The stats behind Arteta’s influence
Everton collected fewer points without Arteta in the team
The closing of the transfer window tends to throw up a collection of statistics aiming to quantify the influence of certain players on the teams they have left.
Here’s a few of them: Samir Nasri, Cesc Fabregas, Denilson and Phil Jones. Opta have also tweeted similar statistics on non-movers Gareth Bale, Antonio Valencia, Nemanja Vidic and Frank Lampard.
These stats are often misleading. Take Phil Jones:
34% – Blackburn’s PL win percentage with Phil Jones playing since he made his PL debut, compared to just 25% without him in the team. Hole.
In absolute terms, Blackburn won 12 of the 35 matches Jones played for the club, and won 3 of the 12 in his absence since debut. Statistically speaking, however, this difference is insignificant because the win frequency and sample size are both small; there’s nothing to suggest Phil Jones was the reason for the fall in win percentage.
In fact, the fall is largely attributable to Jones missing harder matches. Whilst injured during the winter last season, his team played a number of solid mid-table and top half teams, matches they would normally expect to struggle in anyway.
So naturally, when Opta tweeted:
10 – Since January 2005, Everton have averaged 61 points per season with Arteta playing, compared to 51 points without him. Lynchpin.
I was sceptical as to how significant this difference was. Can one man really have that big an effect on the team? The stats will never give the whole answer, but they can give a few pointers.
At first glance, it seems as though Arteta’s presence is all-important. As alluded to the in Opta statistic, a win percentage of 43 equates to about 16 wins in a season, level with fifth-placed Spurs in 2010/11. The Arteta-less win percentage equates more to a mid-table finish with about 13 victories.
But like Phil Jones above, it’s important to know which games Arteta missed:
Everton not only tended to play the majority of their games without Arteta away from home, but they also tended to play marginally more difficult opponents. As such, games with Arteta were largely games Everton might’ve expected to win, and games without him were matches they might’ve expected to be tougher.
In fact, taking into account the location of matches and the strength of the opposition (measured by their season points total), Arteta’s playing presence had statistically no significant impact on Everton’s win percentage.
As always though, statistics don’t provide the full picture. It’s obvious Arteta was a key player for Everton, if perhaps less so since his cruciate knee injury in February 2009. But what the stats do suggest is that Everton perhaps shouldn’t get too carried away with his absence.
More likely, defeats and draws can be attributed to a series of injuries, suspensions and sales, as opposed to the loss of a single individual.
In the end though, this is as much an exercise in understanding the deceptive power of statistics as analysing the influence of Mikel Arteta on Everton’s results. Time will very much tell us the true effect of Arteta’s departure.
5 Added Minutes 
Interesting read. Ultimately pre 2009 Everton’s strategy was based on Arteta – post 2009 and his cruciate it wasn’t, Baines became the main man and Arteta, whilst remaining an important player, was no longer the key man he once was as proved last season when we had a much better win% with him not in the team than when he was playing.
I’ve crunched a few numbers on that cruciate injury – before the injury & with Arteta playing (130 games) Everton’s win percentage was 43.1%, without him (23 games) 30.4%.
Upon his return in January 2010, the respective numbers are 43.1% (again!) and 38.5% (44 and 13 games).
The differences are still insignificant, probably due to the small amount of games played when broken down as such, but I think it still very much supports your point.
You have a point here, but your reasons are just as empty as these OptaJoe’s what-people-want-to-listen stats. I guess because this is a quite meaningless topic.
I’m not suggesting it’s a particularly important topic, just pointing out a few things to bear in mind with these stats.
Very good point, I think win % in isolation can be misleading for just the reason you stated – it ignores the difficulty of the opposition
Especially with low frequencies – if Blackburn get a fortunate late winner in one of the games in Jones’ absence, their win percentage rises to 33% (4 wins instead of 3 from 12), only 1% less than with Jones.
Thankfully Arteta’s played a few more games, but yes, opposition strength and perhaps more crucially home advantage must be accounted for.
I’ve heard a stat and i don’t know if it’s true: Bolton’s win percentage halves when Stuart Holden is not in the team. Do you happen to know if this is true and is it insignificant like these other differences?
Had a quick look at the numbers. From what I can see, Bolton’s win percentage with Holden was 42.9% (12/28). After his injury in March, Bolton then won 2/8 of their remaining games last season and 1/4 this year (though the losses have come against Man Utd, Man City and Liverpool), meaning their win percentage has only been 25%.
It’s a bit more time consuming to test for significance, but from the eye I’d say the difference is insignificant for 3 reasons – too few games without Holden since his debut, difficult opposition and the FA Cup ‘hangover’; 5 of their 6 losses after Holden’s injury came after the cup semi-final defeat to Stoke, where their season just fell away from them. Whether that was specifically because of Holden’s absence we’ll probably never know, but certainly for the other two reasons I’d say there’s no real conclusions to be drawn yet about Holden’s influence.
It’s probably a similar case to the conclusions I’ve drawn on Arteta!