The Problem with Soccer by the Numbers

By Paul Gardner

Spurs coach Andre Villas-Boas was no doubt overstating things when he recently condemned soccer statistics as "useless." But his opinion is refreshing, to say the least.

Because it is not often that a top coach speaks out, forthrightly, against any of the various systems and methods and plans and operations that now infest the sport. Virtually all of these programs rely on statistics, which means that they ultimately reduce the playing of soccer -- all that movement, all those decisions, all that stress and sweat energy -- to numbers.

Right from the start, this sounds like a pretty questionable, not to say outright loony, notion. But we must bow to the eternal plea of the statisticians -- that their stats are, indeed, simply numbers. Neutral numbers. They are not tainted by any sort of bias -- political or religious or sporting -- they are simply bland measurements. The problems with stats arrive with how they are interpreted, or rather, with who is doing the interpreting. That will be a real, live human person, and he, or she, almost certainly will have biases and prejudices. Some of them overt, but others rather well hidden, quite possibly hidden from the interpreters themselves.

Soccer stats, from where I’m sitting, seem to be ideally formulated for misinterpretation. I can begin with some absurd goalkeeper stats. Stats that are likely to be used in a way that defies common sense -- yet, used they are every day. Thus, it is common practice to credit goalkeepers with shutouts. If you look among the official MLS stats for goalkeepers, you will find a column headed ShO.

A low shutout total is considered to be a good thing, a stat that really says something about a keeper’s ability to keep the ball out of the net. What a farce. Everyone in the game -- and most of all, the keepers themselves -- knows that repeated shutouts come not from goalkeeper heroics, but from sturdy team defending. A goalkeeper with a strong defense in front of him is like to have very little to do, only the occasional shot to save. Yet the sole, individual, merit for the shutouts goes to him.

Despite the fact that the stats don’t say anything of the sort. The stat merely shows X number of shutouts for a particular team. Linking the number to the goalkeeper -- well, who the hell had that idea? A goalkeeper, I’d guess. But it’s obvious nonsense.

There is also this business of possession time, which has come to loom large among the soccer stats. And no one is quite sure what it means.

It seems logical to assume that more possession means more likelihood of winning the game. That in turn, should mean more goals scored. But I’ve never seen a possession stat linked to a goals-scored stat, so that may not be the case.

Because the possession stat is dubious right from the start. It sounds good as a measure of a team’s dominance, but even a quick investigation of the figure tells you that it’s pretty meaningless -- because there is possession, and then there’s possession. Passing the ball idly around in your own half, laterally and backward mostly, will up your possession stats, but as a guide to game dominance, or to scoring goals, that sort of possession doesn’t mean too much.

No doubt soccer stats will get more and more sophisticated -- that is to say complicated -- quite possibly to the point where only an approved and duly diploma-ed member of the Association of Soccer Statisticians (that’s ASS, acronymically speaking) will be allowed to interpret them.

For the moment we have a coach like Villas-Boas saying he doesn’t pay much attention to them -- and his Tottenham Hotspur is playing well and lie third in the English Premier League. We also have a coach like Sam Allardyce who does like to use computer-based stats -- and his West Ham United are playing poorly and are in 14th place in the EPL.

That comparison also highlights what I consider the biggest objection to the use of stats as a coaching tool: That they create an impression that everything in the game can be measured and reduced to numbers -- which can then be “worked on” to make things better.

The enormous fallacy of that argument being that stats cannot measure the quality of play. This is an aesthetic judgment, way beyond the reach of numerical classifications. The aesthetics have to do with human qualities of human beings, they often defy norms and expectations. I recall the answer that the great ballet dancer Nijinsky gave when asked how he was able to make such prodigious jumps, seeming to defy gravity by the length of time he stayed airborne. He replied “I merely leap and pause.”

Maybe Ronaldo does the same. But no stat, nor any clever interpretation of a stat, is ever going to capture the essence or the beauty of those actions.

A coach who relies on stats is unlikely to be too concerned about the aesthetics of the game. Which is why I find Villas-Boas’s put-down of stats so appealing. Whatever may be his reasoning it is encouraging to have a coach remind us that there is more to soccer than the arid regimentation of numbers.

