Six top flight managers have been relieved of their duties so far in this 2013/14 Premier League season. Some of those dismissals, at the time, seemed fairer than others. What I am going to look at in this short piece is how those six teams who replaced their managers fared before, and after, the management change.

To examine how those teams fared I am going to look at each teams Shots On Target Ratio and also their PDO. I will include each teams points per game number and their goal share%, but I want to get a feel for how these teams performed by the underlying numbers. I shall list the teams rolling numbers in each category and the new managers rolling numbers in each category for comparison.

Shots on Target Ratio will be used as a proxy for team skill and control on games. PDO will be used as a proxy for for ‘luck’ or, as is another way of putting it, by how much did each team outperform its Shots On Target Ratio.

If team X has a PDO 10 points above average but a Shots On Target Ratio ten points below average it may be fair to say that team X has outperformed its SoTR. That ‘over performance’ is, commonly, generally and perhaps not entirely fairly, referred to as ‘luck’

Issues With Method

Strength of schedule, injury, home/away splits and especially score effects can, and will, skew the numbers I am going to look at here. There is no real way of factoring all these effects into this study. The study isn’t perfect, I neither have the time or the skill to make it perfect. I have what I have and that is all.

We shall start with the first manager to be dismissed.

Paolo Di Canio

 

Mad dog Paolo di Canio was dismissed after just 5 games of the 13/14 season. An abrasive temperament, an inability to coach his team in the basics of game control, an inability to outshoot the opposition and a failure to fix the gigantic fucking hole in Sunderland’s midfield were just a few of the myriad reasons that Di Canio may have been dismissed for.

But really this was about Ellis Short fixing his mistake in appointing Di Canio in the first place. And I’m fine with that. If you have made a mistake don’t be stubborn or proud, but instead fix it. The fix here for Short was Di Canio’s rapid dismissal.

Wowy_pdc_medium

WOWY: Di Canio

With Di Canio W/O Di Canio
SoTR 31.4 41.7
PDO 81.4 100.2
PPG 0.2 1
Goal Share% 21.40% 41.80%

 

Not entirely fair to judge Di Canio on a tiny sample of just 5 games, but the information Ellis Short had after just 5 games may well have been more than enough.

Di Canio’s team had no control (SoTR), no luck (PDO) and the points and goal share were abysmal. Things have improved slightly under Poyet but 1 ppg may not be enough.

You’ll notice how Poyet’s Goal share matches his SoTR number which tells us that scoring% and save% haven’t been too cruel or kind to his team. Sunderland are exactly where they should be under Poyet and that is a 1 ppg team.

Di Canio was terrible, Poyet is a touch better but he likely doesn’t have the horses to improve this team beyond the numbers posted.

Holloway/Pulis

Insufficient samples here. Holloway lasted 8 games. Pulis has been in charge for 9 games.

Pulis_wowy_medium

So, the first cut-off line indicates Holloway’s time in charge. The second cut-off point indicates the start of Pulis’s time in charge. Pulis has posted a better PDO and a better Shots on Target Ratio.

WOWY: Holloway/Pulis

Holloway Pulis
SoTR 34.4 53.8
PDO 86.8 91.66
PPG 0.37 1.44
Goal Share% 26.09% 43.75%

 

Holloway’s numbers were pretty tragic. Goal share is lower than shots share which points to an under performance in PDO. Alas, this is the case.

Pulisball has seen Palace take a staggering ~54% of the shots on target, but register just 43% of the goal share. Again, PDO is the culprit.

No matter, Pulisball has magical powers and those powers are strong enough to record a ppg of 1.44.

Pulis has been an inspired choice to replace poor Holloway. 53.8% of the shots on target, just let that sink in.

 

Martin Jol

Fired after 13 games with what is, as far as I know, the worst SoTR on record. Not even Sunderland under O’Neill were this insipid.

Wowy_jol_medium

Since Jol’s departure Fulham have posted improved SoTR numbers (just under 50%) but PDO is a real issue.

WOWY: Jol

Jol W/O Jol
SoTR 30.3 46.7
PDO 101.6 79.89
PPG 0.77 1
Goal Share% 31.43% 31.43%

 

Jol’s numbers were a fair reflection of performance: SoTR matches Goal Share%. thus 0.77 ppg is a fair reflection of his utter ineptitude in managing this football club.

Since Jol’s departure Fulham have posted relatively strong SoTR numbers but that hasn’t led to a goal share% indicative of the strong shots performance. Why is that? PDO is damn crippling this football club.

Jol deserved to be sacked, but the decision to sack him was likely 5 or 6 games too late, which was probably caused by the belief that the PDO spike around game 8 was actually real talent instead a temporary variance. Meulensteen has Fulham posting good shots numbers but PDO is a real issue. A point per game pace is fine, and it may improve if PDO regresses.

 

Andre Villas-Boas

I really didn’t like Villas-Boas’s sacking, but something was broken at Spurs during those last days and we will likely never find out just what that something was.

Wowy_avb_medium

Villas-Boas posted strong SoTR numbers but the PDO was crippled by a low scoring% early in the season. Once the scoring%  started to improve, along came the save% regression. During the last few games of Villas-Boas’s reign Tottenham’s lowly PDO was finally destroyed but two alarming blowout losses at the hands of Man city and Liverpool.

