The Premier League table is starting to take shape and while results are obviously what matter most, they don't tell the whole story in terms of how teams have been playing so far. Expected goals (xG) tells us how much teams have been dominating games by creating lots of good chances up front and restricting opponents to few, bad ones at the back. While it is still early days and some teams have had more difficult starts than others, Premier League history shows that the more a team dominates games, the more successful they are in the long run. Penalties are excluded because they distort a team's numbers on how threatening or vulnerable they are in general, especially at this early stage. Below is a graphic showing how good teams have been in attack and defence, with the most dominant teams in the top-right corner and the least impressive in the bottom-left corner. The higher up the chart, the more dangerous a team's attack is while the further to the right, the more solid they are at the back.

Arsenal and Manchester City have been the most dominant sides, with the Gunners having the best defence and fourth-best attack and City the joint-best attack and fourth-best defence. High-flying Crystal Palace are the only side to rival City's attacking threat, while Newcastle are living up to their name by being almost as impregnable at the back as Arsenal. Both Manchester United and Chelsea have one of the top five attacks in the league but worryingly also rank in the bottom five defensively. Liverpool's back-to-back defeats following five straight wins gives a more accurate picture of how they have been playing, while Tottenham have needed to be very efficient to make up for their lack of dominance in games. Aston Villa's early-season struggles are no longer to do with finishing - they have now scored six non-penalty goals from an xG of 6.0 - but simply because they have not been playing well.

While promoted sides Leeds and Sunderland have adapted quite well to the top-flight, Burnley have struggled to make the jump so far with both the worst attack and the worst defence. Looking at teams' attacks specifically in the next graphic, United, Arsenal and Liverpool have taken by far the most shots, averaging around 15 per game. But the reason why City and Crystal Palace have the best attacks overall is because they are creating much better chances, averaging around 0.14 xG per shot. What that means is that historically Premier League sides have converted the chances City and Palace are creating at a rate of 14%, compared to 10% of the shots that United, Arsenal and Liverpool are. Brentford are a strange side because while they are averaging just eight shots per game, the quality of those chances has been so high that their attack still ranks mid-table overall.

When we look at how teams are defending, Arsenal and Newcastle have been so impressive because they are giving up both the fewest shots per game and the hardest chances to score on average. But as the graphic shows, there are two teams that stand apart from all the rest for having key defensive problems: Manchester United and Burnley. If you look simply at the number of shots that teams are giving up, Manchester United look good defensively since only Arsenal and Newcastle are facing fewer than their 9.1 per game. But not all shots are the same and Ruben Amorim's side are currently giving up by far the easiest chances to score in the league, averaging 0.16 xG per shot faced. Burnley meanwhile are facing by far the most shots at 18.4 per game, four more than any other side, although thankfully the quality of those chances is not that high.

Finally, while most teams' expected goal difference marries up quite closely to their actual one, it is clear that Tottenham are by far the hottest side in the league in terms of form. Not only have they been the most ruthless side in front of goal by scoring five more goals than expected given the chances they have had, at the other end goalkeeper Guglielmo Vicario has had to be at the top of his game, preventing more goals than any other keeper in the league (2.9). While this is obviously great in the short term, Thomas Frank will want to quickly improve his new club's dominance in games as the former manager of stats-savvy Brentford will likely know that significantly overperforming your xG does not last forever.