Newcastle v Sunderland-Premier League

Goddamn it. I try so hard not to write about numbers ‘n crap outside of Tuesday, and then I read some thing that makes me get out of my chair and walk around a bit. James Grayson did it today. So, sorry for those who’d rather read about whatever else it is I cover these days. Football finance? Sure.

Before I dive in, just a primer: Grayson’s approach incorporates a lot of the work done in hockey circles. Both hockey and soccer correlate well on a lot of shot/goal dynamics, seeing as they’re both “team invasion sports.” One of hockey’s long-time stats is PDO, which is a pretentious and meaningless term to describe shot percentage plus save percentage times a thousand (see? I’m not even going to write out the formula because I respect you so much).

What does PDO tell us? Well we know from regression analysis that team shot and save percentage are largely a function of random variation, what the world generally calls luck. If a team has a really low PDO or a really high PDO, chances are the numbers will adjust accordingly. Teams with mediocre underlying numbers with a high PDO are generally going to get crappier, and teams with good underlying numbers but a low PDO are generally going to get better.

Things get a bit weird on the high end of the scale, and there’s just a little bit of controversy as to why that is: a post for another day (or likely one I’ve already written).

But PDO is fun because here we have a simple, easy-to-calculate metric that can be used to make a reasonable prediction about a team’s future success. No cumbersome Manchester City data dump required! And Newcastle’s PDO has been interesting for two seasons running now. Grayson tells us why:

As director of football is Kinnear going to be directly responsible for an improvement in Newcastle’s save percentage (reanked 20th out of 20 in ’12/13) or their shooting percentage (ranked 17th out of 20 in ’12-13)? Almost certainly not. However it’s highly, highly, likely that they will regress in Newcastle’s favour, and Newcastle are almost guaranteed to get better results the moment Kinnear walks in the door. Correlation doesn’t mean causation, but Kinnear is going to get a good portion of the credit here, and were I a numbers-savvy member of the MSM who had confidence that this arrangement won’t blow up mid-season I’d be swimming against the tide and saying what a good appointment this is, then making sure to link back to that piece throughout the season.

I love the delicious cynicism here of a potential pundit singing Kinnear’s praises fully-well knowing Newcastle will likely improve as a matter of simple regression and not because Joe’s got a keen eye for great footballers that the Toon can afford. Because what matters “is being right,” not pointing to an objective set of facts that really have nothing to do with Kinnear at all. Because anyone could do that, really. Even you, the unwashed football fan who likely knows as much about this sport as anyone paid to write about it (likely, but not certainly).

Anyway, there is I think a major lesson here for all of us stuck watching this sport until kingdom come. In football, not everything good that happens to a team in football is necessarily the cause of some deliberate good decisions on the part of a single manager or football director. Sometimes it’s just a lucky streak coming to an inevitable end. The flipside of this is is that not everything bad that happens to a football club is the fault of a single manager or football director either. We have here with PDO a means to try and figure out the difference a bit.

As ever nothing is definitive. PDO is one clue among many, but not “the thing in-itself.” One still needs to watch Newcastle, to see what Pardew is doing or how the French midfield is performing, etc. etc. But there’s little excuse for just ignoring these clues altogether and just concocting a story that connects the dots.

Oh, and PS: Kinnear is going to help improve Newcastle for 13/14 with his astute signings and iron-clad leadership. You heard it here first. I know because I’m an expert.