Premier League Gameweek 8 Takeaways: What the Data Says
Updated October 26, 2026
Updated October 26, 2026
Gameweek 8 was a tale of missed chances, and of teams with the better numbers finding ways to lose. Manchester City out-created Aston Villa 1.56 xG to 0.61 and still suffered their first league defeat of the season. Brighton won the xG battle at Anfield and lost 2-0. Crystal Palace racked up 17 shots and 1.59 xG at home to Newcastle and lost by the same score. Nottingham Forest piled up 2.29 xG at Ipswich and left with a point.
I’ve gone through the expected goals, big chances and shot data from all 10 fixtures and lined them up against the picks and futures positions we’ve published on ATS.io this season. Some of those calls look sharper today. A couple... not so much.
The gap between points and expected points (xPts) is the thread running through this week. Nottingham Forest are 18th on six points with the third-best xPts total in the league. Newcastle are eighth with the worst. Let's dive into the data.
City won almost every underlying number at Villa Park except the scoreline. They had 64% of the ball, 15 shots to eight, seven on target to two and 56 touches in Villa’s box to 12. They also missed three big chances.
Nicolas Jackson won it with two goals from three shots worth 0.37 xG. Erling Haaland’s header in between cashed the headline pick from our Aston Villa vs Man City props.
Don’t read too much into Villa’s win. Their 12.82 xGA is the fifth-worst in the league, and their 9.3 expected points rank 15th, although it must be said that they managed to pull of a similar trick last year. At the other end, City’s 21 points sit 5.6 above their expected tally, the biggest overperformance in the division.
The football isn’t driving the futures markets, though. Arsenal are level with City on 21 points and now trade at 66¢ (-194) to win the title on Polymarket. That’s up from the 61¢ I passed on in our title futures.
City Yes in the relegation market has climbed to around 41¢, which puts City No near 63¢. That’s below the 77¢ entry in our relegation futures article.
Until we know the outcome of City's appeal, and the timeline for any sanctions they should face, leave City alone entirely in both title and relegation markets.
Forest drew 1-1 at Ipswich on Friday after out-creating them 2.29 xG to 0.46. They had 17 shots and missed three of their four big chances. Morgan Gibbs-White alone got through 0.76 xG without scoring.
Over eight games, Forest have six goals from 13.79 xG, a 7.8-goal finishing gap and the widest in the league. They’ve also allowed just 8.79 xG, second only to Arsenal. That adds up to the third-best expected points total in the league and 18th place in the real one.
Oliver Glasner’s side trade around 17¢ to go down on Polymarket. That’s too high for a team creating chances like this, and I’d rather be on Forest No at roughly 85¢ (-567). The risk is time. If the goals don’t come by the new year, the table starts to matter more than the chances.
João Pedro wasn’t in Chelsea’s matchday squad on Saturday, and it showed. Chelsea produced 0.47 xG at Stamford Bridge and lost 2-0 to Tottenham. That’s less than the 0.52 xG João Pedro had been averaging per appearance on his own.
He’d been the focal point of almost everything Chelsea Chelsea have done well this season, scoring 5 and assisting 3 of the 17 goals (from 10.33 xG) the Blues have scored this season. Without him, Xabi Alonso's men were almost completely blunted, with Morgan Rogers struggling to fulfill the role vacated by the Brazilian.
Chelsea had just 17 touches in the Tottenham box, seven shots and no big chances. Four of those shots hit the target, but at an average of under 0.07 xG per attempt, they were mostly hopeful efforts from range. Cole Palmer, usually at his best playing off a striker, managed two shots worth 0.06 xG before he was sent off on 80 minutes.
Spurs controlled 69% of the ball and completed 569 passes to Chelsea’s 215, but they only created 0.74 xG themselves, but their victory was richly deserved.
João Pedro’s absence on its own doesn’t explain Chelsea’s season, though. They’re fifth on 13 points, but their 10.3 expected points rank 12th. That’s a mid-table side that’s been finishing like a top-four one, and its best finisher is the reason why.
Until there’s clarity on his fitness, Chelsea’s goal output looks the most fragile in the top half.
Spurs, meanwhile, look to be trending upwards. On October 1 we flagged Tottenham Yes to be relegated at 15¢ (+567), with Roberto De Zerbi’s side bottom on two points. They’re now 14th on nine, and Spurs Yes trades between 5¢ and 8¢. That trade has gone against us. I wouldn’t add to it, because Spurs’ 9.93 xGA is the fifth-best in the league.
My Liverpool vs Brighton props preview argued that Brighton’s four straight shutouts were hiding 10.12 xG conceded, and Liverpool broke the run inside 21 minutes through a Pascal Groß own goal. 4 minutes later Alexander Isak’s finish from Bradley Barcola’s pass made it 2.
After that, Brighton had 64% of the ball, took 17 shots to 13 and won the xG 1.32 to 1.20. They put one shot on target, with Yasin Ayari's miss at 0-0 particularly egrigious.
Brighton’s season has been a story of overperformance at both ends. They’ve scored 23 from 16.22 xG and conceded seven from 11.36 xGA, a +16 goal difference on a +4.9 expected goal difference. At 3.6¢ in the title market, I’m not interested.
Liverpool’s title price sits at 7.7¢ (+1199), barely moved from the 7.4¢ we flagged. That trade was always about City’s sanctions. Their +2.5 xG difference, sixth in the league, doesn’t change that.
Bournemouth’s 3-0 win at Old Trafford wasn’t a smash-and-grab. They created six big chances to Manchester United’s none and won the xG 2.15 to 1.17 with only 37% of the ball. Rayan, Marcus Tavernier and Evanilson scored.
The season numbers back it up. Bournemouth’s 14.47 xG is the third-highest total in the league, and so is their +5.2 xG difference. That works out to 13.6 expected points, fifth-best, against the nine they’ve banked in 13th.
Their relegation Yes sits around 6¢, and nobody’s pricing them as a contender either. The value will be in match markets, where a 13th-placed side tends to get mid-table prices. That’s where I’ll be looking over the next few weeks.
Liverpool vs Arsenal (Sunday, 11:30am ET) is the headline fixture of Gameweek 9. Arsenal arrive with the league’s best defense by a distance, 6.19 xG allowed in eight games. Liverpool’s 9.61 xGA is the fourth-best. Polymarket has Arsenal at 42¢ (+138), the draw at 27¢ (+270) and Liverpool at 32¢ (+213).
That’s close to where the underlying numbers put it, so I’m not forcing a side. The bigger impact is on the title market. An Arsenal win at Anfield would push their 66¢ price higher still, and I’m happy to keep passing on it.
Brentford vs Nottingham Forest (Saturday, 11am ET) is where I’d look first. Forest trade at just 27¢ (+270) away from home against a Brentford side priced at 50¢ (-100).
On expected points, Forest rank third in the league and Brentford rank eighth. A finishing streak this cold can’t last forever. At 27¢, I’d rather be on Forest than pay even money for Brentford.
Chelsea vs Manchester United (Saturday, 8:30am ET) hinges on João Pedro. Chelsea are 53¢ (-113) to win, a price that looks like it assumes he’s back leading the line.
United’s 10.9 expected points are slightly better than Chelsea’s 10.3, so this is closer than the table suggests. If João Pedro misses out again, United at 24¢ (+317) become the more interesting side, despite them seemingly staring into the abyss after a miserable week. Wait for team news before taking a position.
