Statistics Can Be Made to Prove Anything - Even the Truth
Football stats - they're cool. Sunderland stats - they're often crap.
An inconvenient truth perhaps, but stats are everywhere, and you can usually find one to prove any point you wish. Football statisticians are everywhere too. Clubs employ whole teams of them who pore over the minutiae of every move that every player in every team makes, to eek out that 1% which might make the difference between success and failure.
But, anyone can do football stats. Even me, with my GCSE in mathematics earned the year Sunderland last reached an FA Cup final. Allow me to illustrate, with two devastatingly simple, but telling, metrics and a couple of graphs.
If the international break afforded us a chance to take stock and reflect on a dismal opening three months of the season, it also enabled us to wallow in the glow of victory longer than is healthy, as a rare three points at Bournemouth has us filled with a little optimism. But, strap yourself back in because it's about to start again - with Hull the visitors in a proverbial six-pointer this Saturday.
David Moyes' Sunderland team has, thus far, been largely devoid of an identity. Whether it be through choice, stubbornness, pragmatism or a lack of options, Moyes has yet to find exactly what that je ne sais quoi which will define his team, actually is.
Regardless, the failings of the current side in their eleven games so far this season have been glaring, and can be summed up with 4 'Ps':
1. Possession-less, Sunderland have an average possession of 41% so far this season, but it hides a bleaker picture - see 26% possession away at Tottenham or 33% at home to Everton.
2. Position-less, players on the pitch who have appeared to not know where they are supposed to be and when. Plus, A hotch-potch of zonal and man-marking systems.
3. Pressure-less, with a midfield seemingly incapable of retaining the ball, or closing down opponents, Moyes' side has exerted far too little pressure on opponents whilst inviting a bombardment onto its shaky defence.
4. Pointless, ultimately bottom of the league with one win in eleven.
So it was with a sense of wonderment that some listened to Moyes three weeks ago, following defeat at West Ham, as he proclaimed that he could sense his Sunderland team were improving. And not just sense it - evidence it:
Our stats are getting better. To a lot it won't be noticeable but we're looking at it closely and the players are positive.
Immediately following those words, his team were knocked out of the EFL Trophy and smashed to smithereens by Arsenal at home.
But, perhaps he was right. Here's two cool metrics, often used by clever people, that add some weight to the cloying hope that the green shoots of recovery have some base and Sunderland are indeed improving.
The Pressing Problem - PPDA
Pressing is considered a fine-art in today's game of football. The high-press, the medium-block or a low-press/counter-attacking style are the three broad methods of dealing with the opposition.
In recent years, Sunderland's variety has tended to be a kamikaze creation which has swung between some trouncings and some stormingly effective play during Sam Allardyce's latter few months.
It's been a backs-to-the-wall, men-behind-the-ball, occasional counter attacking ploy at Sunderland; and it can be devastatingly effective when done right, but you still need to actually get the ball occasionally and if you get it wrong you get tonked.
So, pressure on the ball is important if you want to get a hold of it and do something with it. Statisticians use some simple measures to try and quantify this aspect of football.
One such measure is PPDA (Passes per defensive action) and we've borrowed this version from Ted Knutson, former stats-guy at Brentford and genius behind some top betting companies, and cobbled together a plausible version which works.
Simply, PPDA is a ratio of the number of opposition passes in a game against any act in which you put pressure on the ball when the other team has it - be it a tackle, a challenge, an interception or even a foul.
Here's how Sunderland have shaped up this season; remember - the lower the figure, the more effective you are at harassing your opponent:
Like all good stats, it tells you exactly what you thought you already knew. So Sunderland were appalling against Everton and Arsenal (the two spikes), but much better against West Brom and, finally, another dip at Bournemouth for that all important maiden victory.
The graph is a wibbly up-and-down thing which reflects the inconsistency with which Moyes' Sunderland side have performed. But, that big red trend line is showing in the right direction.
To put this PPDA-pressure stats stuff in some context:
- First 11 Games of Last Season, David Moyes, average 6.9
- Final 11 Games of last season, Sam Allardyce, average 5.7
It might not sound a big difference, but it is. Allardyce's run-in side were 18% better at exerting pressure on their opponent than Moyes' season-starters.
Defence & Attack - PDO
All the best stats have an acronym, and this one is no different. Stats-dudes have used PDO for years, but it doesn't stand for anything - the bloke who dreamt it up just used his username off a hockey forum.
It's a metric of a team's effectiveness at defence added to its ability in attack. They use it in Hockey, Baseball and the NFL all the time to illustrate commentaries and analysis. 'Proper' soccer analysts hate it and dismiss it all the time because it isn't 'clever' enough.
But it works. Studies have shown that PDO almost always eventually regresses to the mean - which basically indicates that it can be used to tell a tale of what went before, and to forecast the future.
So in our GCSE maths version, PDO is the sum of a team's shooting percentage and it's save percentage:
(goals scored/shots on target) + (saves/shots on target)
Here's Sunderland's graph of the season so far:
Again, the line is all over the shop. For one thing, the important thing with PDO is where its average is going, more so than the spikes and dips of individual games; and for another, Sunderland's inconsistency this season is just glaringly wobbly. But, the big red trend indicator is clearly on the up, which is a great thing.
Even more interesting, Sunderland's average PDO value is currently 97.8 and as stated above, PDO nearly always regresses to a norm, in this case 100. Which means that big red line is likely to continue rising a bit longer yet (to 100 at normal performance, or higher if we push ourselves).
Again by way of comparison, in the final eleven games of last season, Sam Allardyce's Sunderland were way over-performing with a whopping average PDO of 112.8.
And, if you still don't believe me - hark back to the highlights of last season, and consider that at Norwich, Sunderland's PDO was a huge 185, and the home victories against Chelsea and Everton saw values of 150 and 137.5 respectively. Told you it worked.
Stats - aren't they great when they tell you what you already knew? And, with a huge game on Saturday, them big red lines better keep on tracking in the right direction.