It’s PFA Team of the Year time (well, the actual day as of writing), and team of the year type awards and articles will be two a penny. There will undoubtedly be a few stats-based awards too, but most of these will probably be bad (no offence), although probably no worse than the non-stats award pieces too.
I’m aiming to put together a Team of the Year, based on stats. But it also needs to be a collection of players that will work as a team (I’m not Garth Crooksing it). This, as well as common sense, means that I don’t have to just pick players at the top of the list of ‘tackles made on a Thursday’ or ‘headed clearances in the middle third’. The stats have also got to be public stuff, that anyone can find, otherwise things turn into a cabal of evil number wizardry. Anyway, here it is.
(The entire team is at the bottom if you just want to skip to the end without reading my reasons).
Goalkeeper: Adrian (West Ham)
This was tough, both because he was up against David de Gea (I am, full disclosure, a United fan) and because it’s very difficult to get something meaningful from goalkeeper statistics. A pure percentage of the shots they save isn’t repeatable. This difficulty makes this part of the team probably the longest, so bear with me and the pace will pick up after it.
However, some people have been working on this. At the mid-way point of the season, Thom Lawrence (@deepxg on Twitter) looked at keepers, adjusting saves made for the quality of the chance. Of the players who’d played meaningful minutes, Adrian came out on top.
And this was also backed up by Paul Riley’s (@footballfactman) similar ratings, based on shots on target faced from 2010/11 to the present. Although this isn’t solely for this season, it suggests that the midseason ratings were on the right tracks and are likely to have held out.
However, Sam Jackson (@Sam_Jackson94) has looked at whether saves made are saved safely instead of parried dangerously (using data from Dan Kennett (@DanKennett), but the graph is public, so it qualifies for this). Adrian’s saves (again, not solely for this season) are a bit more dangerous than the average, but not by too much, so it isn’t enough to make me give the shirt to De Gea.
Back-line: Francis (Bournemouth), Van Dijk (Southampton), Alderweireld (Tottenham), Fuchs (Leicester)
I basically gave these to you as a back 4 to help you get past Simon Francis at right-back. Of right-backs who have made more than 10 appearances*, as per WhoScored, he has the most Key Passes per 90 minutes at 1.7. [*this ‘more than 10 appearances, to cut out players who’ve just had one good game, will be a given from hereon in]
This brings me nicely to Fuchs. I figured that a Team of the Year is inherently going to have an attacking slant, and that Fuchs had played well for Leicester this season, and that he offers something different to the typical bombing wing-back that we have nowadays. He (along with Nathan Aké) have the highest amount of blocked crosses per 90 minutes out of left-backs. He (along with Aké again) has also made a lot of tackles and interceptions, although these are measures that should be taken with a pinch of salt in terms of quality.
While both are fairly defensive, Francis seems to go forward more, with more dribbling and more key passes than Fuchs. Francis is also versatile, having played competently at CB for about half of the season, which is a useful quality considering both of my centre-backs are also capable on the ball and able to step up.
I wanted one centre-back to be safe, a safety net. Toby Alderweireld seems to be this, by certain stats. Though part of this will be down to the system he is in, he’s one of the centre-backs who loses the ball the least while on it (ie not through bad passes, though his pass percentage is quite good too), only 0.3 times per 90 minutes, and also commits the second fewest fouls per 90 of Premier League CBs at, again, 0.3.
At his side, I wanted someone who was good in the air, and who could also pass, and Van Dijk fits the bill. He goes for a total of 6.6 aerial duels per 90 and wins a pretty damn good 74% of them (4.9 per 90). For a comparison, Sakho is 3rd on the list of aerial duels won at 3.7, but this is only 64% of his total 5.8 per 90 attempted.
Central midfield trio: Gueye (Aston Villa), Alli (Tottenham), Eriksen (Tottenham)
Right, I know. Fight me. Gueye has basically the same stats as Kanté though which might be relevant to a primarily defensive-minded central midfielder. The teams have had similar amounts of possession (45% for Leicester, 47% for Villa), although differences in how much pressure the teams are under may have had slight effects. Gueye loses the ball and fouls a bit more, but I figure that’s because he’s coming under more pressure in a bad Villa team, and we should remember that Leicester did reportedly want Gueye in the summer.
Also, I’m playing the Team of the Year system – I figure you have a couple of shots to make unusual calls, I used one with Francis and I’m using my second now.
Alli and Eriksen are more conventional. I’m a big fan of the kind of rotating central midfield trio system, probably because I started paying slightly closer attention to football matches around the time Manchester City had their title-winning De Jong-Barry-Yaya Touré triangle.
When in attacking (central) midfield or central (conventional) midfield, Eriksen has 3.5 key passes per 90 minutes, 2nd only to a man who will be mentioned later. Alli for his part is 5th in this category on 2.2 per 90, while also having posted some nice defensive numbers (which, again, shouldn’t be treated as a be-all and end-all). The offensive work of Alli can make up for my full-backs if they’re not wanting to go forwards. I also figure the two are willing to work defensively as they have done for Spurs this season, which will be important if my front three are feeling a bit energy-conservingly self-indulgent.
Wide attacking midfielders: Mahrez (Leicester), Özil (Arsenal)
Özil is the man previously alluded to, blowing apart the competition on key passes of people in every position, on 4.4 per 90. This string of tweets from Joel (@MessiSeconds), starting with the one embedded below and linked here, nicely sums up how Özil would be getting even more assists than he is now if the guys around him were finishing at normal conversion levels. Oh, and he also tops the assist list for this season already. By 6. (His 18 being trailed by, oh, Eriksen on 12).
This stat that’s going round is drastically misunderstood in like every possible way. https://t.co/GzPmZttMPF
— Joel (@MessiSeconds) April 21, 2016
Mahrez was up against Payet and David Silva for his spot. Silva only had a half-chance, not really having had a *shining* season, and I’d also be worried about him defensively, as I’ve noticed he basically does none of this for City.
Payet has more key passes per 90 than Mahrez, but we already have the top 2 assist makers this season already, and I think the balance of my side would prefer the player who is more likely to act as a second striker, which is Mahrez. Mahrez has had more shots than Payet per 90 (2.7 vs 2.2) but, crucially for my preferred possible second striker role, Mahrez has far more in the box – around 1.5 vs 0.8 per 90.
So, Özil on the left, where Fuchs can carry him defensively, and Mahrez (who from watching a bit of both of them seems more likely to do running defensively) gets the right with the slightly more attacking Francis.
Striker: Kane (Tottenham)
This one isn’t too hard to dispute.
He ranks top of Paul Riley’s public shots on target expected goals model (which you can play with here), although he does occasionally take shots from angles and positions you probably wouldn’t recommend, as you can see. But it works.