Sign in Register

← All posts

Recap Saturday 10 October 2026

Football Results & Model Review Saturday 10 October 2026: Chelsea's rout, Ipswich's late twist and a Premier League day of draws

Saturday 10 October was the sort of football day that keeps a statistical model humble. We had 135 completed fixtures with pre-match predictions on the board, and by the final whistle our winner calls had landed correctly in 63 of 135. That is a respectable showing without being a victory lap, and we would rather tell you so plainly than dress it up. The Premier League supplied the drama, the Championship supplied the volume, and La Liga, Serie A and the Bundesliga supplied a few reminders that favourites do not always read the script.

Beyond the winner calls, our both-teams-to-score calls were right in 71 of 135, and our Over 2.5 goals calls were right in 75 of 135. So the goal-market style calls edged ahead of the straight winner calls on the day, which tells its own story: plenty of matches had goals at both ends, and several finished with late drama that no pre-match probability could reasonably have flagged. Our season-to-date winner accuracy across all leagues stands at 47%, and it is worth keeping that figure in mind when judging any single Saturday.

Below we go through six Premier League matches in detail: Arsenal against Leeds, Chelsea against Bournemouth, Sunderland against Brighton, Aston Villa against Brentford, Ipswich against Fulham and Manchester United against Tottenham. For each we set out what happened, what our Dixon-Coles goal model expected, and whether our call was a hit or a miss. Then we sweep through the rest of the day's results, before drawing a few honest conclusions about what it all means for the model.

Premier League (England)

Arsenal 2-1 Leeds

Final score
2-1
Model's call
Arsenal win 47% · HIT
Predicted score
1-1
Both teams score
Yes · HIT
Over 2.5 goals
Yes · MISS

Arsenal beat Leeds 2-1, and they had to come from behind to do it. Our model had Arsenal as the most likely winner at 47%, with the draw at 28% and a Leeds win at 25%. The expected goals read 1.49 to 1.00, so we saw an Arsenal edge rather than a procession. The most likely single scoreline in our model was 1-1, at 12%, which shows how thin the margins are even when the call goes our way.

Leeds struck first. James Justin scored on 55 minutes, with Gabriel Gudmundsson providing the assist, and the visitors led 0-1. That is the sort of moment that makes a favourite's 47% look flimsy, and for a few minutes our winner call was in real trouble. Arsenal did not leave it long, though. Riccardo Calafiori equalised on 62 minutes, set up by Ben White, to make it 1-1.

Then came the decisive blow. Bruno Guimarães scored on 70 minutes, with Declan Rice supplying the assist, and Arsenal led 2-1. There were no further goals in our record, and the home side saw it out. So the match tracked the shape our model suggested: Arsenal slightly ahead on the numbers, Leeds a live threat, and a goal at each end. Interestingly, the 1-1 scoreline we rated as most likely was exactly where the match stood after 62 minutes, before Guimarães moved it on.

Scoring our own homework: the winner call, Arsenal at 47%, was a HIT. Our both-teams-to-score call was yes at 50%, and both teams did score, so that was a HIT too. The Over 2.5 goals call was the blemish. The model put it at 46% and called no, but the match went over, so that one was a MISS. Two from three, then, and a reminder that a 46% chance is a long way from a 0% chance.

Full match analysis, xG and correct-score grid β†’

Premier League (England)

Chelsea 5-1 Bournemouth

Final score
5-1
Model's call
Chelsea win 40% · HIT
Predicted score
1-1
Both teams score
Yes · HIT
Over 2.5 goals
Yes · HIT

Chelsea 5-1 Bournemouth is the scoreline of the day, and it is the one that looks least like our model's picture. Before kick-off we had Chelsea win at 40%, the draw at 24% and a Bournemouth win at 36%, with expected goals of 1.73 to 1.62. In plain English, we expected a close, open game with goals for both sides. The most likely scoreline was 1-1 at 10%. What we got was a Chelsea romp, and we will own that gap.

The game was effectively settled before most supporters had found their seats. Jordan Henderson scored on 2 minutes, with Cole Palmer assisting, and Morgan Rogers made it 2-0 on 3 minutes, with Henderson this time the provider. Two goals in the opening moments turned a game our model saw as nearly even into a one-sided affair. Chelsea added a third on 46 minutes, when João Pedro scored from a Pedro Neto assist, to make it 3-0.

Bournemouth did not fold entirely. Francisco Evanilson pulled one back on 51 minutes to make it 3-1, which kept the visitors' goal threat alive and, importantly for our records, meant both teams had scored. But Chelsea restored control. João Pedro scored again on 67 minutes, assisted by Pep Chavarria, for 4-1, and Henderson completed his brace on 70 minutes to make it 5-1. It was a thoroughly emphatic finish to a match that began with a flurry.

