Sell De Gea, Hull On Earth, Small Margins And Other Premier League Stat Stories
And so the Premier League meanders dreamily off into the off season with a series of entertaining friendly matches, fond and not-so-fond farewells and recollections of thrilling climaxes sadly not equaled this time round. Gerrard's stuttering and misfiring long farewell felt like an apt reflection of a season that whilst ever intriguing, never quite hit the heights of the goal laden craziness of last year. Too early for farewells? For everyone except Hull, Newcastle and Sunderland, no. The inquests are firmly under way and in many cases the variation of a couple of wins either way has been enough to define a successful or unsuccessful campaign. As ever the divide between success and failure is relatively thin, little comfort to the disappointed underachievers (perhaps Liverpool, Man City, Tottenham, Everton, the doomed and... er... Newcastle?) and dream land for those who came out on the right side (Southampton, Swansea, Chelsea plus if you listen to Sam Allardyce, West Ham and whoever survives). These margins can be deceptively small, see here a case study of two Premier League teams this year, one deemed to have been comparatively successful and one with a more mixed record (data prior to this weekend):
Have a guess, see if you can identify them.
Okay, maybe that's a tricky task if you're a well rounded individual with just a casual interest in football statistics, but from the chart and the data we can see two very similar teams in all but one measure, save percentage, in which team A suffers in comparison to team B by a margin of seven percentage points. In this case study and over the course of a season, that's about ten goals on the against side. With save percentage being predominantly determined by luck rather than skill, we can see that amongst the strongly repeatable key shooting metrics laid out here we have broadly two teams of similar underlying abilities.
Team A is Leicester and team B is Swansea.
Of course, Swansea and Leicester approach the game with very a different ethos, Swansea eternally a passing side and Leicester very much more direct, but the two outlined styles are leading to similar outputs.
There is another level of analysis available here, but one which is blighted by the small sample created by the limited experience both Nigel Pearson and Garry Monk have had in top flight football. It is conceivable that either of these managers has incorporated into their methods tactical nuances that are enabling them to exceed predictable measures. Tony Pulis has regularly overseen notably poor shooting rates yet has found a way of generating points regardless. His super combative direct style with a heavy reliance on set pieces has repeatedly overshot expectations and we have also seen aspects of Chelsea's play this year, particularly in the second half of the season, that have emphasised control over dominance and led to a positive outcome. As ever, a 38 game sample, in the case of Pearson and around 50 in Monk's case give us a limited window into how much their influence is felt.
For now each team has reason to feel pleased with their efforts, they have exceeded expectations, but moving forward, it would seem likely that Swansea will regress next year, much as Newcastle did after their 5th place in 2011-12 or West Brom did after their 9th in 2012-13. Only very occasionally does a team with numbers akin to Swansea's or Leicester's manage to reach the top half, and I would suggest that Swansea are beneficiaries of a positive skew here and to repeat such a placing will require tangible improvement next season. Leicester found a way to do just enough and will need to look to replicate their figures next time, but there is no huge hole there, both teams project below average but without terminal issues. Trying to promote a wider narrative that these two teams are equivalent is not easy!
Hull on Earth
This week Michael Caley looked at how Leicester have improved as the season progressed in this article in which he discussed how the phantom "sacking" of Nigel Pearson had the appearance of a turning point and seemed to impact positively on the club's performance, when in fact, it was quite conceivable that Leicester were experiencing a degree of regression. And that got me thinking about Hull. Again.
I suggested last week that there was potential to deem Hull's plight as unfortunate and having dug a little deeper, I am ever more convinced by this idea. In an attempt to draw further insight, I split the numbers I've got into the season's two halves. In the first 19 games, Hull were terrible. Nearly all aspects of their numbers were bad: they averaged under a goal a game, weren't breaking ten shots or three shots on target per game and had 40 % shot ratios. Backing up this ineptitude were generally par conversion, shooting and save rates. All the indicators suggested that Hull were just bad, and their solace could only be found in the five or six other bad teams around them. These levels of performance would have preempted a sacking in many clubs and Steve Bruce would have had little to complain about if his record of three wins in 19 games had cost him his job. But Hull persisted and most of the numbers improved, as we can see here:
Steve Bruce can rightly be considered unfortunate if he fails to secure Hull's top flight status. Presumably responding to board concerns, he has effectively improved the performance levels in his team as the season has progressed but has been killed by hard to control factors. That Sunderland certainly and Newcastle arguably (but definitely since Christmas) have been worse will be of little solace if they fail to beat Man Utd and are consigned to the Championship.
Games 1 to 19 versus games 20 to now
Plenty more intriguing nuggets of information can be gained from splitting the league this way. Most notably, the team that has reduced it's shooting levels by the highest margin between the two halves is Chelsea (-3.3 shots per game, -1.1 Shots on target) and it's not just the fancy numbers, their goal decline ranks 19th (-0.5 per game), again good old Newcastle score worse here. Their total shot ratio decline ranks 19th and Shot on target ratio decline ranks 20th. All ample evidence that they coasted in. It also underlines the fact that Chelsea's first 19 games showed them to be overachieving at an unsustainable rate. They did not need to continue to perform at that early level to win the title, but nor was it likely, even disregarding any tiredness concerns, that they were going to be able to maintain that pace. The numbers predicted a significant drop off, and so it came to pass.
A similar tale can be told regarding Southampton though their defense has remained remarkably consistent throughout. Early on they were hitting solid top four pace and weathered a fixture storm reasonably well throughout late 2014 to maintain touching distance between themselves and other contenders. That their rates declined was widely predicted (not by me, I gave them a good chance after 12 games) and it was nearly all shaved off the front end. Their attack, never carefree outside of hammering Sunderland and Aston Villa, has consistently struggled to generate sufficient goals to continue to contend and a drop off of 0.5 goals per game in the second half is shown to be a function of five percentage points disappearing from both their shot ratios. Regardless, there is plenty to build on, providing their playing corps is not feasted upon in the same voracious manner as last summer.
Lastly for now, a few words on Man Utd and David De Gea. With the rumour mill pitched into reality by Louis Van Gaal's post match comments after the Arsenal game, it seems ever more likely that De Gea may be returning to Spain. Widely praised during the first half of the season and credited with a huge influence on Man Utd's mid-season successes, a combination of his contract situation and his high current reputation suggest this may be a good time to move on. "But, no!" cry the Utd fans, so enamoured are they with their talismanic keeper. "But, yes!" say I comparing an unsustainable first 19 game save rate of 76% with a below average rate of 67% charged to the second half. He is a good keeper that has experienced a season that contained a run of strong form. There are plenty of other comparably talented keepers available in the market too. Whether the realms of error and misjudgement are lurking in the near future can never be predicted and Man Utd should extract as high a fee as they can, and pronto.
Now how much is Petr Cech?
Thanks for reading!
Follow me on Twitter here: @jair1970