Weekly review
Week of 28 September
28 September 2026 – 4 October 2026 · 1,240 words
Between 28 September and 4 October we had published a prediction before kick-off for 115 fixtures, across 7 days and 9 competitions. We called 58 of them correctly, 50.4%.
Does our confidence mean anything
Taking every call in the period and sorting it by how confident we were: we said 40 to 50% on 42 fixtures, averaging 45.0%, and were right 50.0% of the time; we said 50 to 60% on 33 fixtures, averaging 54.2%, and were right 51.5% of the time. At the other end, 50 to 60% came out within 2.7 points of what we claimed, on 33 fixtures — there, the figure on the card means what it says.
The outcome we almost never call
Of 115 fixtures, 36 ended level — 31.3% of them. We called a draw in none of them. Not rarely: never, and that is worth explaining rather than apologising for. The three probabilities on each card come from simulating the two sides' scoring rates independently and counting how often each outcome appears. That method has a built-in ceiling on the draw: across this period the highest draw probability we published was 32.2%, and the nearest a draw ever came to leading its fixture was 6.3 points short. For the draw to top a card, two teams would have to be so exactly matched that neither win is likelier than the tie — which is rarer than a tie itself. So the headline call cannot name the outcome that happens a quarter of the time, and roughly a quarter of our misses are lost before kick-off. Independence is also the part we would change first. Real football scores are not independent at the low end: 0-0 and 1-1 happen more often than two separate rates predict, which is why the standard correction in the literature exists and why our draw numbers are likely a little low across the board. Until that is fixed, the probability we print for the draw is the figure worth reading on a tight fixture — not the call above it.
Home, away, and which we get right
We backed the home side 82 times and the visiting side 33, so 71% of our calls went to the home team. They landed 48.8% of the time. The away calls landed 54.5%. That gap runs the wrong way for the shape of our output: the calls we make least often are the ones that come off most. The reading we would defend is not that away teams are undervalued in general, but that we only name the visitor when the case is strong, and are readier to give the home side the benefit of a close one. The fix is not to back more away teams; it is to be less willing to call a fixture that is genuinely tight.
The goals we expected
Underneath the win probabilities sit two expected-goals figures, and they are the part of the model that is easiest to check and hardest to fudge. Across 115 fixtures we expected 1.60 goals for the home side and 1.17 for the away side, a total of 2.77. What actually happened was 1.30 and 1.15, a total of 2.44. The home figure is the one that drifts: we were over by 0.30 of a goal per match. It is small in absolute terms and it is systematic, which is the more interesting property — a bias that holds across hundreds of fixtures is correctable in a way that a wide spread is not. Rounding each expectation to a whole number reproduced the exact scoreline 10 times, 8.7%. We mention that as a curiosity rather than a claim: nobody should be predicting exact scorelines, and a model that appeared good at it would be a model that had been overfitted.
The high-scoring flag, tested
One model on each card does nothing but estimate how likely the home side is to score four goals or more. Over 115 fixtures that happened 10 times, 8.7%. The flag was raised on 17 of them, 15% of what we published. Where it was raised, the home side reached four goals 29.4% of the time; where it was not, 5.1%. So it is separating the two, by a factor of about 5.8. That is what the flag is asked to do, and worth stating plainly given it is also the part of the output we have most recently had to correct.
The calls we backed hardest
12 calls were published at 70% or above. 10 of them came in, 83.3%, against an average stated confidence of 79.1%. The most expensive of them was Azerbaijan to beat Liechtenstein at 74.6%, which finished 0–0 — a draw, which is the outcome we structurally cannot call.
The call we would point to
The call we are most pleased with is not the most confident one, it is the one furthest from the price. Belgrano (Córdoba) to beat Talleres (Córdoba) in Liga Profesional Argentina: we published 44.5% where the odds implied about 33.9%, a gap of 10.6 points, and it finished 0–1. 3 calls in this period were that far above the market and came in. We list one rather than all of them because a page of the ones that worked, with the ones that did not left out, is how a track record gets quietly turned into advertising — the figure for how that whole group did is in the section above, and it is not flattering.
Where we were most wrong
Azerbaijan against Liechtenstein, in Int: UEFA Nations League, finished 0–0. We had published Azerbaijan to beat Liechtenstein at 74.6%, the largest confidence we placed on anything that did not happen in this period. It is worth noting what our five worst misses have in common: 5 of them ended level. The model's most confident errors are not upsets, they are draws. It stays on the record. A site that reports only its hits is not reporting a record, it is reporting a selection.
Against last week
Against the week before: 86 settled fixtures then at 41.9%, 115 now at 50.4%, a move of +8.6 points. Volume changed more than accuracy did — up 34% — so the two figures are not measuring quite the same thing: a week with more fixtures is usually a week with more of the competitions we read least well. More useful than the headline: in the 50 to 70% range, where most of our calls sit, we were 3.5 points below our stated confidence last week and 7.8 points below it this one. That gap widened, which matters more than the headline percentage moving the other way.
What we covered
The period spanned 9 competitions, none with enough fixtures to say anything useful about individually.
How to read all this
One week of 115 fixtures settles very little on its own. A percentage point either way on a sample this size is noise, and we would rather say so than dress it up. What these reviews are for is the shape that emerges when several of them sit side by side: whether a stated 65% keeps landing near 65%, whether the gap between our number and the market's keeps costing us, whether a competition we read badly this week is still being read badly in three. Those are answerable, and they are answerable in public, which is the only reason to publish the bad figures alongside the good ones.
Every figure here comes from predictions published before the match day — never from a re-run after the results were known. Results come from public fixture feeds. Every period we have reviewed → How the model is measured →