Lech Poznan supplied the round’s loudest correction, and the model still found enough to like

The round was mostly right about direction, less so about scale

This play-off round did not ask Fipster to be perfect; it asked it to be broadly sensible. On that count, it was. The model found plenty of home wins, a fair few away wins and enough correct scorelines to keep the week respectable. Where it stumbled, it usually did so by underestimating how far a tie could open up once one side got control. That is a more useful mistake than being wrong for no reason. It means the read was close, just not quite large enough.

The biggest clue came from Lech Poznan against FC Thun. Fipster had the home side winning comfortably enough, but nothing in the prediction suggested a seven-goal afternoon. The model had the right winner and the right general mood: Lech Poznan were in charge from early on and never really let go. The clean part of that call mattered more than the exact scoreline because it rested on something real — Lech Poznan’s recent scoring run and defensive control — rather than wishful arithmetic. Once they started scoring, they did not stop for tea.

The neat hit: Ferencvarosi TC did exactly what the model expected

Ferencvarosi TC against Trabzonspor was the round’s cleanest example of a call coming off for the right reasons. Fipster backed a home win and named Bamidele Yusuf as a scorer. Ferencvarosi TC won 4-0, and Bamidele Yusuf did score, from the spot. That matters because the model was not merely guessing at a result; it had identified a forward profile that fit the game state Ferencvarosi TC were likely to create at home. The scoreline ran away from the forecast, but the underlying idea held: pressure, territory and one of their own attackers finding his moment.

Kairat Almaty–Anderlecht is where we overreached for caution

The miss that stung most was Kairat Almaty against Anderlecht. Fipster called a draw and leaned into a goalless script; Anderlecht won 3-0 instead. The problem was not that the model fancied caution in general — that would have been defensible — but that it over-weighted low-scoring signals while underplaying Anderlecht’s attacking streak noted in the pre-match reasoning. Once Jaakko Oksanen put through his own net early, the whole thing tilted sharply away from the forecast. After that, Lukáš Ambros and Danylo Sikan made sure there was no polite retreat back into 0-0 territory.

The names that actually shaped things

Bamidele Yusuf deserves his mention twice over: he was one of Fipster’s pre-match calls who came good, even if Ferencvarosi TC ended up scoring far beyond him as Callum O'Dowda, Marius Dumitru Corbu and Krisztián Lisztes joined in. Jesús Imaz also mattered because he turned Jagiellonia’s tie into something much more comfortable with an early penalty before Jeremy Nosakhare Agbonifo and Sergio Lozano added to it; none of those extra scorers were called beforehand, which is exactly why this page exists. Jens Wålemark was another useful reminder that one player can keep turning up in different ways across a round: he scored three for Lech Poznan across their two legs and helped turn a decent forecast into an emphatic one.

There were other useful names too. Rafa scored twice for Benfica after Fipster had backed Aarhus to keep things tighter than they managed; Robert Gumny scored twice for Lech Poznan without being on any pre-match list; Danylo Sikan added another goal Anderlecht could enjoy without much stress by then. The pattern is fairly clear: when Fipster had read the broad attacking edge correctly, individual scorers often followed even if not always in order or quantity. When it missed, it tended to miss on who would supply the decisive punch rather than on which side would carry most of it.

A decent week, then: keep trusting the strong reads and trim the timid ones

So this round leaves us with something useful rather than merely flattering numbers. The model handled most of the obvious favourites well enough and found several correct scorelines along the way, but its weakest calls came when it mistook restraint for certainty. That is fixable. Going into the next set of ties, Fipster should keep backing sides whose attacking patterns are already doing visible work — especially when players like Bamidele Yusuf or Jens Wålemark are already showing up in advance — and be less eager to pin everything on a tidy low-scoring finish when one team has started kicking doors down instead.