Saturday, 1 January 2011

Mid-season analysis: What's wrong with Arsenal's central defence?

We're exactly half way through the Premier League season and after another 2 cheap points were dropped at Wigan recently, it's worth exploring what's going wrong at the heart of Arsenal's defence, especially when it's occupied by the French pair of Laurent Koscielny and Sebastian Squillaci. The goalkeeper and left back situations have been scrutinised in great detail by Arsenal fans but the purpose of this article will be to examine the centre backs alone through a combination of statistical and video analysis.

For the sake of a fair comparison, Champions League and League Cup games will be ignored due to the weakened nature of the teams fielded. Tackles and interceptions made by centre backs in the final 3rd of the pitch will be excluded from the stats as these generally only occur during corner kick situations and do not reflect defensive ability. The 3rd goal conceded against West Brom in the 2-3 home defeat will also be ignored as the team was playing a 3-man defence from the 66th minute onwards (tackles/interceptions/shots during that time will also be ignored).

Partnerships

Firstly, let's take a look at the defensive efficiency of the 5 different centre backs pairings that have been fielded during the Premier League campaign so far:

Table 1: Centre back statistics for all 19 Premier League games to date

The Djourou-Koscielny and Song-Vermaelen partnerships have had little game time so can be ignored from this comparison. With that in mind the key points are that:

-The Koscielny-Squillaci pairing is the leakiest and concedes goals at almost 4 times the rate of the most secure pairing, Djourou-Squillaci.
-Despite conceding the least goals, the Djourou-Squillaci partnership actually concedes the most total shots and the 2nd most shots on target.
-The Koscielny-Squillaci partnership concedes fewer total shots than any other partnership, but the most shots on target.
-The Koscielny-Vermaelen pairing concedes the fewest total shots and the fewest shots on target.

These stats are for all Premier League games and so do not take into account the differences between home games and away games (where defences generally come under extra attacking pressure). Therefore here is a further breakdown of these stats beginning with home games:

Table 2: Centre back statistics for all 9 Premier League home games to date

From this table the only 2 partnerships worth comparing are Djourou-Squillaci and Koscielny-Squillaci. Again, key points are that:

-The Djourou-Squillaci pairing is the most secure and concedes goals at almost 3 times a lower rate than Koscielny-Squillaci.
-The Djourou-Squillaci partnership concedes more total shots and more shots on target than Koscielny-Squillaci.

And finally the away performances:

Table 3: Centre back statistics for all 10 Premier League away games to date

Key points here are that:

-Djourou-Squillaci once again emerges as the most secure pairing, despite conceding the most total shots.
-Koscielny-Squillaci is the least secure and concedes the most shots on target (as opposed to the least in home performances).
-Koscielny-Vermaelen concedes the least total shots and the least shots on target.

With the exception of shots on target conceded by Koscielny-Squillaci, the pattern of home and away games is quite similar.

Individuals

So what is the reason behind these trends? To understand we need to take a closer look at the strengths and weaknesses of each individual centre back, so that we can recognise the dynamics of each separate pairing. Here are individual stats for interceptions, aerial duels and tackling:

Table 4: Individual defensive stats for the 4 centre backs

The points we can draw from this are that:

-Koscielny is the most pro-active defender, attempting the most tackles and aerial duels, and the 2nd most interceptions. His tackling is by far the tidiest.
-Squillaci prefers to keep his position rather than close down, attempting fewer interceptions, aerial duels and tackles.
-Vermaelen and Djourou are somewhere in between these 2 varying styles, Vermaelen tending more towards Koscielny's game while Djourou is more akin to Squillaci.
-Despite attempting the most aerial duels, Koscielny is joint weakest in this area along with Squillaci (who attempts by far the fewest).
-Vermaelen is the strongest in the air, with Djourou not far behind.

