Studying the Sackings: Statistics for Managers and Tactics

When Martin Jol was removed from his position as Fulham's manager and replacement by René Meulensteen, the Cottagers seemed to respond immediately. They were unlucky to lose home to Tottenham Hotspur, then solidly beat Aston Villa four days later. Since those matches, however, Fulham have looked a lot like Jol's Fulham—pretty passing up front giving no support to an aging swiss cheese defense. By contrast, when Sunderland fired insane fascist Paolo di Canio, his replacement Guy Poyet promptly lost 4-0 at Swansea City. The club has been building from there, picking up results home to Manchester City and recently away to Everton, and my projections now show Sunderland climbing out of the drop zone. And of course the improvements at Crystal Palace under Tony Pulis have been recently documented.

What I want to do here is look at what the managers have done, tactically. Have Sunderland, Crystal Palace, or Fulham started playing football differently under the new bosses, and do these tactical changes show up in the stats? This is a question that I will return to later in looking at Andre Villas-Boas and Tim Sherwood once the sample for Sherwood's Spurs gets a bit larger.

Tactical Effects of Managerial Changes

I'm going to look at three different statistical markers of tactics. These are surely not the only statistical markers of tactics. I like to look at statistics directly related to the kinds of shots a club takes and concedes, since the general goal of tactics is to create good chances while preventing the same. The sorts of chances, good or bad, that a club creates should be a good marker of their tactics.

I'm going to be looking at three basic stats. First is the shots and shots on target taken and allowed from different areas of the pitch. This gives a general sense of shot distribution. I'm most interested in the percentage of shots taken from the danger zone (the close/central area of the box, zones 1-3 in the map to the right).

I also will look at how these shots are created, most importantly whether they are created by crosses or through-balls. Shots from crosses have significantly lower expected goal value than shots from other passes, while shots from through-balls have much greater average value.

Almost all shots from crosses are taken from the danger zone, so I will be looking at the percentage of shots from the danger zone which are assisted by crosses. Likewise, almost all shots off through-balls are taken from either the 18-yard-box central or the wide areas of the box (zones 3-5), and so those will be my denominator for through-balls.

Sunderland: Taking Better Shots

It is striking just how differently Pulis and Poyet have their sides playing based on the numbers. And more than that, they seem to have improved their clubs' play by instituting quite different changes which nonetheless have both led to improved shot quality. Meulensteen, by contrast, seems to have changed little at Fulham and to little overall effect.

First, here's Sunderland, with stats pro-rated on a per-match basis. That means "DZ SiB" is the number of shots from the danger zone taken per match. Di Canio managed Sunderland for five matches, in which time they attempted 17 shots from the danger zone, so 17/5 = 3.4. Likewise with the other numbers, "Wide SiB" are shots from zones 4 and 5, SoB are shots from outside the box.

The second set of numbers are the tactical percentages. First the percentage of danger zone shots assisted by crosses, then the percentage of wide and central shots in the box assisted by through-balls. Next is just the percentage of shots taken from the danger zone, and finally an overall rating of average shot quality. For each manager, I'm listing the numbers on attack and defense.

Sunderland	DZ SiB	DZ SiBoT	W SiB	W SiBoT	SoB	SoBoT		%Cross	%TB	%DZ	Shot Qual
Di Canio (attack)	3.4	1.0	2.2	0.4	7.0	0.8		70.6%	15.4%	27.0%	0.067
Di Canio (defense)	5.4	2.8	2.2	0.8	5.6	1.0		51.9%	17.6%	40.9%	0.090
Poyet (attack)	3.9	1.5	1.3	0.8	6.3	1.5		40.4%	18.9%	33.8%	0.089
Poyet (defense)	6.0	1.5	1.7	0.5	10.3	3.0		61.1%	6.7%	33.3%	0.072

Sunderland's overall shot ratio under Poyet has actually gotten slightly worse. Under Di Canio they averaged 12.6 shots per match against 13.2 conceded, while under Poyet they're averaging 11.5 shots attempted and 18 shots conceded. However, the quality of these shots has changed massively on both ends.

