Atletico Madrid vs Real Madrid: Reading a Close Forecast
What the archived Madrid derby percentages mean, why a narrow model edge is uncertain, and how to distinguish website forecasts from paid software.
The earlier Madrid derby preview assigned almost equal win probabilities to Atletico Madrid and Real Madrid. The useful question is how to read that narrow gap without turning decimal precision into confidence.
Figures from the earlier preview
| Outcome | Previously published estimate |
|---|---|
| Atletico win | 37.67% |
| Draw | 24.07% |
| Real Madrid win | 38.26% |
| Over 2.5 goals | 58.81% |
| Under 2.5 goals | 41.19% |
The win/draw/win values sum to 100%. Real Madrid's lead over Atletico is 0.59 percentage points. This is a small difference between estimates; the article does not provide an uncertainty interval that would establish it as a meaningful advantage.
Removing the draw and rescaling gives Real Madrid about 50.39% of the non-draw probability. That is a conditional calculation, not a new forecast and not a prediction that the draw cannot happen.
Why 1–1 and over 2.5 are different statements
The previous article named 1–1 as the leading exact score without publishing its individual probability. An exact score is one outcome. Over 2.5 covers all scores with at least three goals, so it can receive more probability than any single score. Neither number establishes that both teams will score.
Evidence and current information
These values are preserved from the earlier article, not newly generated here. The original fixture-specific software output and generation time have not been established. The product image previously used showed another fixture, so it could not validate the derby percentages.
For the current free website estimate, see the Atletico–Real Madrid forecast, including its recorded evidence where available. A refreshed page can differ from this archive. The website methodology explains that separate model.
Mad Football Prediction AI Software is a related paid product. Its listing and demonstrations should be assessed separately; these archived figures do not establish its performance.
Correction, 19 September 2026: Removed unsourced injuries, standings and tactical explanations, and the claim that a different-fixture product screenshot proved these figures.
This article was revised with AI assistance to check the published text, available images and calculations. It is not an independent test of the software. Predictions are estimates, not guarantees. 18+. Support for gambling-related harm.