Conclusion:

With this knowledge, we can roughly rank the centre backs in 2 important areas of the game:

Closing down:
Koscielny = Vermaelen > Djourou = Squillaci

Aerial ability:
VermaelenDjourou > Koscielny = Squillaci

This helps us to better understand the dynamics of each centre back pairing:

-Koscielny-Vermaelen concedes the fewest shots on target because it closes down the best and thwarts attacks before they develop (with a larger sample size it could possibly rank the lowest in terms of total shots too).
-The Djourou-Squillaci partnership concedes the most total shots because it doesn't close down (it is joint worst for shots on target too).
-The Koscielny-Squillaci pairing is the leakiest because it is weak in the air and therefore vulnerable from aerial balls and set pieces.
-The Djourou-Squillaci pairing is the strongest because it is deep-lying and doesn't risk being caught high up the pitch. Instead it looks to soak up opposition pressure, taking advantage of Djourou's aerial ability.

The major anomaly within these stats is that the Koscielny-Squillaci partnership concedes few total shots (due mainly to Koscielny's closing down?) but concedes the most shots on target and indeed the most goals (by a large distance). If the partnership restricts opposition shots so much, why then do a high proportion of these end up on target and in the net?

Comparing the Squillaci-Djourou and Squillaci-Koscielny partnerships, it's difficult to comprehend that the vast difference in defensive efficiency is down to a simple difference in defensive strategy. The fact that replacing Johan Djourou with Laurent Koscielny results in an almost 4-fold increase in goals being conceded, should be cause for concern, and points the finger of blame at the young Frenchman.

So what exactly is Koscielny doing wrong? It could be that he struggles on the left side of defence when partnered with Squillaci, as opposed to the more natural right-sided role he assumes when partnered with Vermaelen. Another factor may be that without the comfort of a Premier League-experienced partner, Koscielny lacks the right support and leadership to effectively function within this defence.

Another more simple explanation would be that Laurent Koscielny is simply error-prone, and that despite all his good work in getting tight to opponents and winning the ball back early, he lets it all go to waste with one or two moments of madness each match, whether it's a positional or technical error.

As usual, this pattern is best demonstrated with actual graphic evidence from the matches. The following video breaks down the 21 goals conceded in the Premier League campaign so far (excluding the 3rd against West Brom), and we can observe that indeed the main problems of the Squillaci-Koscielny partnership are its aerial weakness and Koscielny's tendency to make costly mistakes:


Solution:

So what can be done to remedy the current situation? The obvious decision would be to drop Koscielny in the short term and persist with the Squillaci-Djourou partnership, certainly in away games anyway. At home Koscielny's style of play would be beneficial towards Arsenal's pressing game and he can compete with Squillaci to partner the excellent Djourou (and when fit, Vermaelen).

In the long term, with more Premier League and general defensive experience, Koscielny should mature into a fine modern centre back, combining his excellent technical qualities and intelligence with improved concentration and reliability. It will be of great interest to see which two of Djourou, Koscielny and Vermaelen will establish themselves as the club's first choice centre backs given their similar ages and differing qualities.

Comments (44)

Loading... Logging you in...
  • Logged in as
lol like 15 goals were all kos fault. Djourou a must 2day
@iamozgooner's avatar

@iamozgooner · 742 weeks ago

Well now we all know, what I wonder is how much trust does Wenger put into stats as opposed to his own gut feeling / judgement. He must have this data at his finger tips! My gut tells me TV & JD are Arsenal quality Koz maybe / maybe not and Squilly...well he's another Cygan! I think a new center back should be signed in January that can step in straight away until TV is back and partner JD, then choose the best pairings from the 3. What are your thoughts about Samba's ability? Or Cahill from Bolton? Those are the 2 that appeal to me and for the right price will be available, also available for CL games. Happy New Year.
1 reply · active less than 1 minute ago
Kos is not good at CB position. He may be a better LB bcoz he's good at tackling and have a good recovery speed.
Great work here - but (and it pains me to say it) - I think you credit Koscielny with a few too many errors. For instance, on the 2nd Chelsea away goal, I distinctly recall Nasri just passing the ball to Anelka, putting him 1 vs 1 against Kos. If Anelka had beaten Kos there, it probably would have been a 1 on 1 against Fabianski, as the other CB was too far away to intercept. I think he could have been let off the hook on that one - he 'took one for the team' as the cliche goes.