First, as you can see, Di Canio's attack was built heavily on low-quality attempts. His Sunderland took barely a quarter of their shots from the danger zone, and 70% of those shots were created by crosses. In five matches, Sunderland attempted a total of five shots from the danger zone which weren't created by crosses. He did have his club attempting a good number of through-balls.

Poyet has changed just about everything about Sunderland's shot selection and creation, with the exception of the through-balls. Sunderland have retained Arsenal-like rates of shots created by through-balls. But they have increased the number of shots taken from the danger zone to about 1-in-3. And these are on average better shots, assisted by crosses only 40% of the time.

There also appears to be a real defensive effect, as Poyet seems to have helped build a better "shell" for the defense. Di Canio's back line was being threaded by through-balls regularly, and that trend has been stopped. At the same time, Poyet seems happy to allow opposition clubs to shoot from range, as long shots conceded have increased. That looseness outside the box seems to be a function of maintaining tighter marking inside the box, as opponents have had trouble getting off good shots from the danger zone. You can see the lower quality of danger zone shots by two markers, first the increase in the percentage of shots conceded assisted by crosses (up from 50% to 60%) and second the decrease in shots on target allowed from the danger zone.

With samples of only several games, it is difficult to draw strong conclusions just from the numbers. My observation of Poyet's Sunderland has been, at least on the defensive side, of a well-organized unit that closes down good opportunities quickly. I was somewhat surprised by the clear changes in the attack, as I hadn't noticed anything too interesting about the club's attack in the matches I've watched, but perhaps the lesson here is less that Poyet is doing something cool than that Paolo Di Canio had his club playing just terrible attacking football.

For reference, these are league average per match numbers for the stats above:

League Avg	DZ SiB	DZ SiBoT	W SiB	W SiBoT	SoB	SoBoT		%Cross	%TB	%DZ	Shot Qual
Average	4.9	1.9	2.1	0.8	6.2	1.7		47.9%	8.8%	37.4%	0.085

Sunderland aren't good, to be clear. They're still taking fewer danger zone shots than league average and allowing more. But by improving the quality of the shots they're taking while forcing their opponents into lower-expectation chances, Sunderland have improved to a level where they appear capable of escaping relegation once again.

Crystal Palace: Tony Pulis Doing Tony Pulis Things

Everyone knows how Tony Pulis teams play football. It's the "traditional English game" with an attack based on long balls and crosses into a big man in the box and a solid defense in banks of four. He very quickly has gotten Crystal Palace to radically change their attacking strategies, though he does not as yet seem to have affected the club's defensive numbers.

The difference here is particularly notable because previous manager Ian Holloway likewise has a preferred style of play. He did not attempt to play quite the expansive attacking game that he attempted with Blackpool in their one-and-done season in the Premier League, but they weren't playing Pulis-ball either.

Crystal Palace	DZ SiB	DZ SiBoT	W SiB	W SiBoT	SoB	SoBoT		%Cross	%TB	%DZ	Shot Qual
Holloway (attack)	2.1	0.9	1.9	0.8	6.3	1.0		41.2%	0.0%	20.7%	0.067
Holloway (defense)	5.3	2.1	2.4	1.1	5.9	2.3		40.5%	3.9%	38.9%	0.087
Pulis (attack)	5.6	2.3	1.3	0.4	4.9	1.7		56.4%	0.0%	47.6%	0.089
Pulis (defense)	5.3	1.9	2.3	0.9	5.6	1.3		51.4%	4.2%	40.2%	0.080

As you'd expect, Tony Pulis has his club taking shots from high-expectation areas in the central zone of the eighteen-yard box. Crystal Palace now take fewer attempts on goal from either outside the box or the wide zones of the box, and they have more than doubled their rate of shots from the danger zone. These danger zone shots are assisted most often by crosses, so while they aren't the best of shots from that area, they are still better than the typical shot taken when Holloway was in charge.

