Copa del Rey: the favourite wins, just not as often as you think
Since 2019, a LaLiga side visiting a lower-tier club has won inside 90 minutes four times out of five. Against a Segunda side, three out of five. Sixteen of them try again this week.
The big clubs enter the Copa del Rey this week: between Tuesday 27 and Thursday 29 October, sixteen LaLiga sides, Sevilla, Betis, Athletic, Valencia and Villarreal among them, travel to a lower-tier opponent. One match, at the small club's ground. We looked at what happened every previous time.
That is 275 matches from the seven editions played in the current format, 2019-20 to 2025-26. Four times out of five the favourite settles it in 90 minutes. The fifth time it does not, and in half of those cases it is already out at the final whistle. Everything here is the 90-minute result: what happened afterwards in extra time or on penalties is not included.
“If the average holds, of the sixteen, about thirteen will win inside 90 minutes, two will go to extra time and one or two will go out.
The opponent matters more than the ground
The number moves a lot depending on who you visit. Against a Segunda side, the LaLiga club wins inside 90 minutes only 60% of the time and goes out once in four. A Segunda side visiting a lower-tier club does no better. In a one-off match, away from home and with a rotated team, one division of difference weighs far less than it seems.
The last two editions came in lower
The first five single-match editions ranged between 79% and 85%. The last two dropped to 73% and 70%. With around 40 games per edition that is about five fewer wins, and two seasons are not enough to call it a trend. But it is worth watching this week.
And in other cups?
The Copa del Rey sits in the European middle. The German cup, also played this week, is the most predictable for the favourite. The English FA Cup is the least: away from home, the Premier League side wins inside 90 minutes only a little over half the time.
What we learned from our own mistake
We did not look at this history out of curiosity. In August, comparing our probabilities with the market's on a first sample of games, we found that in cup ties between divisions we gave the favourite about 16 points more than the market did, and only there: league games and European competitions showed nothing of the kind. The case that exposed it was Udinese - Padova in the Coppa Italia.
The truth was in between, and closer to us than to the market. Our model measures squad strength, and between the first and second divisions that gap is huge. What it cannot see is who actually takes the field: in the cup the big club rotates, motivation is not the same on both sides, and in a one-off match variance rules. Since then, when the model goes above that cup's historical ceiling, we cut half of the excess. We never raise the favourite, we only lower it.
In the cup, 90% for the favourite is almost never 90%. Away to a lower-tier side, history says four out of five inside 90 minutes; against a second-division side, three out of five. Extra time exists for a reason.
How we measured it: each team's division comes from the league table it played in that season, and the result is the 90-minute one. A team that was in neither the first nor the second division that season counts as lower tier, and we only use seasons where we have both tables complete. No model and no odds are involved.