On the other hand, the CBs were at fault for the Newcastle goal. How does the biggest aerial threat on the field get an unimpeded run on goal at a set piece. They left Fabianski out to dry on that one.

But again, thanks for all your lovely work and analysis. It's appreciated. Happy New Year from the States...
1 reply · active 742 weeks ago
Awesome Analysis!

Just one point, for Goal 15 Villa Away I don't don't think it's entirely Squillaci's fault(poor header), I think Clichy backed out at the last moment making that extra space for the goalscorer.
1 reply · active less than 1 minute ago
great article mate. waiting for more of your videos. happy new year!
1 reply · active less than 1 minute ago
Great work and done so quick
1 reply · active less than 1 minute ago
Would the kos Squillaci work if one of them could head the ball? as have contrasting styles
1 reply · active less than 1 minute ago
Hi, anyone knows how I could forced all Arsenal blogsphere especially Le-grove to read this page before writing another piece of trash in their own website?? damnit!!! Mr Renoog, you just rocks!! I can start to be an expert in football~~ Do more more more like this please!!! Have a great 2011 ahead~~
1 reply · active less than 1 minute ago
Great analysis.
Interesting about Djourou and Squillaci allowing the most shots but the least amount of goals.
I thinking back to some of the games they've played in, Man City and Wolves, without some excellent keeping it could've been a different story, but, although Fabianski made several excellent saves, they weren't impossible for a decent shot stopper. What I'm trying to get here is that Djoruou and Squillaci concede shots that aren't hard to save, but that Koscielny and Squillaci, because of aerial weaknesses and errors, concede shots from close range that are hard to save.
Or maybe I'm overanalyzing.
1 reply · active 742 weeks ago
What an article mate, brilliant analysis. You just conjured up all of our strength and frailties in a subtle way, stats do not lie. Good luck and cheers!
Nice article, but I hope you understand that your sample size is incredibly small and very prone to variance. Kos definitely has some things to work on, as does our entire back line. Interesting read though.
7 replies · active 742 weeks ago
@iamozgooner's avatar

@iamozgooner · 742 weeks ago

I read on an earlier blog reply that you are not involved in coaching, this amazes me considering the knowledge you show. I am sure there are teams that would love your help even on a 'fun' p/t basis maybe with select level teens, from my limited coaching experience at this level a person like you interacting with the coaches and players would be a huge plus. Happy new year.
1 reply · active less than 1 minute ago
Excellent blog piece, you never fail to disappoint with your analysis!

Thoughts on Koscielny, and specifically the Koscielny/Squillaci partnership. I think that although Squillaci has played at the top level for some time, both of them are new to the league, and this is a huge leap for Koscielny. It seems that he isn't getting the guidance he needs. In 4 games with Vermaelen or Djourou he's let in 3 goals, with Squillaci he's let in 16, so I think there's a definite pattern there.
1 reply · active 742 weeks ago
I was watching the Birmingham game and everyone knows that Gunners let in ZERO goals. What was obvious is that Djourou actually played the game and the defence looks rather solid (out of the 'gut' feeling)... somehow, after reading this entry, it appears that Djourou has something in his game that causes Arsenal to let in fewer goals. I suppose its his positioning sense.

But still, Djourou had lost out on 2 aerial battles and one caused Birmingham to hit the crossbar from a corner..(he did well otherwise).. When I watched Rio-Vedic, they lost very few headings and deal with most of the shots well. That has finally convince me that perhaps Wenger do need a 'world class' CB. It still depends on Kyle Bartly if he can fill in the role in the next 3 year.

- just mho, Djourou is extremely good, but i doubt his Aerial game can improve much. He will make an excellence 3rd choice..
- Koscielny is extremely good in his tackles, one for the future certainly. I do like his game, but clearly lacks experience and Wenger is just the person to guide him.

