With a little help from my friends

October 31, 2010 § 3 Comments

Remember the words of the Lord Jesus, how he said, it is more blessed to give than to receive. — Acts, 20:35

Jesus probably wasn’t talking about Nani. He probably wasn’t even talking about football; he was never the biggest fan of crosses. But everybody else should be. Because in the last calendar year, Nani has created more goals per game than any other wide player in English domestic football, a quite magnificent 19 in 35 games. That’s 0.54 assists per game, just a tick over one in two.

By way of comparison, over the same period Andrei Arshavin has created 12 goals in 41 games (0.29/game), Aaron Lennon 8 in 27 (0.3), Luis Antonio Valencia 11 in 39 (0.28), Florent Malouda 13 in 51 (0.25), Theo Walcott 4 in 35 (0.11), and Dirk Kuyt a lethal 3 in 45 (0.07).

Now, obviously a few of those players are more regular goalscorers in their own right, and this isn’t an exercise in A is better than B oneupmanship. But in the words of the excellent Daniel Harris, this time last year Nani was “as frustrating as a pair of rubber pants“. It’s now no exaggeration to describe him as Manchester United’s most dangerous attacking player, and he’s either made or scored roughly a third of the team’s goals this season.

The turning point appears to have been the 3-1 away at Arsenal in January 2010, possibly the first time he displayed decision-making to match his talent, and he’s improved nearly every game since. It could be argued that his display that day owed as much to the vanishing form of Gael Clichy and the careless hands of Manuel Almunia as to Nani’s own performance. But before that game he’d started 7 of 23 games in the Premier League; after that he started all but four, two of those due to suspension, and has barely been rested this term.

That said, the assist is a curious statistic; what we might call a “flat stat”, in that the information presented tells a simple story but doesn’t provide any depth or context. Most common footballing statistics are flat to a certain extent – goals scored, fouls conceded, corners won, and so on – and getting an assist, on a fundamental level, means nothing more than touching the ball last before the person that scores does so. As such, the figure is skewed in all sorts of ways. A player’s assist count can diminish through no fault of his own if he’s playing alongside an out-of-form striker, just as a strikers’ goal return can diminish if his team-mates are off their game.

There is also no qualitative aspect to an assist. In the video below, the assist goes to Dani Alves, despite the creation of the goal belonging to Xavi Hernandez, whose defence-splitting pass tees up Alves to tap across for Leo Messi.

Here, Anderson gets an assist for Cristiano Ronaldo’s wonder strike against Porto, despite the goal itself being wholly contingent on Cristiano Ronaldo’s frankly insulting talent (turn your volume down for this one).

All of which is simply to reinforce the point that assessing football through the simple totting up of the events leading to a goal is to arrive at a partial understanding of the game at best. There is no doubt that Nani has been playing wonderfully well for almost a year now; he’s been a phenomenal creative force for club and country, and had a shoulder injury not ruled him out of the World Cup he’d surely have made the Ballon d’Or shortlist. Yet it seems curious that he could have played just as well, in and of himself, for half or twice the number of assists.

Football is in need of deeper numbers. Bill James described baseball statistics as having acquired the powers of language; to adapt the metaphor, popular footballing statistics are still dribbling on their chin, pointing vacantly at the television, and burbling ‘Goal! Goal!’

Advertisements

Tagged: , , , , , , , , , , , , ,

§ 3 Responses to With a little help from my friends

  • […] This post was mentioned on Twitter by Robert Martinez, Andy Coyle, Andy Coyle, Roy Keane's Gum, Andrew Thomas and others. Andrew Thomas said: NEW POST! Twisted Blood on Nani, numbers, and the importance of helping your fellow man: http://bit.ly/bFXIy8 / comments & retweets welcome! […]

  • I think that the approach you adopted previously (compare team including player X with team sans player x) works well for these kind of comparisons and offers reasonable depth with two big caveats:

    both of your cohorts need to be reasonably big

    there needs to be a reasonable degree of consistency in your other variables.

    Clearly therefore there’s only going to be a small subsection of players which can be analysed in this way. As for the remaining 90% weaintgotnohistory might be onto something when he talks about a system based around how close you were to the ball when good things happen. The fear would be fallling into James’ error error (the best way to avoid committing an error is to be nowhere near the ball) but that problem is not insurmountable (especially as in football, far more than in baseball, your measurable job in many instances is to get close to the ball)

    • twistedblood says:

      Absolutely. The reason team-with-X vs. team-with-not-X worked for van Persie is because he is (a) first choice and (b) injured half the time, giving you two decent samples.

      (Christ, I’ve written “with-not-X” rather than “without-X”. Good to know some of my logic stuck.)

      What I like about the ball location theory is that it would allow to perhaps model the ideal places for the ball to go before the ball goes to the goal (it’s eventual destination, assuming you’re attacking). Then it could give you a map of the pitch, with preferable routes. Which I find very appealing.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

What’s this?

You are currently reading With a little help from my friends at Twisted Blood.

meta

%d bloggers like this: