clock menu more-arrow no yes

Filed under:

Bringing Usage Rate to Soccer

New, 6 comments

Using a basketball metric to measure who ends FC Dallas possessions in the offensive half.

Soccer is still relatively new to advanced statistics, and so often it is advantageous for footie analysts to adapt approaches taken by other sports who have a longer track record of quantifying athletic feats.

When watching FC Dallas play over the last few years, it has often struck me that certain players tend to hold onto the ball longer than they should, force shots from speculative angles and distances, over-dribble, and often make a mess of things even when they do pass.

In basketball, such a player is called a ball hog, and their selfishness is quantified by a stat called usage rate. Usage rate is simply defined as a player's shots and turnovers, divided by the team's total shots plus turnovers while he is on the floor. If a five-person team ended possessions equally, each of their usage rates would be 20%. When an inefficient offensive player uses up far more possessions on turnovers and bad shots, he is colloquially referred to as a ball hog.

Method for Soccer Usage Rate

Hopefully adapting this metric to soccer will be instructive. Shots and turnovers are pretty easy to translate, since Opta Chalkboards count all shot outcomes, and have a master stat for turnovers called "Tackled and Possession Lost." From there, using the timeline below the Chalkboard to select the minutes that each player played at even strength provides the team-level denominator. Periods with man-advantages in either direction are excluded to prevent red cards from skewing the study.

One tricky part of bringing usage rate to soccer is deciding how to weed out turnovers that have little to do with the flow of offense. Goalkeepers and some of their backlinesmen tend to bomb 50/50 balls toward the middle of the field. While these moments are worth discussing at another time, they are not really a sign of the type of selfishness disrupting the attack that we're looking for here. So, for this first attempt at soccer usage rate, I decided to only count turnovers that the player and team originated in the offensive half.

<script type="text/javascript" src=""></script><div class="tableauPlaceholder" style="width:544px; height:815px;"><noscript><a href="#"><img alt=" " src="http:&#47;&#47;;static&#47;images&#47;us&#47;usagerate&#47;Dashboard1&#47;1_rss.png" style="border: none" /></a></noscript><object class="tableauViz" width="544" height="815" style="display:none;"><param name="host_url" value="" /><param name="site_root" value="" /><param name="name" value="usagerate&#47;Dashboard1" /><param name="tabs" value="yes" /><param name="toolbar" value="yes" /><param name="static_image" value="http:&#47;&#47;;static&#47;images&#47;us&#47;usagerate&#47;Dashboard1&#47;1.png" /><param name="animate_transition" value="yes" /><param name="display_static_image" value="yes" /><param name="display_spinner" value="yes" /><param name="display_overlay" value="yes" /><param name="display_count" value="yes" /></object></div><div style="width:544px;height:22px;padding:0px 10px 0px 0px;color:black;font:normal 8pt verdana,helvetica,arial,sans-serif;"><div style="float:right; padding-right:8px;"><a href="" target="_blank">Learn About Tableau</a></div></div>

As Dallas fans would hope, David Ferreira leads this stat, even over a small small sample in which he played only a little over half of the available minutes. But then a couple points of concern pop up. Fabian Castillo is nowhere near Ferreira's equal, but he's ending almost as many Dallas attacks. I discussed some of Castillo's issues last year, and much of it still holds true.

Jackson's numbers here are concerning as well, but mainly his blue bar in the top graph mainly so, showing that 4.3% of FCD attacks in the last few weeks have ended with one of his shots, just behind Kenny Cooper's team-high 4.5%. The second graph, detailing each player's historic MLS strike records shows why this is a problem. Jackson is the only Dallas attacker who's below the strike rate confidence band, meaning that he just doesn't score enough to justify the number of shots he takes.

I know that the young Brazilian scored on Saturday, but unless he has suddenly becomes a dramatically better scores, there is little reason to think that he should be the one shooting unless the chance is extremely promising (like his season-opening goal that was gift-wrapped by a Stewart Ceus brainfart) or if Blas Perez, Cooper, Eric Hassli, and even Castillo aren't options in that moment.

Yes, Jackson and Castillo are thrilling to watch, but in their enthusiasm they simply burn out far too many FC Dallas attacks. On the other side, this team needs to get the ball to Blas Perez more often. His record would indicate that he is the best pure finisher on the team, but lately he's taken roughly the same ratio of Dallas' shots as two defensive midfielders.

Difficulties in Applying Usage Rate to Soccer

In the NBA analysis, usage rate is often a secondary metric. For example, Dirk Nowitzki is jaw-droppingly efficient, as evidenced by his career true shooting percentage of 58.1%, and the Mavericks are smart to design their offense so that he has finished 27% of plays. A usage rate that high can be a major problem, though, if it belong s to an inefficient player like Josh Smith.

In soccer we don't have terribly specific metrics for a player's efficiency. Jackson's shooting statistics certainly point to a lack of finishing ability, but he's shot less in 4 seasons than Nowitzki does in 9 games, there's a wider variety of shot quality in soccer than in basketball, and when the opposing keeper is in position very few shots are as likely to score as an average NBA player's three-point attempt.

It's very likely that Jackson isn't a great finisher, but we have to be aware that his ability in this area could simply be hidden amidst small sample size, and differing quality of shots provided to him. Also, soccer just doesn't have the same type of play scripting or substitution options as basketball, so it's more difficult for a coach to adjust their gameplan to prevent their ball hogs from disrupting things.

It is worth noting that this approach to usage rate for soccer probably inflates turnovers for set piece takers. Michel and Ferreira's turnover figures are likely driven in large part because even well-taken free kicks and corner kicks are still inherently speculative, and are often unsuccessful through little fault of the taker. In that light, Castillo's number looks even more out-of-place.

Finally, the stats in this analysis come from a tiny sample because I didn't have time to gather any comparables for other clubs or look back any further for Dallas.

Where could this stat take us?

If Opta, Infostrada, or another statistics company were to track usage rate, they could systematically deal with some of these issues quite easily. Set pieces could be held out of the analysis, less-relevant turnovers like goal kicks could be excluded automatically, instead of lopping off everything in the defensive half.

Then we can quickly look at different clubs, see where usage rates tend to fall for players in similar roles, and give some real context to our findings.

But this metric would really start to come in handy when we can really quantify the ability to make the final pass, and to score, showing us who we really want to be in charge of how much of the offense.