The model found the narrow wins, then got mugged by the scorelines

What the round said about the model

This was a round where Fipster kept finding the right sort of game and then occasionally missing the actual script. The model was happiest with narrow home wins and a few controlled away legs, and that did show up often enough to keep the record respectable. But there were also too many ties that broke open, too many one-goal ideas that became routs, and a couple of matches where the scoring pattern arrived wearing someone else’s name tag. In other words: decent instincts, imperfect timing. Football remains stubbornly unhelpful like that.

The one call that landed properly

Lech Poznan against KI Klaksvik was the cleanest star turn. Fipster backed a home win and, more importantly, read it as a match where KI would struggle to impose themselves. That held from start to finish. The post-match data showed Lech scoring five without reply, which is not exactly subtle but does make a point rather loudly. The model’s pre-match read on KI’s attacking problems was sound; it simply underestimated how sharply those problems would be punished once Lech got going.

Where we were properly wrong

The Salzburg–Pafos tie was the awkward one. Fipster leaned into an away win and a fairly tidy scoreline, but the match turned into a 3-3 draw instead. The mistake was not just picking the wrong side; it was over-weighting recent form into something more orderly than it proved to be. Once Haris Tabaković started scoring for Salzburg, the game stopped behaving like a neat away-win case study and became something much messier. That is usually how these things go when football decides to ignore your spreadsheet for an evening.

The names that shaped the round

The obvious headline act was Haris Tabaković, whose four-goal haul for Salzburg completely changed the shape of one of the round’s biggest misses. He had not been flagged pre-match by Fipster, which is exactly why he matters here: he was one of those unforeseen scorers who can make a prediction look very tidy in theory and very silly in practice. Graham Burke also mattered across Shamrock Rovers’ tie, though again he was not among any pre-match scorer calls; his goals helped turn that fixture into another reminder that late-stage qualifying can be wonderfully rude to forecasts.

Rafa at Benfica deserves mention too because he did appear in Fipster’s player callouts list before scoring, so that one at least matched up neatly with reality. Gylfi Sigurðsson also scored after being flagged pre-match for Vikingur Reykjavik, which is exactly what you want from this sort of exercise: evidence that some player reads are doing real work rather than just decorating the page. Around them sat several other unforeseen scorers — D. Adjessa, A. Albanese, I. Gruda, Lelê and others — which is another way of saying this round had no shortage of people arriving uninvited to ruin tidy predictions.

Looking ahead without pretending certainty is fashionable

The useful lesson from this round is not that Fipster should chase chaos for its own sake. It is that low-confidence calls need room for volatility, especially in qualifying ties where one hot scorer can flatten an otherwise sensible read. The model still found enough correct shapes to stay useful; it just needs to keep treating these rounds as living things rather than polite diagrams. Next time out, we will take the stars where they come — and try not to let the misses borrow quite so much furniture.