The result and the grades
Real Madrid won 3-0 at Real Racing Club de Santander, and the verdict on our panel is simple: six predictors, six correct outcomes, zero exact scores. Claude Keeper and Kimi Cool both went for 0-2. Gemini Thunder and GPT Striker both predicted 1-3. DeepThinker and the Baseline both chose 1-2. Every one of them under-called Madrid’s margin, and none gave Racing a clean sheet’s worth of credit for scoring nothing.
The Brier scores separate the group, since they reward confidence in the right outcome. Claude Keeper finished best at 0.0296, followed by GPT Striker at 0.0350. Kimi Cool and Gemini Thunder both landed on 0.0818, and DeepThinker on 0.0950. The Baseline came last at 0.2032, which says it backed the away win with much less conviction than the models did.
The gap between 0.0296 and 0.0818 is not about the scoreline. Kimi Cool and Claude Keeper predicted the same 0-2, yet the probability behind each call differed, and that is what the metric measures. A lower number means more of the probability sat on the outcome that happened.
What the pre-match picture suggested
The context sheet made the lean toward the visitors easy to defend. Madrid arrived under José Mourinho with a form line of WWWLW, 24 points from 10 games and a goal difference of 25 scored against 10 conceded. Their recent run included a 2-1 win at FC Barcelona and a 4-1 win over Sevilla FC. The site’s ratings put their attack at 1.4242 and their defence at 0.8215.
Racing’s numbers pointed the other way. They were 15th on 11 points from 10 games, with 15 goals scored and 26 conceded, and a form line of WLDLL. Their ratings were 1.0365 for attack and 1.5388 for defence, and the recent results included heavy away defeats at Barcelona and Celta Vigo.
There were caveats on Madrid’s side. [1] listed Rodrygo and Raúl Asencio as unavailable, and a separate [2] piece described Aurélien Tchouaméni as doubtful with a physical issue carried over from the World Cup and Ferland Mendy as long-term injured. Racing were without Simon Eriksson, according to [3]. Absences on that scale did not outweigh the gap in form and ratings, and the models treated them that way.
Value picks: where the money was made and lost
Three of the six predictors attached a value pick. Claude Keeper and GPT Striker both backed the away win in the head-to-head market at odds of 1.22. It hit, returning +0.22 each. That is a modest profit because the price was short, which matches how heavily the market expected Madrid to win.
Gemini Thunder took a very different route. Despite predicting a 1-3 away win, its value pick was the draw at 7.5. That is the logic of a value pick: it targets a price the model thinks is too generous, not necessarily the most likely result. The draw did not come in, and the pick lost its full unit (-1.0).
Kimi Cool, DeepThinker and the Baseline registered no value pick. The contrast is useful. Two of the three picks on a favourite at 1.22 paid out a little, and the one longshot cost far more than both of them combined earned. It is a reminder that a model can be right on the outcome and still lose on the pick, if the pick targets a different result.
What it means
For Racing, the defeat leaves them 16th in La Liga with 11 points from 11 games (W3 D2 L6, goals 15 for and 29 against). Their last five read 0-5 at Celta Vigo, 0-2 at Osasuna, a 2-2 draw with Valencia, a 2-1 win over Espanyol and now this 0-3 loss. The goals-against column is the worry: 29 conceded in 11 matches is well over two per game. Their next fixture is Málaga against Santander on 7 November, and the site’s forecast for that game locks 24 hours before kick-off, so there is no call to report yet. The question it asks is whether Racing can defend for 90 minutes after conceding at that rate.
For Madrid, the win moves them to third with 27 points from 11 games (W9 L2, 28 scored and 10 conceded). They have now won their last five across all competitions, including the 4-1 over RB Leipzig and the 4-3 at AS Roma in the Champions League. Their next fixture is AEK Athens against Real Madrid in the Champions League on 4 November, with the site’s forecast also locking 24 hours before kick-off. The test there is whether the same attack, scoring close to three a game in the league, can travel again with Rodrygo, Asencio and Mendy still listed among the injured.
Sources
- Marca: Mourinho recupera a Vini para el Leipzig marca.com ↩
- Marca: 100 días de Mou: 81% de triunfos, azote del caso Negreira, pacificador y sanador de una gran 'herida' marca.com ↩
- Marca: Racing de Santander - Espanyol | Reacción o caída en picado: previa, análisis, pronóstico y predicción marca.com ↩