That leaves Squillaci. He makes excellence (and expensive) 4th choice.. He lost out in many one-to-one battles against BPL strikers and that proofs La Liga is pure crap.. Judging from his age, I doubt he could improve much.. Should a new defender arrives, perhaps he will be the one to leave.

Wenger knows..
2 replies · active less than 1 minute ago
As others have mentioned, it's difficult to draw conclusions from this rather small sample size especially regarding goals conceded, as goals are relatively rare events. Six of the goals scored vs the Squil-ceilny partnership were against WBA and Tottenham, two games I'm sure most agree were lost more as a result of total team complacency and lack of effort. WBA was an especially abject performance for the whole Arsenal team, with a striking lack of midfield cover allowing WBA's attackers to run at our back four constantly. With Vermalaen injured and Wenger bringing Djourou back slowly, Squi-cielny have been the de facto starters and they've been involved in all five of our losses. However, my recollection of those losses was that defensively, lack of midfield cover was also a major issue. I don't mean to minimize your efforts in bringing these numbers to attention but just to point out that multiple factors come into play. I think much of our defensive problems stem from the midfield, specifically the adjustment to Song's new role in making forward runs. It places the responsibility of maintaining proper spacing between the midfield and back four on the other midfielders. Against Chelski, that balance was spot on with Song scoring the opening goal but also making a great contribution defensively. In other games, however, we've struggled and exposed our defenders to too much pressure. In the games featuring Djour-laci, all victories I believe, the midfield balance was better for the most part plus against Wolves, Fabianski made several great saves to keep a clean sheet. I agree that Squil-cielny are probably our weakest pairing but they've also been involved in games where they've been exposed due to lack of cover. For example for the Blackburn goal, Koz was muscled off the ball too easily however it was also a situation in which our midfielders got caught too high up the pitch leaving our defenders exposed to a 3 v 3 break, which is always a good situation for the offense since it leaves the defense without an extra man.

I think you've pointed out a key factor in why I agree that Squil-cielny is the worst partnership and that is their relatively poor heading. Correct me if I'm wrong but I've read somewhere that Arsenal have the highest percentage of goals conceded vs set pieces in the league you've provided pretty strong evidence as to why. Plus, as you've pointed out, Squillaci is a defender who prefers to stay back and let things unfold in front of him. I think he'd look more solid if he played for a team which didn't hold a high line because he doesn't defend well in space and he's uncomfortable in 1 v 1 situations. With Koz aggressively attacking the ball, looking to intercept passes and pressing the opposition he's often the last defender, a position with which he's distinctly uncomfortable. He's less able to "clean up" after Koz's mistakes. With Djourou, Squillaci isn't exposed as often since Djourou isn't as aggressive. With Koz, Djourou is also a better partner because he's a more able one on one defender plus he's more comfortable with a lot of space behind him. Currently, I think Djour-cielny is the best partnership since it combines Koz's aggressive play with a partner who's more able to stay back and defend in space plus his aerial ability gives us more security at set pieces. When TV5 comes back, I would prefer Djourou to partner him.
Brilliant article. Really good analysis of our defensive problems, though the reluctance of some midfielders in tracking runners often leads to the defenders being forced to make a higher risk challenge. Sad to see so many goals being conceded aerially from set-pieces.
Tee Song's analysis makes much more sense. Yours is too simplistic and ignores the role of the rest of the team. We lose when the whole team plays badly and does not defend from the front allowing too much pressure on our defence. Your approach just reinforces stereotypes and opens individuals to even more scapegoating. The fact that Squillaci and Koscielny are new to the PL, the pace of the game and the liberal refereeing, which allows certain tactics that would be a foul elsewhere, is also something that needs to be considered.
goonerbegood's avatar

goonerbegood · 742 weeks ago

WHY FOCUS ON THE NEGATIVES AND NOT THE POSITIVES

Verm the machine has been out for a while now due to injury AND WE MISS HIM.

Djourou is coming back from a long term injury,COMING BACK POSITIVELY.