Since Villas-Boas’s departure Tottenham have posted a PDO of, wait for it………130.03. In layman’s terms: the most ridiculous, obscenely unsustainable PDO that not only I have ever seen during a short run of games, but the most obscenely unsustainable PDO one could ever possibly imagine.

Poor Andre.

WOWY: Villas-Boas

Villas-Boas W/O Villas-Boas
SoTR 57.5 55.3
PDO 85.4 130
PPG 1.68 2.66
Goal Share% 41.67% 73.68%

 

Villas-Boas’s goal share% was lower than his SoTR due to the poor PDO. This probably had a knock-on effect to the low ppg number.

Sherwood is a genius, the right man for the job, has liberated the Tottenham players, found the right attacking balance a lucky bastard in that Tottenham have posted a 53% scoring% and a 77% save% during his short reign. Goal share% is far higher than SoTR due to that PDO, thus ppg is pretty darn high.

That PDO number won’t last, Tottenham’s ppg number won’t continue to be that high. Six games is a tiny sample, good or bad stuff can happen. In Sherwood’s case, he has either stumbled on a formula which turns Tottenham into BarcaMunich or he is riding some lucky percentages/unsustainable play.

Villas-Boas posted some good fundamental numbers numbers at Tottenham but some systemic and behind the scenes issues meant his sacking was likely a fair one. Jury is out on Sherwood.

 

Steve Clarke

 

Wowy_clarke_medium

I thought the sacking of Steve Clarke was a touch unfair on the dour Scotsman at the time. I may still feel that way for a little more time yet, but Pepe Mel may well be a significant upgrade. We need more data on Mel and for that we must wait.

Clarke was posting decent fundamentals but since his departure West Brom have improved their SoTR despite facing some pretty tough opposition.

WOWY: Clarke

Clarke W/O Clarke
SoTR 43.1 55.3
PDO 100.6 97.7
PPG 0.93 1.16
Goal Share% 43.59% 50.00%

 

Clarke’s numbers were a pretty fair indication of his ability with West Brom: Sub-par shots on target team who posted a sub-par goal share with a league average (fair) PDO. Since Clarke’s departure the SoTR number is has improved (small sample), the goal share has improved and thus points per game has also improved.

The numbers during the six games without Clarke look pretty good. The sacking seemed harsh but the early returns point to the possibility of this West Brom team having more talent that Clarke was able to coax out of it.

 

Malky MacKay

Oh dear, Malky. This isn’t good.

Wowy_mackay_medium

MacKay’s Cardiff were a poor SoTR team under his management, so poor, in fact, they came mighty close to matching the ineptitude shown by Fulham under Martin Jol. The above chart shows us that Cardiff were always a poor shots team but were merely propped up by a high PDO which regressed ever so slowly toward the mean.

WOWY: MacKay

MacKay W/O MacKay
SoTR 31.3 44.9
PDO 100.5 81.14
PPG 0.94 0.25
Goal Share% 31.71% 28.57%

 

MacKay’s numbers were a fair reflection of what he was able to get out of this team: SoTR matches the goal share, while PPG is fairly impressive due to the early season PDO spike.

We don’t really have enough information about Cardiff without MacKay but it may be fair to say that Cardiff will likely not continue to post that impressive a SoTR number, nor will they continue to suffer the cruel blows that their low PDO number currently deals them.

MacKay was likely not a good enough manager to manage in the PL, but we cannot be completely certain of this due to the talent level at his disposal. Cardiff do have some nice pieces at their club but those pieces are likely not nice enough, nor are there enough of those nice pieces to guarantee safety in their battle against relegation.

We wait to see what the super sub can coach out of this team.

 

Final Thoughts

Managers matter, but so do small samples and the variation in not only performance and luck, but in the variation of strength of schedule and injury. Personalities matter too. Just ask Daniel Levy or Vincent Tan.

Managers can be prematurely fired off the back of periods of ‘bad luck’ just as they can perhaps be fired too late due to periods of ‘good luck’. Stuff happens, life for a manager can be as unfair as it is fortuitous.

My take is thus: by all means fire a manager for posting poor numbers in terms of control of the shots count (SoTR) but be wary of dismissing a manager due to a low PDO (the coach killer) unless it is absolutely certain that the low PDO number is caused by long standing system issues.

Personally, my take away from this piece, and other private work that I have been doing, is how well Goal Share% fits with SoTR (and TSR) numbers. If those two numbers do not sit pretty close to each other then the cause is likely due to a high, or low, burst in PDO. Given enough time, shots talent will equate to goal share talent.

And we know what happens to PDOright?

  • Ben

    Hi Ben,

    Yet another wonderfully informative piece, here. I just had a quick query about PDO. I seem to remember reading (I think on a piece about Benteke) that SoT from prime/central positions were significantly more repeatable than those from wide areas and from outside the box. If this is the case, would it be feasible to split the data used in calculating PDO into two sets, prime and non-prime, in order to better evaluate the influence of luck/skill?

    Thanks,

    Ben

  • Pinkybum

    It seems like you haven’t multiplied the PDO number by 10 – is this correct?

    On another blog they put the formula for PDO as:

    PDO = 10 X (Shooting percentage + Save percentage)

    As you can see the “10” is purely arbitrary – I like yours better.

Improve Performance and Productivity in Your Club:
State-of-the-art Football Analytics