Armchair Analyst: What we've learned from Opta
A friend of mine is engaged to a mathematician, a lovely girl who’s pursuing a doctorate in number-crunching.
Over lunch on Thursday afternoon, I brought up the partnership MLS has with Opta, the rise of soccermetrics and some of the stats we’ve put together through the season’s first several weeks.
Rather than react with glee, she spent the rest of the day trying to convince me that 28 games isn’t a large enough sample-size to determine any statistical trends with an appropriate level of authority, that the margin of error would be enormous, and doing something as simple and commonplace as running a binary logistical regression on the data would prove almost worthless.
Luckily for all of you, I don’t know what a binary logistical regression is, and the threat of a gigantic margin of error has never stopped me from speaking out of turn.
Let’s dive into the numbers.
So far there have been 28 games in this MLS season. The home team is 12-6-10 in those games. Nineteen of the 28 games have had at least one first-half goal, and 27 of the 28 games have had at least one goal (the only scoreless draw of the season came when the Red Bulls played at Columbus in Week 2).
And while there have been a number of pretty remarkable comebacks already, the team that scores first is still undefeated at 18-0-9.
What the Opta numbers reveal should be interesting – but please, keep my mathematician friend’s warning in mind. These numbers are fun to look at and play with, but the sample size is woefully small and the predictive value from a mathematics standpoint is minimal.
“Possession” is the first and most basic soccermetric.
[inline_node:324393]If you’ve been watching European or certain international games over the past 10 years, you’ve doubtless seen possession stats flashed across the screen at some point. Over the past three years, as Barcelona have raised the possession game to an art form, “The Cult of the Pass” has blossomed. (Full disclosure: I’m a card-carrying member.)
But how does possession translate for teams that don’t boast the three best players in the world and an academy that’s been simmering to perfection for 40 years?
The answer so far, is “not that well.” Teams holding more possession through the season’s first 28 games have won just 25 percent of the time, clocking in at 7-11-10 overall. The team with the best possession numbers are the Red Bulls, who absolutely dominated in their two home games and did a decent job with a makeshift lineup at Columbus as well. True to form, New York are 1-0-2.
The Red Bulls, like many other teams throughout the world, have had trouble converting that dominance on the ball into goals. They’ve scored just twice in their three games, with neither goal coming off a sustained build-up.
While the Red Bulls are purpose-built to hold the ball, in many other instances it seems that scoreline-specific tactics influence the stats. That’s a fancy way of saying, “If you don’t score first, you’re finished.”
[inline_node:331519]Conventional wisdom says that in MLS, the team that gets on the board first tends to sit back afterward and invite attacks, which in turn gives them more chances to counterattack. The numbers thus far back up that assertion as the team that’s been first to score has won the possession battle in just nine of the 27 games where the goal has been breached.
Another bit of conventional wisdom is that home teams chase the win, while road teams chase the result. This turns out to be another bit of axiomatic thinking that’s backed up by Opta’s numbers.
In the 20 times this season that the home team has entered halftime either trailing or tied, 15 of those games ended with the home side winning the possession battle. The numbers make it clear that “chasing the win” means getting on the ball and launching attacks as often as possible, while “chasing the result” means sitting back and protecting the 0-0 or 0-1 score line while hitting on the counter.
In the other eight games, the ones where the home team led at the half, the numbers are almost reversed. Home teams have won the possession battle under those circumstances just thrice, or 37.5 percent of the time. That’s exactly half as frequently as they win the possession battle when tied or trailing at the break.
Does this mean Sigi Schmid was correct when he called MLS “a counterattack league” in March? As with everything else, it's too early to tell.
[inline_node:332787]But so far, nobody’s been more guilty of playing by the rule of the counter than the Sounders. In their two home games (both of which saw Seattle tied or trailing at the half), they out-possessed, out-passed and outshot their opponent.
In their two road games (one in which they were protecting a lead, the other protecting a draw), the numbers say the Sounders bunkered for their lives. They allowed 541 passes and 63 percent possession to the Red Bulls in what was eventually a 1-0 loss in Week 1, and 453 passes and 58 percent possession to the San Jose Earthquakes (a much more direct team than RBNY) in what became a 2-2 draw last weekend.
That 1-0 New York win over Seattle last month, by the way, is one of just five games in which a team has attempted more than 500 passes, and the only one of the five to take all three points. The Red Bulls had a league-high 545 passes and the LA Galaxy 538 in 1-1 draws against Houston and New England, respectively. Vancouver allowed 509 passes in countering Toronto to death, 4-2, in Week 1. And the Rapids gave up 540 passes to Chivas USA in a clinical 1-0 Week 2 triumph.
Yes, that shows bunkering really can work, and often does. What’ll be interesting, though, is a revisit of those numbers in a few months when offenses have shaken off their early-season rust. My hunch is that things will even out as the season goes on.
And if I’m wrong, we’ll have the stats to prove it.
Matthew Doyle can be reached for comment at email@example.com and followed on Twitter at @MLS_Analyst.
<p><em><a href="http://www.mlssoccer.com/extratime" target="_blank"><img src="http://www.mlssoccer.com/sites/default/files/new-etr-banner.jpg" alt="" width="620" height="100" /></a><br /></em></p>