And so we reach the end of another season of “The Premier League”, the wildly popular and occasionally thrilling sporting serial transmitted to viewers around the world. The scriptwriters had a high benchmark to match after the goal-laden thrills of season 2013-14 and in truth they came up a little short. Familiar villain “JR” Mourinho was the key player this year, most ably supported by the hugely entertaining new character Mr Van Gaal. Old faces like Tony Pulis brought a familiar brand of stoic solidity and the swashbuckling entrance of last season’s comic relief, “Tactics” Tim Sherwood surprised everyone with an effective cameo. Also riding high were Alan “PDO” Pardew, with his successful return to London and the magic man Dick Advocaat. His miracle rescue of Sunderland was a familiar and reassuring recurring plot.
Less contentment was seen in the North as a brooding Brendan Rodgers “spent £100m and did a Spurs”, a storyline that split viewers almost equally; many had hoped for a more kindly outcome, others watched on in slightly amused awe. Lastly, elite coach John Carver gave some much needed levity as he promised the world but provided nothing and there will be mixed feelings regarding the killing off of show stalwarts Sam Allardyce and Steve Bruce.
Roll on August.
Where have the shots gone?
Earlier in the year, I noticed that there had been a decline in the number of shots in the Premier League. There were less last year than in the four years previous and fewer still in 2014-15. As we can see this trend continued to the end:
With the Premier League widely marketed as the most thrilling in the world, and having just signed a huge new TV deal, it seemed a mite unfortunate that this could coincide with a decline in shooting levels. And these reduced shooting levels have translated into fewer goals:
Whilst goal and shot levels have reduced during the Enlightened Era (2009-15), it is reassuring to note that the levels of conversion, both to all shots and per shot on target, have remained generally consistent throughout. This lightly implies that the most significant driver towards the total of goals is shot volume alone; to investigate a lack of goals, we would need to think about why there has been a lack of shots.
Amongst the big seven clubs (traditional big 5 plus City and Chelsea), the decline in shooting numbers tracks similarly (minus ~3 shots per game between 2009-10 and 2014-15) but it’s all on the front end. Broadly these clubs are creating fewer shooting opportunities but conceding them at the same rate. Particular differences can be noted in the cases of Chelsea, Man Utd and Tottenham, managed respectively by Ancelotti, Ferguson and Redknapp at the start of this period and now with Mourinho, Van Gaal and Pochettino; a significant change from attack focused football to a more pragmatic and prescriptive tactical methodology. That the new Dutch coaches, Van Gaal and Koeman have presided over two teams with extremely low combined shot rates is also noted.
It would be remiss not to mention the loss of Suarez to the league. A one man attacking phenomenon, his shot volume outstripped the rest of the league by some margin in his final two seasons at Liverpool and he was also strongly creative. Liverpool’s goal drop off (-1.3 per game) was higher than the actual goal rate of 13 of the teams 20 clubs this year.
It seems likely that these changes in goal and shot volume are an evolution in tactical thinking and strategy and possibly a general shift in popularity away from two starting strikers to often one. With ever more money on the line, it is logical that caution may be a stronger influence but we should also note the long term trends are more varied. Goal rates between 2009-10 and 2013-14 were the five highest recorded since the 38 game Premier League began and the average between 1995-96 and today is 2.64 per game, far closer to the 2.57 per game we have witnessed this year.
As a counterpoint to the usual goal chart, here we have the top 15 goalscorers ordered by Non-Penalty Goal rate, y’know, the proper stuff. I’ve added in some conversion rates to round out the story a little. Some research i’ve done suggests that raw conversion of 14-18% are likely rates for higher level forwards over longer samples, although a true poacher like Mario Gomez or Miroslav Klose can get over 20%, as seemingly can forwards playing for the Spanish giants. Diego Costa is interesting here, as his long term rate is 18%- towards the high end, and he again posted strong numbers as the focal point of Chelsea’s attack; sure he overachieved a little but nothing suggests he is anything but top class, a charge that can firmly be laid at Aguero’s door too.
Strangely overlooked during player award season, which was presumably a reflection of the prevailing negative view of Man City at the time, he has top scored in the league, ranks 3rd for NPG/90, has significantly the highest shot rate and has converted at 14%, very similar to his long term rate of 15%. There is simply nothing in his numbers that enable you to detract from his evident quality. All this implies a player at the peak of his game, which you would expect at 26, with as ever the kicker being his need to stay injury free. He should certainly be cherished.
