xGunners: North London Derby stats preview

The season’s first North London Derby is tomorrow, so let’s take a stats preview of the match.

Arsenal Offense against Tottenham Defense:

Arsenal have the second best offensive rating according to my simulation model rated at 140.6 (40% better than league average). Tottenham have the fourth best defense with a rating of 121.2 (21% better than average).

Arsenal are taking 6.5 danger zone shots per game, and 4.5 danger zone shots from feet per game (the danger zone is the six yard box extended to the edge of the 18 yard box, and is from where the majority of goals happen). Tottenham are allowing 3.8 danger zone shots per game and 1.9 danger zone shots from feet per game.

Arsenal are averaging 2.6 big chances per game (second in the league) but converting just 24.1% of those chances into goals, compared to the league average of 42.6%. This is a big reason Arsenal have underperformed their open play xG by nearly four goals this season. Tottenham are allowing just 0.9 big chances per game, second best in the league. Even better for them is that Hugo Lloris is a really good shot stopper who has an overall 75% save rate this season, which puts him in the top 5.

I can see the NLD turning out similar to the Manchester City match, where Arsenal don’t have a high volume of shots but do create a couple really good chances. Thing is, Arsenal will need to convert those chances more than a quarter of the time to have a chance at winning.

Arsenal are averaging 2.5 Passing Value Added per game as a team (second in the league), and it will be important going up against the Tottenham press to be able to effectively progress the ball upfield.

Speaking of that press, it will be interesting to see what Spurs do with it against Arsenal on the road. Against both Dortmund and Real Madrid they really seemed to curtail the press, looking to sit back more than they do in the league.

This also seems to be the case this season in their other games against the “Top 6”. They are averaging 31.1 defensive actions (tackles, fouls and interceptions) per game against all teams, and 35.7 defensive actions per game against the big teams (3 matches, so small sample size caveats apply).

The big difference is both where these actions are taking place and in the number of passes the opponent is able to complete per defensive action (PPDA, a proxy for the amount of pressure a team puts on the opposition). Spurs have averaged 9.2 PPDA this season overall 8.4 PPDA against the bottom 14 teams and 10.3 PPDA against the Top 6, suggesting a slightly less intense press against the better teams.

Perhaps somewhat against the stereotype of a pressing team, Tottenham do not press as high as other well known pressing teams like Manchester City and Liverpool. Where Tottenham press intensely, at least according to the data (I can’t say I have watched every Tottenham match this season), is in the middle third of the field.

Overall, Tottenham average 8.5 PPDA in the middle third, which is the lowest number of passes per defensive action in the Premier League. Against the bottom 14 teams they seem to be even more intense in attacking the opponents in the middle third with a 7.7 PPDA.

Against the Top 6 teams, Tottenham are averaging 10.2 PPVA in the middle third, which is consistent with a slightly less intense or effective press against the better teams. Tottenham not pressing as intensely against Arsenal might actually be beneficial for the Gunners, who don’t have the most press resistant midfield in Granit Xhaka and Aaron Ramsey.

If Arsenal intend to try to control the midfield at in this match instead of sitting deep, this might be a time to bring on a third midfielder over the extra man in defense.

Tottenham Offense vs Arsenal Defense

Arsenal have the seventh best defensive rating according to my simulation model, rated at 104.7 (5% better than League Average). Tottenham have the fourth best offense with a rating of 133.8 (34% better than League Average).

If that seems like there is a lot more red dots than Tottenham’s shot chart, your eyes are not deceiving you. Arsenal concede similar numbers of danger zone shots as Tottenham (3.8 per game for both teams) but when they do give up those shots they are much more likely to be classifed as a “big chance” by OPTA. For the season, Arsenal concede 1.6 big chances per game.

Another worrying stat is save percentage. It should be noted that this is a highly variable stat, and 11 games is a very small sample. With those caveats out of the way, Petr Cech is currently only saving 61.1% of shots on target.

At team level, this is the fifth worst mark in the league. When it comes to the big chances Arsenal concede, he is saving 35.7% of shots on target. That number sounds bad but it actually puts him in the top half of the league, with the average being just 31%.

Switching to Tottenham, they are taking 6.1 danger zone shots per game and 4.3 danger zone shots from feet per game, while creating 1.6 big chances per game.

What is also striking is how many shots Tottenham take from outside the box. It seems that Tottenham players haven’t had the shot quality mantra drilled into them quite as well as Arsenal, and have trouble turning down a speculative shot in favor of a higher quality one.

While last year Tottenham excelled at getting these long shots on target, this season they have reverted back toward league average in the percentage that they are getting on target. This year 21.8% of outside the box shots have been on target for Tottenham vs 24.4% for league average compared to 30.5% on target last season.

Tottenham, aka The Harry Kane Team

Controversially, about a month ago Pep Guardiola refereed to Tottenham as the “Harry Kane Team” and caused a mini scandal and some hurt feelings. While this is an exaggeration because they obviously have other really good attacking players in Dele Alli and Christen Eriksen, I do want to highlight how much of their offense flows through him.

The above image illustrates how much non-penalty xG each player on Tottenham has produced this season. Harry Kane has accounted for 36% of the 18.5 xG Tottenham have produced this season and 42% of the 19 goals they have scored.

With Kane’s goal scoring record, for my own sanity I would like to see Arsenal’s defense suffocate him and try to make one of the other players hurt Arsenal.

Simulated Match Odds

I have built a simulation model, and I like to use it to give an unbiased appraisal of things before matches. Here is what my model produced before the North London Derby:

Overall, it looks like a very even match up with a slight edge toward Arsenal (due to home advantage, at a neutral site this is rated as completely even). If you are interested in the methodology of the model you can find that on my personal blog.

It is a very early kick off for me on the West Coast of the United States (4:30am), so I really hope Arsenal don’t make me wake up that early in vain.