Mailbag 8.26.21
18 min read

Mailbag 8.26.21

Mailbag 8.26.21

The request for questions was originally for filling out a podcast so I didn't have to overreact to the first two matches of the Arsenal season which I don't think will be overly reflective of the season for Arsenal. It made for a very good podcast which you can listen to here:

I also thought that, seeing as I had already written out a lot of my answers it would be wasteful to not post them here, plus some additional graphics that wouldn't translate to the audio platform.

Arsenal are often criticized for going wide to Tierney and crossing. How does this compare to the way City for example attack? – Raj @Clinical_Helb

This is a great question and one that will be helped by having graphics here.

If you look at where Arsenal creates key passes and add to their probability of scoring the hot spot is in on the left edge of the box.

The other teams are a bit more balanced, Chelsea creates quite a bit from the central spaces but creates the most threat from the final third-half spaces.

Liverpool is similar but tends to skew more to the right side with Trent Alexander-Arnold.

Manchester City creates the most from the central zones but has more threat in the right half-space, which is where Kevin de Bruyne likes to operate.

I don’t think that it is surprising or bad to use the width for Arsenal, I do think that ideally, things will need to be able to find more balance so that isn’t the only way that the team creates chances.

A bit outside of your normal area, but interested to see if during the Arteta era that certain formations have / haven’t resulted in better results from a data perspective. Offensively and Defensively. - Josh Lang @joshfredlang

I have the data from last year:

343: 1.25 xG, 0.99 NSxg 30.4 deep touches, 9.6 shots, 12.2 PPDA, 74.5 Possessions – 1.33 xG A, 1.23 NSxG A, 36.9 Deep Touches A, 12.3 shots A, being played 10 times

4231: 1.58, 1.3 NSxG, 37.1 deep touches, 13.4 shots, 12.5 PPDA, 77.8 Possessions – 1.21 xG A, 0.88 NSxG, 26.8 Deep Touches A, 10.9 Shots A, Played 26 times

Can you analyze Auba's data when he plays with and without Lacazette? For example, xG/90 and maybe something like average position/touch. I'm curious if playing with Laca means that Auba gets less final third touches or box touches or quality shots. - djbowen3

With Laca, 1148 minutes (12.8 90s), 3 goals 0.23, 2.9 xg 0.22, 21 shots 1.64, 11 key passes 0.86, 0.86 xA 0.07, 1 assist 0.08, 74 prog passes rec 5.8 66 deep touches 5.2, 1.88 AVA 0.15

Without Laca, 1212 minutes (13.5), 8 goals 0.59, 7.8 xG 0.58, 37 shots 2.74, 11 key passes 0.82, 1.58 xA 0.12, 2 assists 0.15, 83 prog pass rec 6.15, 94 deep touches 6.96, 3.34 AVA 0.25

Lots of this is also the argument for why Auba should be primarily CF in my opinion.

Do we have statistics that assess style of play? Say possession vs. counterattack? If we can quantify this (or not): Why do you think we've come to identify Arsenal as a possession team? Has Arsenal had success as a possession team? Any non-cup success? - patos heridos @mcd410x

Yeah there are things that you can look at, I’d generally put teams into slow/defensive possession, direct possession, proactive counter-attack, and passive counter-attack.

Each of these would come with stats that would describe the teams inside and some variation around a general theme.

So why is Arsenal a possession team? I think mostly because Arteta wants them to be, and we have a squad that is kind of built to play that way. I think that you can see a team inside of Arsenal that kind fo straddles different categories. I think that Auba and Pepe would thrive in a more direct or counter-attacking team, I think that if Arsenal leaned into being able to be a full transition team that would be better for them.

I think that Xhaka is a perfect slow possession midfielder and Partey probably fits better in the transition-type team.

Our defensive line before really seemed to work well in a deep block that is more passive and just tries to keep teams out of dangerous spots when out of possession. I think the Ben White is a nod towards perhaps being able to do defending higher to allow for faster transitions.

Arsenal has had success with possession and with their deep buildup play as a tool to create space to attack, however that is heavily rehearsed and I think teams were able to pick up on how to frustrate those moves. I think that we saw a bit more success when we will still having more of the ball but also had a bit more freedom to attack in less structured ways. Like everything in life it is a bit of trade-offs, the slow methodical possession was boring but also protected our defense effectively. The first early signs of trying to play a bit more proactive have not had the best results for the defense and have not produced any additional attack. It will be about finding the right balance.

Seen a lot of people saying teams like West Ham and Leicester play better football than us. Since style of play can’t be quantified, is it the case that they do the same things we do, but just look better doing it. Or is there an actual gap developing between them and Arsenal? - Siddhant Bhatia @sbskelpie

I think style can be quantified and described, it is hard to say a value judgment beyond does this produces results. We all have our own opinions on that, but even within that, I think that there is a certain style that people do generally favor in seeing from their teams.

