CSKA 1948 and Spartak Trnava produced one of the most unusual Conference League encounters in recent memory, with a 0-0 draw that somehow masked an extraordinary penalty shootout in extra time. The designated home side dominated normal play without finding the breakthrough, but the drama exploded in the 120th minute when Dvali converted from the spot for CSKA. Spartak answered immediately through Gong, and the penalty sequence escalated frantically: Franco made it 2-1 to CSKA, then Kostrna levelled, Cuellar restored the advantage, Bukovsky equalized again, and Frederic put CSKA ahead once more before Lausic drew Spartak level. Iliev's penalty in the 120+5th minute ultimately settled the tie for the designated home side in one of the strangest finishes you'll see at this level.

Our pre-match model predicted a 3-1 CSKA win and assigned a 44% probability to a home victory, with a 33% lean toward a draw. That 33% draw probability was well-founded—the actual scoreline sits squarely in the plausible outcomes the model had weighted—but the route to it was anything but conventional. We'd expected the designated home side to create clear chances based on their underlying metrics, and CSKA did press throughout normal play without converting. Spartak proved resilient and dangerous on the break, which the model had flagged as a threat, so the cautious final score in open play reflected a tighter contest than our 3-1 projection suggested.

The penalty shootout framework sits outside typical match prediction, so our call can't account for that chaotic sequence. What matters is that we correctly identified a competitive fixture where multiple outcomes were live, even though the exact path to this result was never going to be forecast by pre-match data. It's a sharp reminder that tight, low-scoring matches can still deliver drama—just not always the kind analysts can predict.