Tony Pulis clubs never ever play through-balls.

The defensive numbers are strikingly similar. There's a small increase in the number of shots conceded which are assisted by crosses, but with almost all the other numbers staying the same, I'd guess that's just variation. Can a club escape relegation with a blah defense and a whole lot of crosses in to Cameron Jerome? Pulis looks to be testing that theory.

Fulham: When You Can't Defend, You Can't Win

Fulham started the season on a run of form that seemed to demand new kinds of hyperbole. They were being outshot by an average of 20-to-8, and even worse nearly 8-to-2 from the danger zone. A lucky opening day victory over Sunderland and a few other results against the run of play were keeping them afloat, but the football they were playing was going to lead to disaster sooner or later. Eventually after an abject performance in a three-nil thrashing at West Ham, manager Martin Jol was let go and replaced by his recently-hired assistant René Meulensteen.

I expected Meulensteen to address the basic problem at Fulham, that their defense is horrific. But he has persisted playing many of the same low-effort flair players in attack and the same 30-somethings in central midfield and defense that had failed under Jol. A few pretty good results early on buoyed Meulensteen and perhaps gave him the false impression that if he could just unlock the club's attacking potency, he could build a survivor. While Fulham have improved in the attack, they have continued their barely-there style of defense, as exemplified in the six-nil stomping given by Hull City. Hull had been averaging about 1.1 shots on target per match from the danger zone before meeting Fulham. They put seven on target in one match, which now makes about 25% of the club's total.

Fulham, I think, are terrible.

Fulham	DZ SiB	DZ SiBoT	W SiB	W SiBoT	SoB	SoBoT		%Cross	%TB	%DZ	Shot Qual
Jol (attack)	2.4	1.3	1.1	0.7	4.7	1.0		38.7%	14.6%	29.2%	0.069
Jol (defense)	8.2	3.3	2.4	1.1	9.2	1.8		52.3%	5.2%	41.5%	0.087
M'steen (attack)	5.0	2.0	2.7	1.0	6.2	2.0		33.3%	0.0%	36.1%	0.084
M'steen (defense)	8.0	3.8	1.7	0.3	7.2	2.8		39.6%	0.0%	47.5%	0.103

Meulensteen definitely has his attackers creating more opportunities from all over the pitch. I'm not sure why the rate of through-balls has dropped from a well above-average 15% down to zero, but that certainly looks like a change in tactics.

The improvement in the attack, however, doesn't help much if the club is still allowing four SiBoT from the danger zone every match. Fulham have been allowing slightly fewer shots, but they've been of even higher quality. More from the danger zone, fewer from crosses, and the club's average shot quality allowed has spiked over 0.10.

Conclusions: Managers and Tactics Matter

It seems evident to me that both Gus Poyet and Tony Pulis have put their particular stamp on the clubs the took over this fall, and to positive effect. Poyet has his club attempting better shots and forcing the opposition into lower-expectation tries. Pulis has his club shooting from better areas, even if those shots are mostly created with crosses. My numbers suggest that both Palace and Sunderland have been playing much improved football in the last few weeks, and I think the components of that improvement can be seen pretty clearly here.

I'm somewhat more impressed by Sunderland's numbers. I've watched them play a bit more than I've watched Palace, and while nothing exciting is happening on Wearside, they looked like a competent Premier League side. Palace are still really shot on talent, and I'm not sure that long balls and Cameron Jerome can save them, even though the numbers have been pretty solid recently. '

Fulham look screwed. This is now two managers and 19 matches with one of the worst defenses I've ever seen, and if there's a solution to the problem it's hard to believe Meulensteen is the man to find it.

I'll be looking at doing some more managerial profiles over the next month, so if you have anyone who you think would make an interesting study, let me know in the comments. I'll try to get to as many as I can.