Football Results & Model Review Saturday 31 October 2026: Brighton shock City as Coventry beat Sunderland
Saturday 31 October gave our model plenty to chew on. We logged 119 completed fixtures with pre-match predictions attached, and the headline figure is that 62 of our 119 winner calls came in. That is a shade better than half, and it sits above our season-to-date winner accuracy across all leagues, which stands at 47%. A decent day, then, though not one to dine out on, and the detail underneath is more interesting than the headline.
The goals markets were a more modest story. Our both-teams-to-score calls landed 65 times out of 119, and our Over 2.5 goals calls also landed 65 times out of 119. Those are the same tallies, which is a neat coincidence rather than anything deeper. In the six Premier League matches we look at closely below, our model leaned towards goals in every one of them. It was proved right on both counts in two of them, Coventry against Sunderland and Hull City against Ipswich, and wrong on both in the other four.
That is the honest shape of the day: a model that sniffed out a few of the winners, got caught out by a couple of the shocks, and kept expecting goals that some of the games simply did not provide. Below we go through each of the six in turn, hits and misses alike, then sweep up the rest of the card across the Championship, La Liga and Serie A. Where we got it wrong, we will say so plainly.
Premier League (England)
Chelsea 2-0 Manchester United
Chelsea beat Manchester United 2-0, and our model had the right team in front. Before a ball was kicked we gave Chelsea a 45% chance of winning, with the draw at 24% and a Manchester United win at 31%. Our expected goals read 1.86 for Chelsea and 1.50 for United, which painted a fairly tight picture with a slight lean to one side. That lean was enough for a clear winner call, and the winner call was a HIT.
The goals were a different matter, and the story of the match is in how late they came. Nothing separated the sides until the 50th minute, when Morgan Rogers put Chelsea ahead, with Romeo Lavia providing the assist. That one goal changed the temperature of the afternoon, and it was soon followed by a bigger moment still.
In the 63rd minute Lisandro Martínez was shown a red card for Manchester United, and the very next minute Cole Palmer converted a penalty to make it 2-0. Those two events, a minute apart, effectively settled the contest. Our model could not have known about a sending-off, of course, and nobody should expect it to, but it is worth remembering how much a single red card can reshape a match that looked evenly poised on paper.
Our other calls fared worse. Our most likely scoreline was 1-1, at 10%, and the match finished 2-0, so that was off the mark. We also called both teams to score at 66% and Over 2.5 goals at 65%, and both were a MISS, because United never found a goal and the match stopped at two. So: a HIT on the winner, two misses on the goals markets, and a reminder that a 45% favourite is only ever a modest favourite.
Full match analysis, xG and correct-score grid βPremier League (England)
Manchester City 0-1 Brighton
This was the most uncomfortable result of the day for us. Brighton won 0-1 against Manchester City, and our model had Manchester City as the favourite. We gave City a 44% chance of winning, the draw 23% and Brighton 33%. Our expected goals were 1.92 for City and 1.67 for Brighton, a pairing that suggested a lively game with goals at both ends. Our winner call, City at 44%, was a MISS, and we will not dress it up as anything else.
The match, as it turned out, was settled almost immediately. In the 11th minute Yasin Ayari scored for Brighton, set up by Maxim De Cuyper, and that was the only goal of the game. Among the key events there is nothing after it: no equaliser, no second, no card. A single early strike was all that separated the sides over the full ninety minutes.
It is worth saying that our model never made Brighton an outsider. A 33% chance is roughly one in three, and outcomes like that arrive regularly across a full season. The probabilities were not wild; the result was simply one of the less likely branches. That is why we publish the percentages and not only the call.
The goals markets went the same way. We put both teams to score at 70% and Over 2.5 goals at 70%, both calls were yes, and both were a MISS, since only one side scored and only one goal was scored in total. Our most likely scoreline was 1-1, at 9%, and the real score was 0-1. So a miss on the winner, a miss on both teams to score and a miss on the goals line. A thoroughly humbling set from one match.
Full match analysis, xG and correct-score grid βPremier League (England)
Coventry 3-1 Sunderland
Coventry beat Sunderland 3-1, and here our model went the wrong way. Pre-match we made Sunderland the narrow favourite, with a win at 41%, the draw at 25% and a Coventry win at 34%. Our expected goals were 1.50 for Coventry and 1.67 for Sunderland. The winner call, Sunderland at 41%, was a MISS. It was a close call on paper, but a miss is a miss, and Coventry were convincing winners on the scoreboard.