From that I think that people like to see teams that produce more shots, I do think that people like fast direct attacks, people like dribbles, generally I think people dislike slow plodding possession.

On these Leicester and Arsenal look very similar, they take a similar number of shots, similar direct shots, slightly more dribbles, at a slightly faster pace.

West Ham play a different style, where they cede possession generally and play more directly, with correspondingly more direct shots but less shots overall, and an overall faster pace.

On the question of doing it better, I don’t see a huge gap right now but it is very close. In my current rating system, the teams are all right with each other and I would put them all in the same tier rankings-wise as the teams fighting for Europa League.

Will we ever be good again? – Martyn @MartynHadADream

I feel like I could go a couple of different ways with this. On the one hand, in the total population of soccer teams, Arsenal is good!

In FiveThirtyEight’s global soccer rating Arsenal are the 20thbest team in the world and rank 16th in the Club Elo Ratings. Based on this Arsenal are in the top echelon of teams, it is just that the Premier League is also stocked with better teams.

Within the Premier League Arsenal are around 6th best team, I think that you could make a case for anywhere between 5thand 8th and that’s really why I like to think in tiers rather than hard rankings.

So if you think of good as being back in the top handful of teams in the Premier League where challenging for a title is possible or maybe even the top 10 of Europe, I think that is at best 3 seasons away likely a bit more.

I think this transfer window shows that there is an acceptance within the club that getting back into the Champions League is not something that can be done with a quick fix of one or two veteran signings, and that the overall talent level of the team is a step below the teams they are chasing. Looking at this strategy I think the five-year plan is to look for getting back into Europe this season, so 7th or better, a season to consolidate at this level and work at getting to 5th. 3rd year to challenge for Champions League and then in year 4 that is where we are ready to solidify things.

This is something that would take astute buying of players that really compliment the core of the team while continuing to supplement depth with players that can push the players in the first eleven. When we get to that 3rd or 4th season that is where it might make sense to look at a couple “win now” veteran signings on a value deal.

So to wrap things up, 5 years from now is when Arsenal will be good again, unfortunately, this might also be the answer when you ask again in 2 seasons. Having 3 teams significantly richer than you and then 2 others equal to more depending on European money makes things really hard.

following on from the Athletic Podcast, what are Arsenal actually good at? - Jamie Robertson @jamielrobertson

Slow boring defensive possession. This is what Arsenal is good at.

Also, I would say that Arsenal is good at attacking high lines with minimal pressure on the midfield passers.

What will happen next, Arsenal get relegated or arsenal win the league? - Saturdays on Couch @SaturdayOnCouch

I really badly want to say relegation and lean into the negativity but in reality, it is probably the title but with none of the above the more likely option for the foreseeable future. This all comes down to Arsenal’s revenue gap over the bottom teams just giving such a high floor for the team. I really do think what we have seen the last couple of years is Arsenal banging on that, sure they could finish in that 10 -14 range but that would really take quite a huge underperformance for the talent level the team's revenue can buy. I am not hopeful about winning a title but the pathway of hitting on a few buys, complementing with some big money additions and a season of positive injury luck and maybe some strong finishing, and Arsenal could win a title. The opposite situation for me would require the team to go through a financial crisis of some sort where they couldn’t use the built-in revenue advantage for some reason.

At what does the decline spiral become critical? Poor results lead to lower league position, failure to qualify, less income from comps, sponsors, fans, leads to difficulty recruiting/retaining quality players, leads to worse results. - John Batty @techbatty

This is an interesting one and I think it depends a lot on the team. For me, the biggest issue is with how league finish affects revenue generation and how willing an owner might be to make up shortfalls. I think that generally just by the virtue of being in the premier league Arsenal guarantee that they will be one of the richest teams in world football. I will speculate and say that for the short term I don’t imagine continued massive increases in the tv deal but I do think that the separation between the premier league and the rest of the leagues will only grow. I imagine that on the commercial side there would be some hit to not being in Europe but what I imagine is more commercial revenues stagnate vs grow, we have already seen a cycle of the major deals being re-done and the drop out of the champions league probably hurt but wasn’t massive, if that happens again perhaps it is another where Arsenal still do okay but the gap with the teams above opens but they still remain above the bottom part of the league. Overall this is a team with a big fan base, in a very rich city and that will help cushion the blow for a lot longer than many other teams. Players will want to play in the Premier League, London is a draw, plus Arsenal will be able to pay wages that can only be matched or beat by a handful of continental teams (when things go back to normal) plus City, Chelsea, and United, playing in the best competitions matters but so does getting paid, so as long as arsenal have a revenue advantage they should be able to attract players but maybe not always the top handful.