Kos,is our No 6.the boy is quality..first season in the BPL,HAS MADE LOTS OF INTERCEPTIONS,brillant plays,even got a goal to his name on his first season,which some take them quite a while e.g TH14,Kos Needs to let himself free as our THI4 in defence.(the day henry let himself free,he discovered himself and he went on to be our highest goal scorer till date.LEGENDRY.)
Kos has too much respect for the strikers,he has a better read to the game as most defenders ,he has to start using that as his strenght.ofcourse he will make mistakes,am sure he rues THOSE mistakes and he will get even better as the season reels in.rutheless but intelligently.E.g Vidic.BUT EVEN VIDIC makes the odd mistakes BUT due to his better understanding of the BPL,he has helped MAN U to remain UNBEATEN SO FAR.DEEP INTO THE SEAON-

Squid:Since he came ,he has been exceptional ,needs to be more ruthless with strikers but intelligently.

AT THIS OUR DEFENCE WILL GET THE NIK SOMETIMES,this is not our weapon.

Our weapon is to win the league COLLECTIVELY AS A TEAM.the goals against us against wesbrom,the eir goals it was a collective mistake,clichy should not have allowed that btw his leg flip pass from opponent.,the defence needs more protection and it begin from the goalie.FABIANSKI needs to be commanding more his goal post,Scezny waiting his time.W e need collective effort as FANS , as CLUB ,as a TEAM(like my boss said today,it takes collective effort to one single job.everybody have to be on the same page,this is what champions are made of.)

Last Nine Points-WE COLLECTTED 7pts dropping 2. chelsea,wigan,birm city(out os ,.. 3rd in the league,4pts ahead of chelsea,....THE TEAM AS A WHOLE NEEDS TO TAKE THIS OPPURTUNITY ,Wenger was right,it is ours to loose.

Against Man city,it will be a diff ball game,we have to be ruthless but intelligently.we have to FACE OUR FEARS ,EMBRACE IT ,Nothing is impossible,it depends on the individuals ,on how they want to achieve sucess as a TEAM.

LONG LIVE THE ARSE.
Mr. Renoog analysis though good is too simplistic as others Tee Song and Passenal have pointed out. I particularly do not like the singling out of individuals as it opens them to scapegoating as Passenal has rightly pointed out. Tee Song also shows how it is important to take in other issues particularly midfield cover etc. The Blackburn Goal at home, for example, if others were aware especially midfield, even TV5 (I do not know how he is exonerated for this one; he ran into no man's land letting Diouf pass easily), should not have been a goal. Too often with Arsenal, once one person makes a small mistake e.g. gets skinned, it almost invariably ends in a goal because of lack of awareness and hence cover by others. Unfortunately, blame is apportioned to the first guy and not the others. Watch other clubs CB pairings organisation (they watch each other's back) and MF defense awareness, how they help each other. Arsenal did very well on this count on Monday Dec. 27, 2010 against Chelsea and against the Brum. This is what I see has been the case in most goals that Koscielny has been blamed for. There is no question, however, that he needs to bulk up and improve Mr. Renoog and others have pointed out that correctly.

IMHO, we cannot blame Koscielny and Squillaci too much given where they have put Arsenal so far this season. They have played the most games, and being new to the EPL plus Koscielny's inexperience, I would say they have performed very well for rookies. Yes, there is room for improvement. I have to say that having watched TV5 last season, I am not sure he is heads and shoulders above Koscielny or Djourou at present and improving. He, TV5, tends to commit too many positional errors for me, and forages forward a little too much. If you have been watchful, this season Arsenal are conceding fewer goals on the counter attack than last season. It could be just tactical adjustments by Arsene, but its also clear that the current CB are more defense minded than last season pairing.
Also, blaming CB for Arsenal conceding too many goals through set pieces is not quite fair. Why Arsenal concede so many set piece goals is not just the CB's problem! I think its a team problem, there are too many short people who cannot compete for high balls. So its Wenger's problem, he should try and balance the team with more height, plus of course know how on defending set pieces even if you are shorter than the opposition. Arsenal could also try not to concede too many free kicks in dangerous areas.
Very interesting attempt at analyzing a very complex issue. I appreciate the effort but can't say it's comprehensive or conclusive. Some of the points you've made make a lot of sense but as others have mentioned in some cases it's a bit simplistic.

