The trouble with trying to determine how good Arsenal are
6 min read

The trouble with trying to determine how good Arsenal are

"There are three kinds of lies: lies, damned lies, and statistics." - Benjamin Disraeli

Arsenal are an interesting team. Interesting in probably the bad sense like the "May you live in interesting times" expression.

Depending on what you want to show Arsenal can be anywhere from among the four best teams to among the five worst teams.

Statistics can be a dangerous thing.

I will go through a number of different things but in the end, I come down that Arsenal are roughly the 6th best team, with very little separation between the teams that I think are between 5th and 12th. Given this, the spread for where Arsenal can finish is pretty wide and there isn't anything that we have seen this season that changes that.

Now let's go through things starting with weighted xG (I like using this weighted G + xG because it does help try to capture information that is communicated through goals, rather than just expected goals).

One of the things that this really helps illustrate is the tiers in the League right now.

Elite - There are the three teams in the title race.

Good - West Ham in the chasing "pack".

Medicare Middle - Pretty much everyone and now maybe even Watford

Relegation - Norwich and pre-January transfer window money explosion Newcastle.

Now let's look at things another way, this time just at even game state (tied) and no team up or down a man.

xG difference at even game state and no man advantage. From 

Arsenal look a lot better by this metric and you can tell a story about why this is important. How a team performs when the score is tied is very important, the adage that goals change games is true! Scoring a goal changes a team's win probability massively, so performing well during these moments of matches is very important for points accumulation.

In Arsenal matches, the first goal has come relatively early. The first goal has been in the 22nd, 15th, 7th, 66th, 30th, 12th, none, 8th, 23rd, 5th, 52nd, and 39th minutes. In 8 of the 12 matches, the first goal has come in the first half-hour. This does change the way that teams play (maybe it shouldn't change how Arsenal plays nearly as much as it has but that is another discussion) with teams tending to take less risk.

Looking at how teams are doing at the league level when winning compared to drawing (looking at margins of 1 and 2 goals, to exclude blowouts) teams so far this year are creating 0.34 more xG on a per 90 basis. Arsenal are creating 0.15 less xG on a per 90 basis when winning.

Looking at the other side of the ball, teams winning are allowing 0.2 more xG against on a per 90 basis. Arsenal are slightly better only allowing 0.17 more xG on a per 90 basis when winning.

In totality, teams tend to improve their xG difference when winning, this makes intuitive sense where teams allow a few more marginal shots (why the xG allowed goes up) but with the tradeoff that as teams push forward the team winning creates better chances for a net xG gain. Good teams and Arsenal want to be a good team, tend to continue to be better than the teams they are playing even when winning, rather than going ahead and then holding on.

It is positive that Arsenal are a good team when the game is tied, that will lead to points, but it is troubling that when Arsenal go ahead the numbers aren't in line with the other "good" teams in continuing to perform better than there opponents.

Another area that I have looked at recently is comparing the results this season to what teams did with the corresponding fixtures last year. This isn't a perfect apples to apples comparison because 1) 3 teams are completely different than last year 2) teams have brought in and lost players. It is also better than looking at things from say after 12 matches because you can see a pretty big difference in the strength of schedule faced (Arsenal have currently played the 4th hardest schedule based on my team ratings this year), while also having the same issues above.

Looking at this Arsenal have essentially held on to the same points (-1 technically) but have performed worse to accumulate those points. Looking at things in a little more detail Arsenal have allowed 4 more goals and scored 5 less. From the xG they have allowed 5 more and created 5.1 less.

Here is a detailed look at how Arsenal have performed this year compared to last.

Another interesting thing that I noticed looking at this is that the current run of good form, Arsenal did pick up five points compared to last year. The majority of these matches were also part of Arsenal's good run of form following boxing day last year, with only the Aston Villa match coming before boxing day last year.

If we want to get fun with excluding the second pre-season matches before the transfer window signaled the true start of the season Arsenal have picked up 5 points over last year. Unfortunately, the matches before the transfer window closes do count and Arsenal dropped points that they had picked up last season.

Even with all this, Arsenal are roughly in line with the team that finished with 61 points and given the messy middle 60 points might be enough for 6th place (that is currently what my simulation model projects as the average points total for the 6th place team).

The last thing I wanted point out, and something that really complicates the analysis of at Arsenal, is comparing the teams Arsenal have faced performance vs Arsenal compared to what they have done against other teams.

In attack, Arsenal have created 86% of the expected goals that the teams that they have faced allowed against others (meaning Arsenal are creating less on average than the teams they have faced have allowed against the other teams they have faced). Arsenal have done better 4 times (Brentford, Norwich, Crystal Palace, and Aston Villa), done worse 7 times (Chelsea, Man City, Burnley, Tottenham, Brighton, Leicester, and Liverpool), and been average once (Watford). This ranks 14th overall in the Premier League.

On defense, Arsenal are allowing teams to create 112% of their average expected goals against them (meaning that they are creating on average more against Arsenal than they have against other teams). Arsenal have allowed teams to create more than their season average 6 times (Chelsea, Man City, Burnley, Brighton, Aston Villa, and Liverpool), allowed less than the season average 4 times (Norwich, Crystal Palace, Leicester, and Watford) and allowed the average amount twice (Brentford and Tottenham). This ranks 15h overall in the Premier League.

This is pretty damning of the team for me. Sure maybe teams could be more "up" for playing Arsenal but it is a major concern that with regularity teams have put some of their best numbers of the season against Arsenal.

So where does this leave things for Arsenal? I think this is a young team that is inconsistent, I think Mikel Arteta as a manager can be too conservative at times, I think that Arsenal are roughly above average with some flaws at creating chances without unbalancing other areas of the team.

Where does that mean for where they finish? I think anywhere from 4th to 8th is very realistic (with a real chance of the team finishing 8th to 12th). Arsenal are bunched with the other flawed teams, they have shown some positive things (the teams performance when tied and ability to take a lead) but balanced out by things that cause worry (not showing the same performance advantage at other game states and compared to average performances).

Arsenal have another run of games here through the end of the year that should help shed some light on the team's level, but this is also what I said about the run that just ended and I am just as confused about the actual level of this team now as I was 8 weeks ago.