Commentary

Soccer and Metrics (Part 2): A troubled partnership

By Paul Gardner

OK then. The aim of metrics in soccer is to improve player performance. After that, the problems begin. The miCoach program lists some of the things that it can improve -- player position, power output, speed, distance covered, intensity of play, acceleration and GPS heat mapping. And those awful TV commercials told us that “This is how the fast get faster, the strong get stronger ...”

These are basically physical, measurable attributes. But is it logical to assume that faster, stronger players are going to give us better soccer? How to define “better,” or even “good” soccer anyway?

Which is where miCoach and all similar soccer programs suddenly look rather hollow. Because any definition of soccer based on technology -- and the sort of metrics listed above -- is bound to come up short. Pretty disastrously short, in my view, for certain vital elements of the sport are missing. Are bound to be missing, because they lie beyond technology’s ability to measure them.

For instance. Creativity. Subtlety. Artistry. Can you have soccer without those qualities? Unfortunately, yes, you can. And without them, a pretty dismal sport soccer becomes. In effect, a soul-less sport dominated by metrics, which deal only with Facts.

Charles Dickens’ Mr. Gradgrind would have approved. It was his belief that “In this life we want nothing but Facts, sir. Nothing but Facts.” The words were written over 150 years ago in the novel “Hard Times.” Gradgrind is a heartless, mechanical man, a distant warning of how relying on facts might transmute soccer -- into a heartless, mechanical sport.

I have no doubt that the metrics boffins are aware of the danger. Hence that intimidating claim at the end of the miCoach commercial -- “THIS is smart soccer” -- which seems to imply an intellectual component.

But what’s so smart about running faster or getting stronger or covering more ground? Those are purely physical talents and improving them may mean nothing but a more physical version of soccer. Is that better soccer?

The metricians have an answer to that criticism, one that deserves attention. They avow that the information all their technology makes available is completely neutral. What we’re getting from them are modern, dynamic statistics -- real-time statistics? -- and it has been acknowledged for decades now that stats have no positive or negative value ... until they are interpreted. A state of affairs that has given stats a bad name, as it implies that they can be made to mean anything you wish.

For soccer metrics, we can assume that it’s the coach who makes the judgments, so if things don’t work out, blame the coach. But even that assumption is unsafe. There are metrics people who, having worked so hard to come up with their wonderfully sophisticated programs, feel that only another techie can really understand them.

So, for the job of making them meaningful the coach gets shoved aside and the sports scientist takes over. These days most pro clubs have a sports scientist.

The basic question remains: Are we getting better soccer? Does the growing importance of metrics and the sports scientist ensure a more enjoyable game, both from the playing and the spectating angles?

Well, I don’t know. I doubt anyone knows. My feeling is that it does not, and may well result in a less attractive game, as the importance of intangibles -- including, but not limited to, creativity, subtlety, artistry -- is quietly downgraded so as to increase the importance of the metrics.

Therein lies the most extraordinary aspect of the whole metrics scenario. For its authority, it relies heavily on the formidable weight of being scientific. Yet it lacks one of the most important elements of scientific authenticity. Proof. Is there any convincing evidence that meticulous attention to metrics pays off?

Not that I’m aware of. I’ve never heard a top pro team -- and they’re the ones who spend most money on this sort of thing -- telling us that their success is because of a metrics program. You can be sure that, if such a case existed, we would have heard about it, big time, from the marketeers. And it is a fact that most, if not all, of the stories about these programs fail to cite examples of clubs that have won things based on metrics. Most of the clubs cited, in fact, haven’t won anything.

The scientists have a legitimate gripe here -- they can argue that no coach is going to sing the praises of a set up that suggests he is being replaced by a sports scientist, maybe even by a computer.

I am not about to belittle the scientific aspects of metrics. The figures they come up with are real, and I have the feeling that I think most people have -- that these figures must give us a deeper insight into the game, that they must therefore be useful. In particular, it seems to me that the physiological metrics must prove valuable in the early detection of medical problems. And I do have a feeling that the new rulebook ban on the use of real-time data smacks of an anti-technology attitude that FIFA was thought to have abandoned with its acceptance of goal-line technology.

But the direct application of metrics to the game worries me. I see soccer as a sport, as a human activity. Definitely not a computer game. If metrics become so worshiped, so entrenched in the game that they are not questioned, then I don’t see how soccer can escape being turned into a rather different sport. Something cold, heartless, mechanical. Bleak and much less enjoyable. Something definitely Gradgrindish. A horrible fate.

Soccer and Metrics (Part 1): Beyond the marketing silliness

33 comments about "Soccer and Metrics (Part 2): A troubled partnership ".
  1. Lou vulovich, August 10, 2015 at 10:24 a.m.

    Until FIFA follows the rules and actually forces referees to call the game accordingly nothing will change. Coaches will recruit bigger stronger and faster players as it is much easier to coach a team not to loose than to coach to win. It takes very little coaching genius to find 15 athletes will a little skill and a lot of athleticism and determination
    have them play the ball back under any pressure, rotate always outside and hit it long every chance and high pressure the opponent. OH and foul your opponent every time you get near him, that is coaching and playing soccer the modern way. It reminds me of NHL goon Hockey without the fights. If Barcelona or RM are not playing it is painful to watch.

