Model lab

How it works

What goes in, how it is checked, and where it falls short.

How the forecast works

Every senior men's team has one rating: how many points it would beat an average team by. After each international the rating moves up or down by how far the real margin differed from what the ratings expected. Home advantage is worth about 5.5 points, and is switched off at neutral venues. Win, draw and loss chances come from a bell curve around the expected margin.

Only past results are used: 6,357 senior men's internationals since 1995 from World Rugby. No bookmaker odds, no injuries, no team news.

How good is it?

Held-out 2023 to Nov 2026

+31%
better than a plain home/draw/away guess, 762 matches, settings fixed on 2010-19 only

Winner picked

77%
of the same 762 matches

vs World Rugby formula

-1.6%
slightly behind their ranking formula on 745 non-drawn matches (likely noise, but not a win)

The +31% looks big because most internationals are mismatches, which any rating gets right. World Rugby's own formula is the fair test, and we do not beat it. Result-shuffling checks scored −30% to −42%, so the skill is not an artefact. A different curve shape did not help. This has not been compared with bookmaker odds.

Where the data comes from

World Rugby's public ranking and match feed. It is undocumented, so it could change without notice, and its terms have not yet been confirmed. Club and provincial rugby (URC and others) and player form are next.