Steve Fenn with analysis of who can shoot the rock and who can’t.
What’s more important:
» The “moment of truth” when a striker tries to score and the opposing keeper tries to stop him….
» ….or the contributions and failures of everyone on the pitch that conspired to determine the quality of the striker’s opportunity?
A single match’s shots–and whether they are saved–determines the outcome, but the quality of shots earned and allowed is a more reliable barometer of the team as a whole.
It’s a base paradox of observing this game. Because that final moment is so dramatically important, everyone’s memory clings to strikes and saves, lessening the relative value of myriad brilliances and mistakes which create and allow chances in promising situations.
Right now 14 MLS teams are bunched so closely together in the race for 10 playoff spots that it’s difficult to tell many of them apart if you scale their ranking based on points per game.
As Zach Slaton pointed out this week in Forbes, narrow point leads are unreliable markers of relative team quality, so in the present MLS situation we need a better way to gauge clubs’ strengths.
Thanks to shot location data patiently catalogued by American Soccer Analysis, we can see every 2013 MLS club’s shots and shots allowed broken down into 6 zones.
From the data, it’s not hard to get a scoring expectation (xG) from shots in every zone. This is based on the overall MLS averages and the number of shots-for and shots-against in each zone.
So far this season, players score on 32.6% of shots taken inside the 6-yard box (1 in the diagram below), while those attempting their best Gareth Bale impersonation from 25-plus yards away from goal (5) find the net on a measly 2.23% of such shots.
Attempts from inside the box but wide, and those from a little beyond the box are scoring 6.5% and 5.1% of the time, respectively.
Really wide shots have a low 3.64% strike rate, but this is the most troublesome region since it is often hard for stat keepers to tell a bad cross that got closer than intended to the keeper from an audacious shot. 2011 Brek Shea will take it though. Thankfully, this region features the fewest shots this season, lessening its impact on analysis.
Comparing overall xG to each club’s goals and goals allowed quantifies which clubs are likely hovering above their most likely level, and who’s most likely underrated based on points and goals.
Below we have a visualization of all this, with square sizes based on shots per game. The coloring for MLS averages is pretty straightforward with purpleness indicating likelihood of scoring per shot, but the club-level data can be tricky. On offense, the blueness of a square conveys how much it’s strikers have exceeded the expectation for that zone, while the depth of an orange hue signifying how much worse they’ve been versus the MLS average.
On defense it’s the same for the opposing strikers, so a club whose strikers & keepers have over-performed will have some blue squares on offense and orange on defense. This is usually most striking in zone 2, where the average takes 3.9 shots per match, and 17.81% of them have been goals.
A quick note on the predictive value of xG. Splitting the season between March-May and June-present, expected goal differential (xGD) from the first half correlates to Points Per Game and GD in the second at R² of .2504 and .4316, respectively.
For comparison, PPG predicted at .0131 (PPG) and .0430 (GD) and GD yielded R² of .0332 (PPG) and .0509.
For the math-averse who have made it this far: points and goal differential in the first three months of the season have been almost entirely unrelated to results since, but xGD has been immensely more predictive. Since fans and pundits tend to, consciously or subconsciously, use recent points and goal differential as main drivers of their expectations, this has big implications for how would should be observing the game. But don’t go running to the nearest sport book just yet. This is still a relatively small sample and there are certainly other factors at play, but shot location does gives us a better picture of teams’ levels.
As you can see, LA Galaxy and Sporting Kansas City far and away come out best in xGD.
Their offenses regularly shoot from promising positions (mores so for LA Robbie Keane thank you very much), and both of their defenses prevent the same.
In LA’s case, their scoring is very much in line with xG, but they have allowed far more than shot locations would predict. Galaxy optimists would probably dub this The Cudicini Effect, now that Jaime Penedo has sent the shaky Italian to the bench. (Editor’s note: Love Jaime Penedo. See: June 11, 2011)
On the other end of the spectrum are some of the usual suspects in DC United and Chivas USA, but not far above them lurks a couple surprises. The playoff-contending Vancouver Whitecaps and the much-vaunted Real Salt Lake have been surpassing the expectation of their shot locations for & against.
RSL leads the league in scoring, but this data presents a quandary. Are they excelling in every part of scoring that this model doesn’t capture, or are they due for some harsh regression to the mean?
Because it has to be noted that data on the positioning of the opposing keeper and his defenders isn’t available publicly right now. Obviously, that would be a vital ingredient to shot quality, as would be information on whether the strike comes from a set piece, or is struck with the preferred foot or the head.
Specifically, penalty kicks which have a greater xG than even zone 1 shots haven’t been weeded out yet. Also, if we had more detailed shot locations at our disposal it is almost certain we would see further xG variation within each of these 6 zones.
For now though, the most relevant findings of this study are that Los Angeles and Kansas City have been underperforming and Real Salt Lake & Vancouver are flying higher than they deserve.
Some faults and strengths are certainly beyond the scope of this study, but it seems quite likely that the true level of each club is probably closer their xGD than to their points or raw GD. After all, shot locations are determined in small and large ways by many different players on both sides of the ball, while the resultant “moments of truth” are subject to the fickle skills of finishing and keeping, where chaotic bounces and spins of the ball rule the day.