The match was effectively decided before the hour. Tatsuhiro Sakamoto opened the scoring in the 17th minute, and Taiwo Awoniyi doubled the lead in the 37th. Early in the 52nd minute Jack Rudoni added a third, with Victor Torp providing the assist, and at 3-0 the contest had a very one-sided look to it.
Sunderland did pull one back. Enzo Le Fée scored a penalty in the 67th minute to make it 3-1, which is where the score stayed. Late on, in the ninety-ninth minute of the clock as listed, 90+9', Omar Alderete was shown a red card for Sunderland, a footnote in a match that had long since been decided.
There was some consolation for us on the goals markets. We called both teams to score at 64% and Over 2.5 goals at 61%, and both were a HIT, thanks to Le Fée's penalty and a total of four goals. Our most likely scoreline was 1-1, at 11%, and the match finished 3-1, so that was not close. One miss on the winner, two hits on the goals, and a good reminder that a correct goals call does not mean we read the game properly.
Full match analysis, xG and correct-score grid βPremier League (England)
Hull City 2-2 Ipswich
Hull City and Ipswich drew 2-2, a result that made our model look a little too confident in one direction. We gave Hull City a 50% chance of winning, with the draw at 25% and an Ipswich win at 25%. Our expected goals were 1.81 for Hull City and 1.24 for Ipswich. Our winner call, Hull City at 50%, was a MISS, because the match ended level. A draw was always on the table at 25%, but we did not call it.
The game had a good shape to it. Ipswich took the lead in the 22nd minute through Florentino Luís, who scored from an Abdoul Ouattara assist. Hull City needed until the 58th minute to respond, and when they did, it was Tim Iroegbunam who levelled, set up by Oliver McBurnie.
McBurnie then turned from provider to scorer. In the 68th minute he put Hull City 2-1 ahead, and for a few minutes it looked as though our call would be rewarded after all. It was not to be. In the 74th minute Julio Enciso equalised for Ipswich, with Azor Matusiwa providing the assist, and that was the last of the key events.
The goals markets were kinder to us. We called both teams to score at 60% and Over 2.5 goals at 59%, and both were a HIT, with four goals shared between the sides. Our most likely scoreline was 1-1, at 11%, and the match finished 2-2, which is the right result shape but the wrong number of goals. So a miss on the winner, two hits on the goals, and a draw that our model gave only a one-in-four chance.
Full match analysis, xG and correct-score grid βPremier League (England)
Brentford 2-0 Nottingham Forest
Brentford beat Nottingham Forest 2-0, and our model had this one right. We made Brentford favourites with a 52% chance of winning, the draw at 24% and a Nottingham Forest win at 23%. Our expected goals were 1.84 for Brentford and 1.16 for Forest. That 52% was the highest winner probability among our six featured matches, and the winner call was a HIT.
It was a quiet match in terms of key events, with only two goals to talk about. The first arrived in the 30th minute and was an own goal, recorded against Nikola Milenković. That is a scrappy way to break a deadlock, and it is the sort of goal no model can pencil in. Nothing else was recorded until the 76th minute.
That was when Igor Thiago scored, with Jaidon Anthony providing the assist. Those two goals were all there was, and the final score was 2-0 to Brentford. The match had no cards listed among its key events, and it never threatened to become the end-to-end game that our goals numbers had hinted at.
That is where our model came unstuck. We called both teams to score at 59% and Over 2.5 goals at 58%, and both were a MISS. Our most likely scoreline was 1-1, at 11%, which also missed. So the lesson is a familiar one: we picked the right winner, but we leaned towards goals when the match gave us two. A HIT on the result, and two misses on the markets that came with it.
Full match analysis, xG and correct-score grid βPremier League (England)
Tottenham 2-0 Crystal Palace
Tottenham beat Crystal Palace 2-0, and our model picked the winner. We gave Tottenham a 48% chance of winning, the draw 25% and a Crystal Palace win 26%. Our expected goals were 1.73 for Tottenham and 1.23 for Palace. The call, Tottenham at 48%, was a HIT, and the result was a good one for our record on the day.
What our model could not have foreseen was how early the match would change. In the 13th minute Jean-Philippe Mateta was shown a red card for Crystal Palace, which reduced them to ten players for most of the afternoon. It is the sort of event that reshapes a match before it has properly begun.