Most extreme on the longer list is Swansea’s Ki. His 8 goals scored at a 29% conversion rate from midfield is extremely unlikely to repeat, so any potential suitor alerted by his goalscoring prowess may do well to do further analysis. Papiss Cisse is a fascinating case given that he appears to have two modes: off and on. There is simply no way of predicting what kind of season he will produce; this was a good one, apart from the disgracing himself part. It will also be of interest where Charlie Austin ends up. Whilst it is easy to presume a new player in the league on a hot scoring run is benefiting from an element of surprise, there is nothing, statistically at least, that suggests he could not repeat his efforts for another team.
More thoughts on the second half of the season
As has been widely noted, at least amongst stat types, West Ham have been in utterly miserable form since the autumn. Over the course of the last 19 games, they have posted a 42% shot ratio and 43% shot on target ratio whilst converting shots at a rate of ~6% (league average ~10%). Their only saving grace has been an above par save percentage of 74% (league average ~70%) but without their early season run in which they vastly exceeded their underlying numbers, they would have been in serious trouble. Following a 2013-14 season in which they projected extremely badly, it comes as a wonder how Allardyce retained employment throughout. The comfortable hum of league safety must have resonated up to the boardroom early on and it will be interesting how a new coach approaches the team. Historically, teams have struggled in the aftermath of Allardyce- he is known to a degree to be statistically minded and focused towards the measurement of individual tasks. Whilst such drop-offs are maybe circumstantial- he enjoyed success at Bolton then was ousted by new owners at Blackburn and Newcastle, it should be remembered that whilst posting very bad numbers often, his teams’ points totals have usually exceeded a flat projection. A new coach will need to be mindful that improvement may well need to be wholesale to match Allardyce’s final league positions. Otherwise, trouble may brew. We shall see.
The Pulis effect has been more pronounced than ever before. All of West Brom’s shot numbers now look dreadful. In the second half of the season they have taken fewer shots than any other team and posted the second worst shot ratios yet have scored points at a rate of 1.42 per game a vast improvement on the 0.89 per game prior. As ever, Pulis’ super defensive, high set piece reliant methods have defied conventional shooting analysis and his position as a semi-ironic poster boy for stats types remains solid. Going forward, what will intrigue here is how much trust he is given with West Brom’s transfers, especially given the large amount of money he spent on unwanted Tottenham players whilst at Stoke, with mixed success. I suspect Daniel Levy has made a phone call this morning.
There are strong similarities between Chelsea’s first half of the season and the second half that Arsenal have just posted. Both teams have looked extremely strong during these periods but have also benefited from favourable aspects in the numbers. Each team converted shots at a significantly higher rate than the opposition (~+6 percentage points), each found their shot on target ratio far in excess of their total shot ratio, something that usually tracks closer together and is prone to regression. This has given the impression of superiority in excess of reality. During the autumn, some commentators viewed Chelsea as a team for the ages, there was talk of “Invincibles” and suchlike, and a combination of tiredness, a lack of necessity and simple regression of hot-running metrics has cooled such talk. Chelsea were very good for large parts of the year but never as good as was being mooted earlier on. Arsenal’s run has not garnished the same amount of coverage and retained squad fitness has helped a great deal, but there is little doubt that whilst performing well, they have also benefited from the rub, and long term, it is unlikely to continue at that level. An 80% save percentage just the most obvious outlier here.
Obligatory Tottenham finale
A huge dose of recency bias has led Tottenham fans to sail off into the summer with a spring in their step and cast Liverpool fans into despair. Neither team has enjoyed great prosperity this season, and given the hope that contending for top four places is a positive expectation in any given year, they have both failed to enjoy a beneficial skew. Tottenham’s 5th place certainly flatters as amongst the mini-leaguers of 5th to 7th they easily have projected worst. I talked earlier around their issues in creating and preventing chances in the box and there is plenty of squad work to be done if Pochettino’s system is to be realised successfully, especially in defense and central midfield. It is a young squad and the plan is long term, so for now i’ll retain some hope, but there will be little wriggle room for the Head Coach if he doesn’t get a fast start next season.
Thanks for reading!
Follow me on twitter here: @jair1970
Summer content will follow!
*Quick thanks to everyone who has read and enjoyed the column this season. It’s been a pleasure to write and despite occasional criticism, it seems to have been generally well received. I have a simple intent: to widen to the interest in stats work and analytics. Also thanks to Ted Knutson for the offer to bring it to Statsbomb and particularly the loyal band of people who have regularly promoted it via social media. It is very much appreciated. See you for 2015-16!