I've always believed, and even Arsene acknowledged, that most goals come from 4 or 5 mistakes. At times it's the players, at times the ref, and on other occasions a lot of quality from the opposition. We also see a combination of these factors.

For instance, in the first goal against Liverpool I'd say Vermaelen is really at fault. Having seen that goal a number of times at reduced speed and various angles, it seems to me that TV5 just didn't read the danger and allowed the striker to gain a couple of yards on him. If Vermaelen had sensed the danger earlier he would have moved across and not straight in front thereby getting tighter on N'gog. I agree the midfield was at fault for that goal and even the Keeper can be questioned. So a number of errors leading to that goal.

Similarly, if we look at the blackburn goal. Samba brings the ball out without any real pressure. RvP just jogs around him without closing him down. Song gets sucked in but doesn't get into any kind of a meaningful position. Sagna is just hanging around without closing the passing channel or sensing the danger in the space behind him. Vermaelen doesn't cover for the mistake by Koscielny well enough. Clichy lets his man go past him. I agree with you observation that Koscielny was out-muscled but I can't see how a player losing a challenge on the touchline should ever lead to such a soft goal.

We can look at all the goals and there will be a number of culprits. I'm sure you know this but this article, especially the video, seems to put a lot of blame on Koscielny (perhaps you didn't mean it that way) and it's coming across as extremely biased.

I do agree with your general analysis of the strengths and weaknesses of the defenders although Vermaelen seems to be a special case. He only played in the first three games so his numbers are rather flattering - unsettled Liverpool down to 10 men for one half, blackpool out of their depth at the emirates, Blackburn was a strong performance. I don't think we can make any definitive conclusions based on these numbers for Vermaelen.

For the other defenders some of the numbers are skewed by one or two poor team performances and it would be difficult to think that other defenders (vermaelen and djourou) won't end up with similar numbers if they played for a long time when the team was not completely confident and tactics weren't always correct.

While the minutes seem like a lot, on the whole the sample size isn't big enough. Games like the Tottenham loss (Central defenders hardly at fault for any of the goals - whether they should have gone up or not and who marks whom on set-pieces is decided beforehand by the team management?) tend to make Koscielny and Squillaci look a lot worse when compared to Djourou and Squillaci who got some breathing space against City when they went down to 10 men so early on.

I don't think the solution will come by changing personnel or by giving Koscielny a break. A lot of work has to go towards perfecting the details while defending set-pieces. We also need better organization and defensive contribution from the midfield and wide players, something that was seen in the Chelsea game. Everyone at the club has to understand that defending is an 11 man operation and each player has to pull his weight.

Also, are these stats from some source or have you worked on them yourself like you do for the assists and key passes?
SharkeySure's avatar

SharkeySure · 742 weeks ago

Excellent work putting the video together and writing upi your analysis. Commendable.

However, I cannot look beyond the fact that analysis is primarily about goals conceded and not all round play, Ie Koscielny outmusckled by Diouf, but Clichy gets caught ball-watching very badly, and would ordinarily be expected t prevent BMDiouf from scoring.

No amount of stats will flag that sortif situation for me.

Lastly I watch Kozzer play and am very very surprised to read that you doubt hius aerial ability. I think he's very good in the air, the fact that he has the odd slip shouldnl;t detracft fronmhis all round good play, aerially or otherwise.

Still, its a very good article. One of the best I've read in a very long time !! Well played
Carlito11's avatar

Carlito11 · 737 weeks ago

Just be directed here via a blog on our defence. Outstanding analysis really sheds factual light on the debate we've been having. Thanks for your work on this- I have now bookmarked your site!
However, I cannot look beyond the fact that analysis is primarily about goals conceded and not all round play, Ie Koscielny outmusckled by Diouf, but Clichy gets caught ball-watching very badly, and would ordinarily be expected t prevent BMDiouf from scoring.

Post a new comment

Comments by