15 comments about "The Problem with Soccer by the Numbers".
  1. Jon Deeny, February 27, 2013 at 4:40 p.m.

    I agree with the logic behind not giving the credit to the goalie for cleansheets, but by the same token many goals are not the keeper's fault yet the goals against average is a keeper statistic, good or bad.

  2. Paul Lorinczi, February 27, 2013 at 4:51 p.m.

    Sorry Paul, I am going to have to disagree with you on this one.

    If a manager is using stats as a black and white decision making tool, it is a wrong way to use them.

    Stats in the modern game are needed because players are indeed assets to the company. There are key indicators that help a manager understand a players fitness. Why do you think Man United is getting high productivity out of players like Ryan Giggs and Paul Scholes? In the old days, they would have run the old horses into the ground.

    Stats do help provide a complete picture of what is happening. Stats in of themselves do not help. It is like anything else in life, they provide a picture when experience alone can't answer a question.

  3. Miguel Dedo, February 27, 2013 at 4:59 p.m.

    Not a statistic, but another misleading part of soccer reporting is "Goal of the week." To me, a great goal is a team effort; nine passes twisting defenders around until at the end the "goal" is a easy tap across the goal line.
    A bicycle kick form 20 meters is a great circus act, an unusual goal, but not a great goal.

  4. Steve Gabbard, February 27, 2013 at 6:20 p.m.

    I teach statistics and research methods. I was a soccer referee. There is a reason that Twain described lies in 3 categories. Lies, Damn Lies and Statistics. It is not inherent in the stats, it is the issue with what constructs/concepts are proposed and what statistical measures are used as surrogates to quantify them. Construct = goalkeeper quality; stat = ShO. OK -- simple but ultimately useless as pointed out. Saves/(attempt*difficult index)? Saves/PK attempt? Net punt yards? Some amalgam? If the RIGHT set of stats were established that the majority of the sport agreed were meaningful to describe a particular quality parameter...fine. Otherwise we reduce soccer to stupid stats the way we reduce politics to sound-bites. On the other hand Gardner is hoisted by his own petard, citing 2 anecdotes about coaches stats use and the EPL placement. Better he should correlate the entire leagues use of stats and placement to show its folly. Using an counter-example to a 'rule' could negate virtually any useful heuristic, because ultimately they all have contrary examples.

  5. beautiful game, February 27, 2013 at 6:58 p.m.

    Stats can be subjective. Yet,one stat that is completely missing in MLS, are unforced turnovers/giveaways...commentators never mention it; as if that is a taboo.

  6. Ken Jamieson, February 27, 2013 at 7:04 p.m.

    The increasing amount of statistics kept in soccer should be seen as a positive sign for soccer enthusiasts on this side of the Atlantic. American sports are stats-driven, without stats baseball is a meaningless and boring waste of an afternoon. Every NFL position is scrutinized by stats, especially QBs.
    The fact that soccer is being subjected to more and more statistical analysis means it is being accepted as a mainstream sport. All we need now is a soccer equivalent to On-Base-Percentage.

  7. Daniel Smith, February 27, 2013 at 7:36 p.m.

    I agree with you on THIS one Paul! Trying to measure effective and beautiful football with stats is like trying to use statistics to determine beautiful art or a great movie! If we pursue the results at all costs approach - out game will become the equivalent of the 'formula blockbuster movie'... Throw in a star or two, some great action CGI, a hot woman or two, and $20m in marketing!!!!
    Pursue the game as an art....

  8. Kent James, February 27, 2013 at 8:45 p.m.

    Stats are only as good as the use to which they are put. Many years ago, my college coach recorded how many "unforced errors" each player made (good call, I w Nowozeniek; I thought of my example before I saw your post, but that just confirms its relevance). Having statistical measures can document tendencies players have, good or bad. And it's the coaching that decides which tendencies are good, and should be encouraged, and which should be discouraged. Yes, stats can be misleading, or useless, but used properly, they're just additional information to help coaches and players (and fans) figure out this simple game, that somehow eludes being figured out!

  9. david fernandes, February 27, 2013 at 8:48 p.m.

