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 | 257 | 0.613 | 71% |
| Base rate | 257 | 0.678 | 66% |
| Recent form | 257 | 0.629 | 68% |
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 |
|---|---|---|---|---|---|
| Polymarket | 13 Feb 2026 | 83 | 0.609 | 0.583 | −4.6% |
Ratings use every game in the related competitions (see Sources); only Super Rugby Pacific games are scored in the track record.
The simulated table adds bonus points the union way: four for a win, two for a draw, one for losing by seven or less, one for four tries (at each side's usual rate).
International call-ups during the Six Nations and test windows weaken clubs unevenly; the model doesn't know who is away.
Questions and answers
How accurate are PodiumOracle's Super Rugby Pacific predictions?
On 257 games from held-back seasons (2024 to 2026), which the model never saw while it was tuned, the side it favoured won 71% of the time. Its log loss was 0.613, against 0.678 for the base rate and 0.629 for recent form. Against the betting markets on the same finished games, its log loss was 0.609 against Polymarket's 0.583 over 83 games (lower is better).