What team metrics most need to improve (and what level do they need to reach) in order for Arsenal to have a statistically-probable chance of finishing top 4? - Rambling Gooner @rambling_gooner

The easy answer is goals. In the Emirates era through the last time Arsenal finished top 4 (2015/16) Arsenal averaged 71 goals a season, last two seasons Arsenal have been at 55 and 56 goals, so scoring more would go a long way towards getting back there.

The slightly more complicated answer is that they need to balance the tradeoff in attack and defense. In the first half of Arsene’s time as manager, Arsenal was pretty consistently in the 40+ GD range, with goal-scoring rates that were good but not great, especially by modern standard but those teams did really well at keeping the ball out of the net enough to finish in the top spots of the table. If Arsenal got back into that range, they would start putting together title challenges again.

So if Arsenal do really want to get back into the top 4, getting back to goal differences of 30 or better is a good goal to shoot for. The other thing to remember is that scoring is generally better than not conceding, if you want to maximize points you do it through scoring rather than limiting goals, variance is a killer and if you are a low scoring team too often that will turn into 0 goals scored and that comes with a maximum of 1 point, and the more goals a team scores the higher probability they turn 1 point matches into 3 point matches.

Plus it is just more fun.

Is there any way to use location tracking stats to measure the average number of players we have in the box when the ball is in the final third (vis-a-vis other teams)? My general sense is that we simply do not get enough players into the box, or perhaps quickly enough. - Sam Handlin @shandlin

Yes, this would be a great use of tracking data because I too have the suspicion that for a lot of the crosses that Arsenal attempts there are not that many players in the box. Unfortunately tracking data is not public and I am not nearly advanced enough to use computer vision to do my own tracking via match broadcasts.

Are there any stats that are similar to box plus minus in Basketball? Maybe like plus minus on expected goal differential? If so, is it useful or telling? - Cat In the Wall!!! @CAT_IN_THE_WALL

Yes there are.

These are called with and without stats. FBRef posts them without any adjustments for player quality in the playing time section of the website. They have both goals, goal difference, xG, and xG difference.

One of the things that makes this hard is that in soccer, there is not a lot of scoring and the scoring that does happen is a bit down to chance. I think looking at shots or even better-expected goals gives a bigger sample to compare against and is better. The other thing that makes this really hard is that there just aren’t that many substitutions and therefore combinations that you can use to tease out impacts. If they stuck with the 5 subs rule maybe you would be able to see more but with the current rules, a lot of things are just the team stats.

There are people that have tried to take this and make adjustments for these issues, the oldest is GoalImpact and a newer one that I have seen and one that I really like by Hugh Klein called Macro-Football (he would make a good guest for this podcast IMO) where his model uses xG and does adjustments for things like score, manager, and teammates.

I am still skeptical to a certain degree again because of what I said above but it is an intriguing course of study for me.

Can you take a look at how we compare to other sides physically? Heights, speed, etc as I feel we are often getting bullied - Ereek Ronk

I don’t have anything particular on speed metrics (I really wish I did) I do have the listed heights and weights that are posted on sites like FBref and WhoScored. What I have on these is that so far this season Arsenal has the 3rd tallest team (adjusted for minutes played) but only the 13th heaviest team. Arsenal average 5’11” (182 cm) and 165 pounds (75kg) so far this season. The biggest teams (average of height and weight rank) are Chelsea, Tottenham, and West Ham

Stats you've come across/calculated that really surprised you. Where do you see the world of stats in football progressing to? Was there any statistical anomalies in the invincible season? - Mark Bløndal @markbloendal

One of the things that I think is really interesting is the stats that try to capture the value of on-ball events. I think that they are the logical next step for event data, this is something that I have been interested in and sort of working on since really the 16/17 season, with my first stab at coming in 2017. I think that these stats can be interesting and surprising, so much of what happens on a pitch you forget about and there are a lot of really valuable actions that a player did that just didn’t turn into end product or a few mistakes that are super memorable that either stick with us or get lost in the fog of memory.

My initial reasoning was trying to help illustrate that the stuff Alexis did around the box was well worth the risk of losing the ball, hint it was. From there I have refined things further where I am pretty happy with how things work to measure attacking contribution. The next steps are things like American Soccer Analysis’ Goals+ metric, which tries to also include defense (I am not 100% sold for reasons of just what defense is) and StatsBomb’s OBV. I think that OBV with the StatsBomb 360 data will be really interesting, we may never see more than glimpses (or more in the discord) but that really seems like the cutting edge of what things are doing.