    I agree mr. Paul people always have that misinterpretation of dominance and superior. A good example would be a.c. milan vs. Barca this past week, where barca, like usual, had that great possession yet could not find the net, while milan not only scored twice also created more attack than barca. Unlike vilas novas i wouldnt ignore stats but i wouldnt rely solely on it.

  10. John Selvey, February 27, 2013 at 8:54 p.m.

    Relative to most of Mr. Gardners opinions, this is one of his better efforts. I agree!!

  11. Saverio Colantonio, February 27, 2013 at 9:19 p.m.


    How many times have told us about how statistically speaking, crosses into the penalty box are of a low percentage when it comes to scoring and in-swingers from corner kicks have a higher chance of a goal being scored. So Paul, its not whether or not to use statistics, however, it is crucial to know what statistics are important.

  12. Ramon Creager, February 28, 2013 at 10:29 a.m.

    Statistics in soccer are a relatively new phenomenon. I remember when virtually the only stat I saw was who scored goals. Then came assists. Next thing you know I see things like fouls committed, corners, distance run, pass accuracy, possession, etc. etc. I don't think knowledgeable coaches and observers have had a chance to synthesize a meaning from all these stats yet, the way they have in American Football. (Compare, for instance, the crude clean sheet stat for GK to the AF QBR for quarterbacks. See also what the Football done with statistical analysis). I foresee a time where this kind of understanding will come to soccer, like it or not, and may indeed be helpful in lending some solid understanding to what many intuitively know. In the end though it will not change this about our favorite game: There are many different and interesting ways to play this game, and each can bring success; and you still need smart coaches and player personnel types.

  13. Kent James, March 1, 2013 at 4:31 p.m.

    Stanley, while I agree that statistics can be useful (especially if they are able to document (or dispel) subjective opinions (it would be hard to suggest you dominated a game if they had more possession that you, e.g.), but even the two seemingly basic stats can be misleading. I don't doubt that Greaves being offside 8x in one game was a problem (though one might not need stats to notice it, though they might be required to prove to Greaves it was a problem). But if you punish people for being offside, they may never get close to offside, which actually may diminish your teams ability to get behind the defender's line, while players who are caught offside occasionally, may be getting behind the defense 3 out of 4 times (which is worth being caught offside for). And while shots on goal is one of the more obvious things to work for, a shot right at the middle of the goal is on target, but not a good shot. And I've wondered (as a defender who is trying to coach attackers), do the best attackers aim for just inside the post, which will increase the chances of shooting wide (with no chance to score), but increase the chance of scoring if the shot is on goal, or a yard inside the post, which puts more shots on goal, but also more saves. It would be hard to measure statistically (since you can't tell where a player is aiming, only where they send the ball), but you might be able to show that a shot within a foot of the post as x% chance of going in, 2' inside has x-y% chance to go it, etc. So I think statistics can be valuable tools, but as Ramon said, it will take a while to see how useful they are. I sense, however, that as a devotee of the artistic side of soccer, Paul Gardner hopes they won't be useful...but I'd sure use them if they were. And maybe, PG, statistics will demonstrate that greater artistry leads to greater success!

  14. Martha Diop, April 8, 2013 at 6:01 a.m.

    We as human being make decision in hope of a positive. Since we do not know the future, using stats for decision making come handy, but also may just provide comfort, as we rely on the simple assumption that what happened in the past is likely to happen in the future.
    If a player has scored many goals in the past, is not it obvious that if I buy him, he will score many goals for my team? Or maybe it is not obvious?
    Once they were two roads going through the forest haunted by a fierce wolf. Stats showed that 90% of those who traveled on road A were eaten, while 90% of those who traveled on road B escaped the wolf. Pretty obvious that one should take road B. My 10 year old keep telling me that I still have 50-50 chance of being eaten. I still have not figured out if he is right or wrong

  15. Henry Lamontagne, June 2, 2013 at 6:42 p.m.

    The problem with soccer by the numbers is that you are looking at the wrong numbers. If we want to evaluate a Goalkeeper we may look at saves as a % of shots on goal. If we want to evaluate the effectiveness of our defense we might look at how much time the opponent spent in our defensive 1/3. If you pick the right numbers to track, they can be very meaningful & useful.

Next story loading loading..

Discover Our Publications