Arsenal is in an awful place – there is no argument about that. For this piece I’m going to look past contract negotiations, locker room disharmony and public displays of protest in favor of statistics. Statistics that I hope will shed light on the issues my eyes have been telling me are the two biggest deterrents to Arsenal’s success over the past 5+ years:
- A lack of ball recoveries in the center of the pitch – ball winning
- A lack of progressive, incisive passing – direct play
Ball Winning
A gaping void has seemed to exist in the center of the pitch tracing back to the Everton game in December. Second and third balls and 50/50s were dropping in the heart of the pitch and we lacked the fight to consistently win them. For a possession-based team that struggles with an “off the ball” identity, these are massive.
Extra possessions for the opposition mean more opportunities to exploit our team shape and/or break at pace to hit us on the counter attack. We had been losing the ball recovery game for months, as I noted in my article a few weeks back. Let’s see how we’ve fared the last 3 matches and compare them to prior performances:
We miss ball winners like Gilberto.
Ball recoveries the past 3 matches:
Man City 59, Arsenal 68 —- +9
West Ham 52, Arsenal 66 —- +14
Crystal Palace 54 , Arsenal 47 —- -7
Let’s compare these +/- statistics with the prior 12 league games dating back to Everton (a):
-1 (WBA), -16 (LIV), +5 (HULL), -1 (CHEL), +2 (WAT), +6 (BURN), -14 (SWAN), -1 (BOU), +1 (PAL), +11 (WBA), -22 (CITY), +3 (EVE)
The only game Arsenal have won and been in the red in ball recoveries is the 4-0 win against Swansea (a) in January.
What these stats say: There is a direct correlation between ball recoveries and Arsenal success. If we lose the midfield battle, we generally lose the game. Give the opposition more possessions, the greater the chance our off the ball deficiencies can hurt us. It’s why so many great Arsenal minds on social media have been clamoring for 3 in the midfield and/or differing tactical approaches.
I’d recently thought Arsenal had turned a corner after winning ball recovery battle against Man City a few weeks ago. I compared our recent +9 performance to our -22 margin in December. It seemed impressive considering City is a team who has come to embody the phrase: the best form of defense is attack. They regularly deploy 5 midfield bodies behind their striker and have been overrunning opposition in an attempt to initiate play in advanced areas. All that optimism I had was put to bed yesterday as we lost by a -7 margin to Crystal Palace.
Direct Play
I love when Arsenal play direct football with a clear sense of urgency. All too often over the last 5+ years our poor performances have come against teams that sit off us, let us play our possession game, and exploit our shape when we do lose the ball.
It’s common sense really, let the opposition organize behind the ball and it makes the game simpler for them. From an Arsenal standpoint, we no longer have an athletic or stylistic advantage, which limits space to exert our technical brilliance. For the opposition, it slows the game, makes defensive distances easier to manage, and helps get their play makers into space and isolated against our defenders when they get service. Furthermore, these problems are magnified when we play with such little width.
Raise your hand if you like a direct pass.
I collected data that I felt would reveal the nature of Arsenal’s passing recently. I took total forward passes (TFP) and divided that sum into total passes (TP) to determine the percentage of attempted passes that are direct.
| Opponent | Total Forward Passes(TFP)/Total Passes (TP) |
% Forward Passes |
| 4/10 @CP(L 3-1) |
342 / 627 |
55% |
| 4/5 vWHU(W 3-0) |
365 / 680 |
54% |
| 4/2 vCity(D 2-2) |
246 / 405 |
61% |
| 3/18 @WBA(L 3-1) |
372 / 752 |
49% |
| 3/4 @L’Pool(L 3-1) |
236 / 405 |
58% |
| 2/11 vHULL(W 2-0) | 263 / 483 |
54% |
What these stats say: There is a minimal correlation between direct passing and Arsenal success.
The game against West Brom recently showed that Pulis’ side let us have the vast majority of the ball and forced us to play mostly sideways and backwards passes. We were thoroughly outperformed that day and played into their hands. The Crystal Palace game yesterday seemed the same way and was surprised at our 55% rate.
Once again, I was tricked into thinking we had made the necessary tactical adjustments against City recently. There was a clear emphasis to play direct (61% FP). Maybe i’m collecting the wrong data or maybe forward passing percentage is simply too context dependent (ie: game flow, opposition, tactical setup, personnel, etc.). Either way, I’d like to dig deeper in the future and look at a larger sample size of past results to see if forward passing percentage is truly a success indicator.
At the end of the day, stats are just stats. You can take them or leave them. I was pleased to have them somewhat confirm what my eyes have been telling me. The next few months will be a roller coaster ride. Whether big changes at Arsenal happen in the form of a new manager, players, or to the club hierarchy, I just hope what I see on the pitch represents stylistic change.
– Dougie Cazorla
Follow me on Twitter @dfresh10
31 year old based in the U.S. I am an elementary school teacher by day whose passion for teaching is only rivaled by a love for the Arsenal. I release a weekly piece each Tuesday that is oftentimes analytical but sometimes eccentric. Founder of the U.S. supporters group Syracuse Gooners (www.cusegooners.com). I enjoy interacting with any and all Arsenal-based opinions on Twitter. Have a younger brother who chose to support Spurs. Fielding suggestions for ways to disown him.
A couple of thoughts Dougie:
1) Wouldn’t it help with analysis a bit more if we could pinpoint where on the pitch the recoveries took place?
2) Does forward pass necessarily indicate a direct approach? My understanding is, and please correct me if I’m wrong, is that a forward pass is any forward pass and most of ours are probably over 5 yards as we build up. So, not as direct as we’d like perhaps.
Thanks.