Scorigami Watch: Matchweek 6

The model's record

After each round, the odds posted before it are checked against what happened, and the round is added here, starting with Matchweek 6.

How the odds are worked out

Every Premier League result since 1992 decides which scores have never happened. The last three seasons go into a goals model: each club gets an attacking and a defensive rating, fitted to its results with older matches counting for less (a match a year ago counts half as much as one today). Home sides get a boost, and a newly promoted club starts from a typical promoted side's ratings until its own results take over.

The ratings give each side's expected goals for a fixture, and treating each side's goals as a Poisson count gives the chance of every possible score, from 0–0 up to 15–15. The odds of a never-seen result add up the chances of every score that has never happened with the same side at home.

What counts as never seen?

A result counts as new when that exact home–away score has never happened, the empty squares on the home/away grid. That includes a classic scorigami (a score never seen either way round) and a result that has only happened the other way round, such as the first away 2–7. The first-of-the-season figure is the chance of a result that hasn't happened yet this season.

How accurate is it?

Tested on every match from 2016/17 to 2025/26, using only the results before each one, the model expected 3.3 never-seen results across those ten seasons; there were 3. It expected 360 first-of-the-season results; there were 345. It can't know about injuries, red cards or weather, so treat a single fixture's odds as a guide rather than a forecast. Since 2026/27 every round is also scored live, in the model's record above.

When is it updated?

Every morning, from the latest fixtures and results. Before each round, @footygami posts the round's odds, and when a never-seen result happens it posts it within minutes of full time.