The result and the grades
Valencia won 1-0 away at Elche on 1 November, and no predictor saw it coming. Every one of the six entries, Claude Keeper, Kimi Cool, Gemini Thunder, GPT Striker, DeepThinker and the Baseline, forecast a 1-0 home win. The outcome was wrong for all of them and the exact score was missed by all of them too, so the grading separates them only through confidence.
The Brier scores show how far each one leaned toward the wrong side. Claude Keeper was the least exposed at 0.6584, followed by Kimi Cool at 0.7646. Gemini Thunder, GPT Striker and DeepThinker all landed on 0.855. The Baseline was the worst at 1.0357, which suggests it put the most weight on an Elche win. Lower is better in this measure, and none of these numbers is a good one.
The value picks fared no better. Five of the six entries flagged the home side on the h2h market at odds of 2.325, and each of those lost its unit for a profit-and-loss figure of -1.0. The Baseline did not place a value pick. A unanimous direction, a unanimous miss and a flat -5.0 across the five staked entries make this a clear failure for the field.
Why the models leaned home
The pre-match file gave a reasonable case for Elche. They came in on the back of a 3-0 win at Rayo Vallecano and two goalless draws away at Villarreal and at home to Celta. That run meant three games without conceding. They sat 16th with 10 points from 10 games, a goal difference of -3 (14 scored, 17 conceded).
Valencia’s side of the file looked worse. They were 19th with 6 points from 10 and had won once all season. Their last five read L-D-D-L-W, with a 1-2 home defeat to Villarreal the week before. They had scored only 7 goals and conceded 17. The site’s own ratings had Elche’s attack at 1.0192 against 0.8343 for Valencia, and Valencia’s defence at 1.1117 against 0.9825 for Elche.
The head-to-head also pointed home. Elche won 1-0 against Valencia in April, and the January meeting finished 1-1. With home advantage, a recent clean-sheet streak and a weaker visitor, a narrow Elche win was an easy forecast for six models to share. The shared lean is the real lesson. When every entry reads the same signals in the same way, a single upset exposes all of them at once.
What the context did and didn’t tell us
Some of the background had a bearing on the game, though none of it said how the match would go. Elche were short of Marc Aguado, who [1] reported would miss the next matches until after the international break. A missing midfielder is the kind of detail a model can price in only roughly.
Valencia carried a different sort of pressure. The club has publicly announced that it is for sale, as [2] noted. Guido Rodríguez also admitted he needed to improve personally while he assesses his physical state after an earlier injury, according to [3]. Neither item suggested a team about to win on the road.
That is why the result is a useful reminder. Form tables and ratings capture trends, but a bottom-three side can still take three points from a team above it, and a single goal is enough to decide a low-scoring game. We have no further match detail in the input, so we do not claim to explain how the goal came about. We can say only that the scoreline matched the narrow margin the models predicted and went to the other team.
What it means
For Elche, the defeat leaves them 17th in La Liga on 10 points from 11 played (W2 D4 L5, 14 scored and 18 conceded). The loss ended a short unbeaten spell that included the win at Rayo, and the goalless draws at Villarreal and against Celta had suggested a side that was hard to beat. Their next fixture is at home to Real Betis on 8 November. It asks whether Elche can turn defensive solidity into points at home, where the Celta draw and now this loss have brought one point from two games.
For Valencia, the win is their second of the league season. It lifts them to 9 points from 11 (W2 D3 L6, 8 scored and 17 conceded), but they stay 19th. It also followed a 1-2 home defeat to Villarreal, so one result does not yet make a run. Their next game, at home to Real Madrid on 7 November, is a very different test from this one, and it will show whether the win was a turning point or a one-off.
The site’s forecasts for both of those fixtures lock 24 hours before kick-off, so the models will have another chance to be graded in public.
Sources
- Marca: Anselmi, sobre el penalti anulado al Elche en Villarreal: “Los errores de procedimiento son completamente evitables” marca.com ↩
- Marca: Por fin ha comenzado la salida de Peter Lim del Valencia marca.com ↩
- Marca: Guido: "En este momento los débiles se rinden y nosotros no nos vamos a rendir" marca.com ↩
