Stats cannot explain football’s
beauty or its artistry
Paul
GARDNER
WORLDWIDE VIEWshooting was hopelessly feeble?
I’ve been discussing team stats,
and I’ve been inescapably drawn into
single-player stats. Now, I’m suddenly
talking about the strengths and
weaknesses of individual players.
Which gets us where we need to
be to expose the fallacy of stats.
I am not exactly doing anything
original when I insist on defining soccer
in terms of its players. But that is
something that needs constant
emphasis these days; days in which
the statisticians and the computer are
working to overturn that notion.
Of course, that is not their intention,
but every statistic that submerges
individual skills in a percentage, and
every analysis that ignores artistry
because it is viewed as anarchy, works
to downplay the importance of players
and to exult the idea of a more
programmed activity.
But we’ve just seen how a statistical
approach fails to reflect the reality of
the sport. So, down with statistics?
Death to statistics? Not at all. If we
must have soccer stats at least let
them be intelligent ones.
I have already dismissed the
“clearance” category as a nonsense.
“Possession” too has its problems. At
the moment that stat is pretty useless,
a number that is totally unreliable as a
guide to anything.
But that stat can surely be refined.
Possession in the opponent’s half
would tell us more. As for merely
counting “passes”, again the category
needs to be broken down. The possibleIf we must have game stats could we
at least have intelligent ones?
The more game statistics I see, the
more irritated I get because I’m never
sure what these figures are trying to
tell me. I’ve seen a set of stats that
includes, as a category, “clearances”.
What the heck is a clearance? A hefty
whack, into touch probably. Is that a
positive or a negative stat?
By way of introduction that’s a
frivolous touch, but I have a more
serious aim. I want to question the
value of the soccer stats that are
currently thrust before us.
We get two versions of stats: pre-
game and post-game.
If pre-game stats have any use at
all, it ought to be as a guide to what is
about to happen, to provide a reliable
mathematical basis to work out who’s
going to win.
Who is going to believe that?
Anyone who does will be busy
filling out betting slips all day and
presumably making a lot of money.
A believing coach should be winning
a lot of games. As none of that
happens, it’s worth asking: are these
game stats in any sense reliable?
Let’s switch to post-game stats. Take
a look at the following stats and tell
me who won...TEAM A TEAM B
32% Possession 68%
12 Shots 11
7 Shots On Target^7
3 Corner Kicks 7
8 Fouls 6
521 Touches 901
338 Passes 716
10 Tackles 13
1 Offside 6Team B has huge advantages in
three categories: more possession,
more touches and more passes. But
the game-winning implication of those
solid Team B stats looks iffy when you
see that Team A is very slightly ahead
on shots.
Nevertheless, there is that massive
advantage in ball control. I mean, two
thirds of the game? That has to mean
something. So you pick Team B. And
you are wrong. Massively wrong.
This was a Premier League game.
Team A, Bournemouth, not only beat
Team B, Chelsea, they beat them 4-0.
Looking again at the scoreline, a stat
showing goalkeeper saves might help.
Without it, we might envisage a 7-7 tie.
The Premier League does not issue
that stat. Maybe the Bournemouth
keeper was in great form? Or maybe
he didn’t need to be as Chelsea’sSave...Bournemouth
keeper Artur BorucNumbers game...
Chelsea crowd out
David Brooks of
BournemouthTHE WORLD THIS MONTH