Four gameweeks is not long enough for expected goals to be the truth, but it is long enough for the gap between the table and the numbers to be worth naming. Two sides in particular are living a different season from the one their underlying figures describe, and both of them go into Gameweek 5 priced on the league table rather than on the performances.
Everything below comes from our own Premier League xG table, which updates as each round's data lands.
Hull City are third on eight points. On expected points they are on 4.0, which would put them tenth. That is a gap of four full points, the largest in the division in either direction, and it is being built at both ends: five goals scored from 3.6 expected, two conceded from 5.8 expected against. Their expected goal difference is minus 2.2, fifteenth in the league.
Their weekend was the pattern in one game. Chelsea 2-2 Hull City looked like a point earned at Stamford Bridge, and on chances it was better than that: Hull City created 1.35 expected goals to Chelsea's 1.07. But a side that has now conceded two goals from close to six expected is not doing anything repeatable at the back.
The mirror image is Fulham, eighteenth on one point with 5.5 expected points. Minus 4.5 is the worst luck in the league. They have taken 6.8 expected goals worth of chances and scored four of them, and their goalless draw at Anfield was another round of the same: 0.77 expected goals, no reward, and a point that flattered nobody. A team creating at that rate does not stay eighteenth.
Newcastle United are tenth on five points with 2.4 expected points, the lowest figure in the division. Three league games have produced 3.0 expected goals for and 6.3 against, an expected goal difference of minus 3.3, which is seventeenth. They have scored six and conceded four.
That is the whole of their start compressed into two numbers, and it matters tonight, because they travel to Elland Road to close the round against Leeds United, who have created less (3.9 expected goals in three) but conceded less too. Our model makes goals at both ends 59.6 percent in that game, and the Newcastle United defensive number is most of the reason why. The full read is in our Leeds United against Newcastle United preview.
All nine games now have expected goals in. The two Sunday fixtures landed from the provider after this piece first published, and they change the top and the bottom of the round.
| Match | Score | xG | Total |
|---|---|---|---|
| Coventry City v Brighton & Hove Albion | 0-5 | 1.47 - 2.50 | 3.97 |
| Sunderland v Arsenal | 0-2 | 1.79 - 2.14 | 3.93 |
| Aston Villa v Nottingham Forest | 1-2 | 1.06 - 2.27 | 3.33 |
| Crystal Palace v Ipswich Town | 2-3 | 0.92 - 1.83 | 2.75 |
| AFC Bournemouth v Brentford | 2-2 | 2.09 - 0.52 | 2.61 |
| Chelsea v Hull City | 2-2 | 1.07 - 1.35 | 2.42 |
| Liverpool v Fulham | 0-0 | 1.45 - 0.77 | 2.22 |
| Tottenham Hotspur v Everton | 0-0 | 0.70 - 1.31 | 2.01 |
| Manchester United v Manchester City | 0-1 | 0.84 - 0.77 | 1.61 |
The derby is the number of the round. Manchester United against Manchester City produced 1.61 expected goals between the two of them, comfortably the quietest game of the weekend and below two goalless draws. Manchester City won it on 0.77 expected goals, which is the lowest winning figure of the round by a distance.
At the other end, Coventry City 0-5 Brighton & Hove Albion was the highest-scoring game on chances as well as goals, at 3.97. Brighton & Hove Albion took five goals from 2.50 expected, and Coventry City created 1.47 without scoring in a game they lost by five.
The most one-sided remains AFC Bournemouth against Brentford: 2.09 expected goals to 0.52, and it finished level. AFC Bournemouth sit fifteenth on three points with 6.9 expected points, the second-largest shortfall in the league after Fulham.
Sunderland against Arsenal is second at 3.93, and Sunderland's 1.79 without scoring made that scoreline the round's most misleading. It was also the leg that carried Saturday's winning Daily Acca, backed as no goals at both ends.
Crystal Palace have conceded 9.8 expected goals in four games, the worst defensive number in the division by a distance, and eleven actual goals to go with it. Nothing about that is bad luck.
Aston Villa have created 3.1 expected goals in four games and scored once. The lowest attacking output in the league, and unchanged from our Gameweek 3 review. One point from four games is a fair reflection.
Tottenham Hotspur have not scored in four games, from 3.7 expected goals. That is a finishing problem rather than a chance-creation collapse, but 0.93 expected goals a game is not a platform either.
Manchester United have the third-best expected goal difference in the league at plus 4.6, behind only Arsenal and Manchester City, and sit thirteenth on four points. Minus 3.4 on expected points is the fourth-worst return in the division. Losing a derby on 0.84 expected goals to 0.77 is how a season like that accumulates.
Season-long form ordering for all twenty is on the Premier League power rankings, and the per-player numbers behind these team totals are on the player stats hub.
Two teams are priced wrongly relative to their performances, and in opposite directions. Hull City will shorten as favourites they have not earned; Fulham will drift as underdogs their chance creation does not support. Neither is a prediction about one match. It is where the market and the numbers currently disagree, and the disagreements are the only thing worth betting.
Our day-by-day slips run at the Daily Acca with a full public ledger, and the round-by-round column is at Premier League betting predictions, which settled Gameweek 4 at 6-2 with Monday night still to play.
18+. Please gamble responsibly - begambleaware.org. Expected goals data via our own import, updated as each round lands.