The favourite still has a football match to win

A favourite can score and still lose. The football figures are current as of 5 October 2026. A strong favourite can tempt us to mentally award the league points before kick-off. The smaller probability deserves more respect than that. Surprise is part of forecasting, and a result can be painful without proving that the original preference was foolish. The model gave Monaco an 8.1% chance of beating PSG on 4 September 2026 (Ligue 1). The model gave PSG a 73.3% chance of beating Monaco on 4 September 2026 (Ligue 1). The model gave a draw between PSG and Monaco on 4 September 2026 (Ligue 1) an 18.5% chance. PSG 1-2 Monaco on 4 September 2026. Marquinhos scored in the 31st minute. Flávio Nazinho scored in the 58th minute. Stanis Idumbo-Muzambo scored in the 72nd minute. A narrow margin gives a surprise its own kind of force. The attraction of a favourite is easy to understand before kick-off; afterwards, the temptation is to explain away everything that disagreed with the expectation. A cooler response leaves room for uncertainty without taking any pleasure away from the winning side. The model gave Arsenal a 68.2% chance of beating Brighton Hove on 19 September 2026 (Premier League). The model gave Brighton Hove an 11.6% chance of beating Arsenal on 19 September 2026 (Premier League). The model gave a draw between Brighton Hove and Arsenal on 19 September 2026 (Premier League) a 20.2% chance. Brighton Hove 3-0 Arsenal on 19 September 2026. Pascal Groß scored in the 31st minute. Charalampos Kostoulas scored in the 45th minute. Chema Andrés scored in the 57th minute. The emphatic margin makes this reversal feel different. A winning probability concerns the outcome, without promising how comfortable the football will look. The size of the defeat adds to the sting, but it should not be squeezed backwards into a certainty that nobody held before kick-off. The model gave Lorient a 12.4% chance of beating RC Lens on 5 September 2026 (Ligue 1). The model gave RC Lens a 67.1% chance of beating Lorient on 5 September 2026 (Ligue 1). The model gave a draw between RC Lens and Lorient on 5 September 2026 (Ligue 1) a 20.5% chance. RC Lens 0-1 Lorient on 5 September 2026. Mamadou Koné scored in the 54th minute. Souleymane Sagnan was sent off in the 90th minute. A slender margin offers a reminder about the price of expectation. The favourite can carry most of the model's confidence and still have an uncomfortable contest ahead. A proper reading of a probability means allowing that possibility to remain alive, even when instinct wants to hurry past it. The model gave Hull City a 14.8% chance of beating Coventry City on 29 August 2026 (Premier League). The model gave Coventry City a 63.7% chance of beating Hull City on 29 August 2026 (Premier League). The model gave a draw between Coventry City and Hull City on 29 August 2026 (Premier League) a 21.5% chance. Coventry City 0-1 Hull City on 29 August 2026. Liam Millar scored in the 82nd minute. The timing adds its own sting to a narrow defeat. The pre-match confidence cannot protect a side through the closing stretch. The opposition keeps its right to win throughout the contest, however inconvenient that is for a forecast that seemed persuasive before the ball was kicked. The model gave Athletic a 15.3% chance of beating Celta on 30 August 2026 (La Liga). The model gave Celta a 63% chance of beating Athletic on 30 August 2026 (La Liga). The model gave a draw between Celta and Athletic on 30 August 2026 (La Liga) a 21.7% chance. Celta 0-2 Athletic on 30 August 2026. Álex Berenguer scored in the 20th minute. Roberto Navarro scored in the 22nd minute. An early advantage and a late advantage ask different things of an opponent. The forecast cannot promise which difficulty will arrive. A surprise should make us more careful about certainty, while leaving the football itself room to explain why the experience of defeat felt so uncomfortable. 7 of the 455 reconciled predictions gave the result that happened a probability under 10%. The cut-off describes how unlikely an outcome looked beforehand. A low probability still leaves that outcome possible. A favourite's defeat is a completed test of the preference, and the proper response is to examine the expectation over time rather than rewrite it after the result. A probability describes uncertainty before kick-off. A probability cannot promise the margin, the timing of a goal or the identity of a scorer. That distinction matters when a favourite loses emphatically: the model preferred an outcome, while the contest supplied an experience. The judgement of the preference means keeping those ideas apart, without making the defeated favourite's discomfort less real. A low probability deserves attention precisely because it remains possible. A habit of treating the smaller outcome as irrelevant makes a forecast look simpler, but also makes it less useful when the football becomes uncomfortable. The model has 455 graded football