On the scorecard, this was a clean sweep. Our winner call, Chelsea at 40%, was a HIT. The both-teams-to-score call was yes at 67%, and Evanilson's goal made it a HIT. The Over 2.5 goals call was yes at 65%, and the match sailed past that, another HIT. We would add a note of caution, though: the calls were right, but the scale of the win was far beyond what our expected goals suggested, and we would not claim the model saw this coming.

Full match analysis, xG and correct-score grid β†’

Premier League (England)

Sunderland 0-2 Brighton

Final score
0-2
Model's call
Brighton win 47% · HIT
Predicted score
1-1
Both teams score
No · MISS
Over 2.5 goals
No · MISS

Sunderland 0-2 Brighton went the way our winner call suggested, but not in the way our goal calls expected. The model had Brighton as the most likely winner at 47%, with a Sunderland win at 27% and the draw at 25%. Expected goals were 1.28 to 1.74, a clear lean towards the visitors. Our most likely scoreline was 1-1, at 11%, so even where we backed Brighton, we imagined Sunderland finding a goal.

They never did. The match stayed goalless until 53 minutes, when Maxim De Cuyper scored for Brighton, with Pascal Groß providing the assist, to make it 0-1. Then came the strangest twist of the afternoon. On 57 minutes Pascal Struijk was shown a red card for Brighton, which in theory handed Sunderland a route back into the game. Our record shows no Sunderland goal arriving to take advantage.

Brighton held firm and then sealed it in stoppage time. Jaouen Hadjam scored on 90+3 minutes, with Georginio Rutter assisting, to make it 0-2. So Brighton won despite the red card, and Sunderland's home pressure came to nothing. It was a match that shifted in tone after the dismissal, but not in outcome, and the late goal gave the result a gloss that the closing stages might not otherwise have earned.

Our winner call, Brighton at 47%, was a HIT. The goal calls were less happy. We called both teams to score at 60%, and Sunderland failed to score, so that was a MISS. We called Over 2.5 goals at 58%, and the match finished with fewer than that, another MISS. One from three, then. It is a good example of how a model can read the balance of a match correctly and still misjudge how many goals it will contain.

Full match analysis, xG and correct-score grid β†’

Premier League (England)

Aston Villa 2-2 Brentford

Final score
2-2
Model's call
Brentford win 37% · MISS
Predicted score
1-1
Both teams score
Yes · HIT
Over 2.5 goals
Yes · HIT

Aston Villa 2-2 Brentford was about as even as our model thought it would be, which did not stop us getting the headline call wrong. We had Aston Villa win at 35%, the draw at 28% and a Brentford win at 37%, with expected goals of 1.34 to 1.39. The most likely scoreline was 1-1, at 12%. Brentford were the slightest of favourites in our eyes, and so our winner call was Brentford at 37%, a call we made with very little conviction.

The game began briskly. Yehor Yarmoliuk put Brentford ahead on 10 minutes, making it 0-1, and Nicolas Jackson levelled for Aston Villa on 12 minutes to make it 1-1. That early exchange matches the scoreline our model rated as its single most likely outcome, and for a long spell it looked as if the match might simply stay there. The middle of the game, as far as our record goes, brought no further goals.

The finish was frantic. John McGinn scored on 77 minutes, with Johan Manzambi providing the assist, and Aston Villa led 2-1. Brentford needed only two minutes to reply. Kristoffer Ajer scored on 79 minutes, assisted by Jaidon Anthony, and it was 2-2. A game that had sat quietly for a long stretch produced a pair of goals in the space of a few minutes, and neither side found another winner.

The winner call was a MISS: we backed Brentford at 37% and the match was drawn. The draw itself had been given 28% by our model, so it was hardly a fanciful outcome. The goal calls fared better. Both teams to score was called yes at 56%, and that was a HIT. Over 2.5 goals was called yes at 51%, and that was a HIT as well. So a split verdict, with the goals reading sound and the winner reading wrong, which is about what you would expect from a match this balanced.

Full match analysis, xG and correct-score grid β†’

Premier League (England)

Ipswich 2-1 Fulham

Final score
2-1
Model's call
Fulham win 35% · MISS
Predicted score
0-1
Both teams score
Yes · MISS
Over 2.5 goals
Yes · MISS

Ipswich 2-1 Fulham was the day's clearest case of a match that our model saw as a coin toss, and then flipped awkwardly. Before kick-off we had Ipswich win at 32%, the draw at 33% and a Fulham win at 35%, with expected goals of just 0.97 to 1.04. That is about as tight as a three-way split gets. Our most likely scoreline was 0-1, at 14%, and our winner call was Fulham at 35%.

For a long time it looked as if that call might come good. Ryan Sessegnon scored for Fulham on 55 minutes to make it 0-1, exactly the scoreline our model had singled out as its most likely, and the match settled into a tense, low-scoring pattern. There was no further goal in our record until very late, and for most of the afternoon the visitors appeared to be heading for a narrow win.

