 | 0.687 | 0.671 | −2.4% |
| Kalshi | 13 Feb 2026 | 279 | 0.663 | 0.663 | 0.1% |
The table is Liiga's: three points for a win in regulation, two for a win in overtime or a shootout, one for losing in one.
The playoff odds follow the current format: twelve teams, the top four waiting while the rest play best-of-five first-round series, then best-of-seven rounds re-seeded so the best team left meets the worst.
Liiga has no promotion or relegation, so the same clubs return every season.
Goaltending is not in the rating, and live scores aren't shown for Liiga yet.
Questions and answers
How accurate are PodiumOracle's Liiga predictions?
On 1,253 games from held-back seasons (2024–25 to 2026–27), which the model never saw while it was tuned, the side it favoured won 59% of the time. Its log loss was 0.673, against 0.687 for the base rate and 0.681 for recent form. Against the betting markets on the same finished games, its log loss was 0.656 against the bookmakers' 0.646 over 384 games, 0.687 against Polymarket's 0.671 over 164 games and 0.663 against Kalshi's 0.663 over 279 games (lower is better).