On the Invincible’s, unfortunately, that is before Opta started collecting things, StatsBomb did go back and re-tag most of the matches but I have not had a chance to really look into it. It is one of the things on my list to go back and dig into.

I think you did this stat before but I can't remember? Is there a stat that tracks how fast we move into the final third compared to other teams? – Jay @jaybert84

This is a new stat I am tracking called pace towards goal. What it looks at is how fast a team moves up the field with the ball when they are in control of the ball. Right now Arsenal is a slow team but after doing some counterattacking against Chelsea are the only 13th slowest so far this season. I haven’t gone back and done this for last season's stats but I remember talking with Carlon Carpenter in our podcast that Arsenal was second slowest in the similarity measure for StatsBomb.

I have also started looking at the number of final third passes per shot, basically as a way to look again at directness. Arsenal currently has the 18th most passes per shot with just under 17. The teams that are worse are two poor attacking teams in Southampton and Crystal Palace, but with Chelsea and Man City in the similar range and just slightly fewer.

This is more a measure of style than saying something is good or bad.

Do you think Emile Smith Rowe could play CM beside Partey and behind Odegaard in home games against teams that sit deep? He seems so good in tight areas playing quick short passes. What's his defensive numbers like against some of the more attacking 8s in the league? - Ricky Dobbs

I went and pulled his radar for last season and it actually looked surprisingly good on a midfield radar, much better than it actually did on the attacking mid template. I am far from a scout but my view is that I want him closer to goal, with the ability to come back and get the ball when needed. The big question is always would he be able to help enough defensively to be passable as a midfielder and I think that is still a question for me. I would hate to take him out of a role that he seems to really be settling into.

Why doesn't ESR show good metrics? He looks like he is always doing something good Has there been in a change in that area this season (although its a very small sample size atm) ? - Ganesh Bulusu @GBulusu

This is an interesting question, I think that on certain things he rates well. On the goal probability added type stats he rates very well, on the on/off type stats he looks elite, he is a very secure passer while also completing passes into dangerous areas at a good rate.

Where he doesn’t rate as well is on the end product stats, his shots and xG are lacking, he didn’t create many chances for teammates or xA. He was also not especially good with the ball at his feet.

So far this season he is looking amazing but yes it is very small sample but also a continuation of what we saw in pre-season and when he was in the championship. I am excited that the valent and good decisions will translate into some of the other stats this year.

Do you think we need another CM.. come the AFCON we'll be relying on Xhaka, Lokonga and Azeez for the base of midfield. Or is this a potential January signing - Sam Whittingham @ElRimmo

My answer is a resounding yes, I still see this as a major weakness of the team. We are awfully close to seeing Mohamed Elneny get a good chunk of minutes with this team. Arsenal hasn’t really changed too much from the team where he played 1500 minutes and made 17 starts.

Really all we have done is change Dani Ceballos with Sambi Lokonga, I think that is a great addition but I still look at that depth chart and think that is not good enough. I imagine it is still a lot to do with getting players out but I would have really liked to see someone else come in so Xhaka isn’t first choice at least Elneny is not 4th. Right now improvement in midfield comes down to getting a lot from Partey and I really hope that he can give the team 30 starts because I think that is needed to come close to our goals

if you can see all this data, and the problems shown, how come Arteta and co-can’t? - m shepherd @marcsenal

I am sure that they can.

Everything that I have ever heard about Arsenal’s StatsDNA is that they are top-notch and on the bleeding edge of things. I don’t think that there is anything that we can see that the club isn’t aware of. The biggest thing is that so much of what is going on with Arsenal is the balancing of tradeoffs. I agree that there are times I am frustrated because I think that we are in a transition of sorts between styles and the types of players that we have.

Should feel good about Lokonga, right? - Paul Elliott Johnson @RhetoricPJ

Yes, I think so.

He has shown a lot of the good things that I think Arsenal was after when they signed him, while the growing pains that come with playing a young player are also there.

The thing that has jumped out for me is that his tendency for verticality from the Jupiler Pro League has translated to his first couple of matches with Arsenal. He is passing a bit more at Arsenal but playing a very similar mix of passes, while also doing so more efficiently than he did at Anderlecht.

On slightly less rosy terms I think that you can see the issues that he has with defensive positioning that was mentioned when he was signed have popped up. He can get drawn to the ball, or stay on the player he is marked for too long that it destabilizes the team's shape.

I think that for the most part, we will all be relatively happy to let some of the growing pains on defense go, to get that sweet sweet progressive passing.

I would have said that he had the August player of the month wrapped up but then Arsenal went and put a 6 spot on West Brom’s U23s and that might sway things towards someone who played well there.