Then came an extraordinary finish. Abdoul Ouattara equalised for Ipswich on 88 minutes, making it 1-1. Barely any time later, Dara O'Shea scored on 90+1 minutes, with Jack Clarke providing the assist, and Ipswich led 2-1. A match that had been grinding towards a Fulham win ended up as a home win delivered in stoppage time. It is the sort of finish that makes no allowance for expected goals.

We have to be candid here: this was a clean sweep of misses. The winner call, Fulham at 35%, was a MISS because Ipswich won. We called both teams to score no, at 41%, and both scored, so that was a MISS. We called Over 2.5 goals no, at 33%, and the match went over, another MISS. Our model expected a low-scoring game and saw a tight one decided by two goals in the closing moments, which no pre-match number could sensibly have predicted.

Full match analysis, xG and correct-score grid β†’

Premier League (England)

Manchester United 1-1 Tottenham

Final score
1-1
Model's call
Manchester United win 54% · MISS
Predicted score
1-1
Both teams score
Yes · HIT
Over 2.5 goals
No · MISS

Manchester United 1-1 Tottenham gave our model its most confident miss of the day. We had Manchester United win at 54%, the draw at 25% and a Tottenham win at 21%, with expected goals of 1.78 to 1.03. That was a clear lean to the home side, the strongest favourite among our six detailed matches. The most likely scoreline was 1-1, at 11%, which is precisely where the match ended, though the winner call went the other way.

For a long time the home side's favouritism looked well founded. The match stayed goalless until Bryan Mbeumo scored for Manchester United on 73 minutes, with Matheus Cunha providing the assist, to make it 1-0. That goal was precisely what our model's lean implied. Five minutes later, though, the match shifted. Rodrigo Bentancur was shown a red card for Tottenham on 78 minutes, which on paper should have helped United see the game out.

It did not work out that way. Tottenham stayed in the contest and in stoppage time Dominic Solanke scored on 90+2 minutes, with Sandro Tonali providing the assist, to make it 1-1. So a side that had been shown a red card pulled level at the very end, and the home team's lead was erased. It was a cruel finish for United and a satisfying one for Tottenham, and it left our winner call stranded.

The verdict is mixed. Our winner call, Manchester United at 54%, was a MISS, because the match was drawn. Both teams to score was called yes at 54%, and it was a HIT. Over 2.5 goals was called yes at 53%, but the match finished below that line, so that was a MISS. A favourite at 54% still fails to win a good deal of the time, and the draw sat at 25% in our numbers. That is not an excuse, merely the arithmetic of probability.

Full match analysis, xG and correct-score grid β†’

Elsewhere

Away from the six matches we have dissected, the day was busy and decidedly mixed. In the Championship, Swansea beat Norwich 2-1, Blackburn beat Cardiff 2-1 and Middlesbrough edged Wolves 1-0, all of them matching our winner calls. Preston against Millwall finished 0-1, and our call for Millwall was right too. But the second tier is where models go to be humbled. Charlton drew 1-1 with Bristol City, Derby drew 1-1 with Wrexham, West Brom lost 0-1 to Birmingham and Bolton lost 1-2 to Stoke City, and in each case our pick did not come through. Sheffield Utd beat Lincoln 3-0, and Burnley won 0-4 at Watford, where we had backed the home side at 57%, which was a firm miss.

Elsewhere in Europe the picture was similarly patchy. In La Liga, Barcelona beat Getafe 3-0 and Real Madrid beat Villarreal 1-0, both calls landing, with Barcelona at 73% the most confident of the pair, while Atletico Madrid won 1-2 at Alaves, where we had gone the other way. In Serie A, Inter beat Parma 3-0 and Napoli beat Frosinone 3-0, but Genoa's 2-1 win over Fiorentina went against our call. The Bundesliga was the most entertaining set of scores: Hamburger SV won 2-3 at 1899 Hoffenheim, Bayern München drew 2-2 at FC Augsburg, Bayer Leverkusen won 3-4 at FSV Mainz 05, and Union Berlin beat SV Elversberg 1-0. Plenty more results came in beyond those we list here, and they fed into the overall tallies.

So what does the day tell us about the model? Mostly that it is doing its job as a probability tool and not a crystal ball. Our winner calls landed in 63 of 135, our both-teams-to-score calls in 71 of 135 and our Over 2.5 goals calls in 75 of 135, all against a season-to-date winner accuracy of 47%. The hits were real, the misses were real, and the biggest lessons came from matches such as Chelsea's rout, Ipswich's stoppage-time swing and Manchester United's late draw, where the numbers pointed one way and football went another. We would rather show you all of that than only the flattering parts. If you want to dig further, the individual match pages on Fixture360 carry the full probabilities and expected goals, and our model accuracy record keeps the running score, wins and misses alike.

← Previous: Football Results & Model Review Thursday 8 October 2026: Borchers hat-trick, Neymar's penalty double and two Brazilian draws

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+).