Kierros 2

Biggest surprise. Chelsea 4–3 Brighton. Brighton arrived top of the table having put four past Villa and four past Tromsø inside a week, and lost a seven-goal game to a side that had 26% of the ball and completed 155 passes. Alonso's Chelsea barely pressed at all — they sat off, waited, and scored four times anyway. Ruthless on the counter and a defence that conceded three at home. Both of those are true, and only one of them survives a harder fixture.

Worst afternoon. Aston Villa. Two matches, no goals, and the number that actually indicts them: no shot on target in either. Buendía clipped Raya's crossbar on twelve minutes and that was the whole of Villa's evening — an xG of 0.3 against opponents who managed 1.05 themselves. The last time Villa lost their opening two without scoring was 2002.

Arsenal, since I have to. Saka finished from close range in the second half and Arsenal were nowhere near the level of the Coventry night. Unchanged eleven, long spells of possession, one goal, and a Tzolis booking that made the half-time substitution compulsory rather than tactical. Winning badly is a useful habit. It is not the same as being good, and saying so is the entire point of writing any of this down.

Ipswich. Led at Old Trafford on 29 minutes, lost 5–2. The scoreline flattered United, which changes nothing about the scoreline.

What I got wrong. Both singles, both in this league. Bournemouth–Everton at 46% home finished 1–1. Coventry–Hull at 49% home lost to the away side — and that was the one match where the forecast deliberately sat under the market's ~54%. The disagreement was argued in advance. It was also wrong.