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 | 209 | 0.898 | 59% |
| Base rate | 209 | 0.975 | 45% |
| Recent form | 209 | 0.966 | 54% |
Every nation is rated on all its men's Tests, Elo-style, with the margin of victory (a win by wickets counts about 20 runs a wicket, a draw half a win).
A season here is a World Cup cycle: every international from the day after one final to the next. Bilateral series are forecast and rated but aren't in the event's table.
The 2025–27 World Test Championship is simulated in full in every run: every Test left in the cycle between the nine nations is played with today's ratings (12 points a win, 4 a draw), the table is ranked by the share of points available, and the top two meet in the final.
The field isn't settled yet, so this is a provisional one: the hosts and the nations that would qualify on today's ratings, seeded into groups by rating. It changes as results come in and will be replaced by the real draw once it's announced.
The 2025–27 final is at Lord's in June 2027. Slow over-rate deductions, which the ICC takes off a side's points, aren't included, so the real table can differ a little.
Every international is treated as neutral: home advantage isn't modelled, because the archive doesn't say who was at home. That flatters touring sides a little.
Squad selection (rested stars, part-time associate sides) is not in the model.
A Test's result depends on the pitch and the weather more than a one-day match's; the draw rate the model learns is the average across grounds.