Tottenham took the lead in the 25th minute. Archie Gray scored, with Mohammed Kudus providing the assist, and the score stood at 1-0. The second goal came in the 73rd minute, when Jan Paul van Hecke made it 2-0. Those were the only goals in the key events, and Crystal Palace never found a reply.
As with several results on the day, the goals markets let us down. We called both teams to score at 59% and Over 2.5 goals at 57%, and both were a MISS. Our most likely scoreline was 1-1, at 11%, and the match ended 2-0. In fairness, a red card that early makes a 1-1 draw a harder ask. So a HIT on the winner, two misses on the goals, and an early dismissal that rewrote the script.
Full match analysis, xG and correct-score grid βElsewhere
Away from the six matches we have looked at closely, the Premier League card had a few more results worth noting. Bournemouth won 2-0 against Leeds, and our call of a Bournemouth win at 40% was a HIT. Aston Villa and Fulham drew 1-1, and we had called an Aston Villa win at 45%, so that was a MISS. In the Championship, the picture was largely a pleasing one. Most of the winner calls we made there came in, including Norwich's 3-2 win over Watford, Stoke City's 2-0 win over Preston and Bristol City's 3-2 win at Bolton. Middlesbrough beat Charlton 3-0, Millwall beat Derby 4-1, Southampton beat QPR 4-1, Swansea beat Blackburn 4-0 and Wolves beat Cardiff 3-0. Lincoln against Birmingham finished 1-3, another HIT for us.
The Championship misses were Burnley's 0-0 draw with Portsmouth, where we had called a Portsmouth win at 43%, and West Ham's 1-3 defeat by West Brom, where we had called a West Ham win at 49%. In La Liga, Barcelona's 6-1 win over Alaves and Real Betis's 2-1 win over Malaga were both hits, but Levante's 0-2 defeat by Atletico Madrid and Rayo Vallecano's 3-0 win over Celta Vigo both went against our calls. In Serie A, Bologna's 2-1 win over Monza and AS Roma's 5-2 win at Udinese were hits, while AC Milan's 3-1 win over Inter was a miss. There were also 93 more results we have not listed here, all of which count towards the day's overall record.
- Bournemouth 2-0 Leeds · model called Bournemouth win (40%) · HIT
- Aston Villa 1-1 Fulham · model called Aston Villa win (45%) · MISS
- Lincoln 1-3 Birmingham · model called Birmingham win (50%) · HIT
- Norwich 3-2 Watford · model called Norwich win (58%) · HIT
- Stoke City 2-0 Preston · model called Stoke City win (64%) · HIT
- Bolton 2-3 Bristol City · model called Bristol City win (56%) · HIT
- Burnley 0-0 Portsmouth · model called Portsmouth win (43%) · MISS
- Middlesbrough 3-0 Charlton · model called Middlesbrough win (61%) · HIT
- Millwall 4-1 Derby · model called Millwall win (50%) · HIT
- Southampton 4-1 QPR · model called Southampton win (59%) · HIT
- Swansea 4-0 Blackburn · model called Swansea win (47%) · HIT
- West Ham 1-3 West Brom · model called West Ham win (49%) · MISS
- Wolves 3-0 Cardiff · model called Wolves win (49%) · HIT
- Levante 0-2 Atletico Madrid · model called Levante win (40%) · MISS
- Rayo Vallecano 3-0 Celta Vigo · model called Celta Vigo win (34%) · MISS
- Real Betis 2-1 Malaga · model called Real Betis win (61%) · HIT
- Barcelona 6-1 Alaves · model called Barcelona win (62%) · HIT
- Bologna 2-1 Monza · model called Bologna win (52%) · HIT
- Udinese 2-5 AS Roma · model called AS Roma win (66%) · HIT
- AC Milan 3-1 Inter · model called Inter win (52%) · MISS
So what does Saturday 31 October tell us about the model? That it is useful, and not infallible. A winner record of 62 of 119 is respectable, and it is a reminder that a favourite at 45% or 50% is only ever a modest favourite. Brighton's win over Manchester City and Hull City's draw show how often the less likely outcome turns up, while the goals markets were a gentler lesson, with 65 of 119 on both. If you want to see how any of these matches were rated, head to the match pages on Fixture360, and take a look at our model accuracy record for the full season picture, misses included.
Explore the markets
This post was written with AI assistance from Fixture360's own model data β every figure, score and event comes from our database, and the stat cards are generated directly from it. Predictions are statistical estimates for entertainment and information only, not betting advice. Please gamble responsibly (18+).