  2. Santiago 1314 replied, August 11, 2015 at 9:01 a.m.

    Excellently put, Lou

  3. Kent James, August 10, 2015 at 10:27 a.m.

    While I agree that a mechanical version of soccer is less entertaining than the artistic version that Paul advocates, sometimes power and speed are also entertaining (I like a balance). And just because you can measure something, doesn't mean you have to become a slave to it (though again, PG has a point that this sometimes happens). So PG's concern about the use of this technology has a whiff of the anti-technology sentiment he rightly attributes to FIFA's banning of the real-time use of it (though, to be fair to FIFA, one might see that as an effort to keep anything from outside the field of play from influencing the game). So while I agree that the marketing sounds ridiculous, if the technology works, I could see it being useful (for example, does the team's pace slow as the game progresses, or remain the same; the answer might determine how much fitness training the team needs (or doesn't need)). I think the more interesting question, is is it possible to use "moneyball" type analysis to soccer? In other words, can the teams with less financial backing overcome their financial deficiencies by using statistics to identify undervalued players?

  4. Santiago 1314 replied, August 11, 2015 at 9:08 a.m.

    I think you could look to Atheletico Madrid, as a Team that takes "Undervalued" Players and gets the most "Money Ball" out of them...Southampton,would be another one, but their model is more if Development vs Buying...The MLS Dynamo has had a Pretty good "Money Ball" program over the years

  5. Kenneth Barr, August 10, 2015 at 11:46 a.m.

    Paul Gardner is absolutely correct when he says this is a game played by humans, not computers. At its best, advanced metrics allow humans to maximize performance. But, as with all attempts to place numerical values on human performance, the object becomes eliminating variables and uncertaincies and that is where metrics fall down. We often see analysts talk about possession as a vital clue as to who should win a match. This use of a quantity, percentage of possession, fails to evaluate the effectiveness of that possession. Last season both Crystal Palace and Manchester United out-possessed their opponents that day, West Brom and Chelsea, by better than 2 to 1. Yet, they both lost and neither scored. After their match, Louis van Gaal bemoaned Chelsea's "negative" tactics and said his side should have won because they had more of the ball. Alan Pardew, Palace's manager, admitted that West Brom frustrated his side's attacks and hit them on the counter effectively. Pardew's evaluation shows he got what happened, LvG either was sandbagging or really felt amount of possession should determine victory. It is the quality, not the quantity, of a side's possession that wins matches.

    At the end of all this, the old bromide still applies. There are lies, damn lies and statistics.

  6. Santiago 1314 replied, August 11, 2015 at 9:14 a.m.

    And LvG, will never admit that it's "him" that stifles and intimidates his players into robots... I doubt he will make it through the year...ManU looks just like it did last year. .. $240 million just doesn't buy you what it used to...

  7. Raymond Weigand, August 10, 2015 at 11:59 a.m.

    Metrics are a 'controlling' tool ... and with any tool - they can be used to build or to destroy (inadvertently). The team manager is responsible for having a final product in mind for the team - based on the capabilities of the current resources - including time. Sometimes the creative side of soccer is at odds with the fitness side ... which, may be why DiMaria was 'given' to PSG at a huge loss to MU. For all the creative play - some coaches would still like the player to be able to participate at their best for 2 sets of 45 minutes. For example, Arjen Robben is the Energizer bunny - and is quite creative - quite fit - technique is impeccable - and is paid $MM. You may wonder - how does Robben pull the same predictable move and still succeed as often as he does? Fitness / Technique -he understands when his opponents are tired. (and he also has a creative side - to the anger/frustration of the Mexico side in the last world cup)

  8. Bob Ashpole, August 10, 2015 at 1:05 p.m.

    Information (metrics) is fine, but to be useful to a coach, the information has to be both accurate and valid. It is easy to collect information, but most information is useless. Coaches have collected and used information for generations. Certainly any improvement in information gathering and processing (over stopwatches and clipboards) should be welcomed, but I am very skeptical of IT companies. I am also very "skeptical" of any "scientist" claiming advantages for an IT product. The goal is not to gather information, but rather to gather useful information. What scientists ought to be doing is studying what information is predictive of events, not suggesting that more information generally is desirable. Misuse of "Metrics" is exactly what lead to the England FA promoting kick and rush soccer.

  9. Joe Linzner, August 10, 2015 at 2:12 p.m.

    Exactly correct, stats and metrics based soccer is a shortsighted approach to all sports. Possession is a stat that is useless unless it produces goals, pass completion/pass attempts is another stat that i find suspect. It is easier to complete a backpass but in the opponent's third,attacking passes are generally not so easy. While physical condition plays a large part in developing speed/strength in soccer that is less important than quickness and speed of thought. Vision as to available space (advantageous space) is often more important than individual skill. Metrics in misused more often than applying it usefully. The USA has a tendency to use taller, bigger, athletes that actual good soccer players. First touches are still uncertain, one on one dribbling skills sorely lacking and as far as our talent pool being deeper, it just may be, but in no way is it better.

  10. Lou vulovich, August 10, 2015 at 3:43 p.m.

    Sorry Paul, you are too late the direct application of metric has been going on for a while, with the exception of Spain, Portugal and a few South American countries it has been applied for a while. That is why we have, cold,mechanical, heartless soccer
    90% of the time.... What will happen when Leo is done.

  11. Ric Fonseca, August 10, 2015 at 4:15 p.m.

    As early as the "late" 70's, a coach in my region (LA/San Fernando Valley) was starting to deal with our sport and engaging some sort of computer programming that was short of obfuscating the playing and participation in futbol-soccer. As some - and perhaps many - of you know "back in the day..." we relied on recognizing a players' ability on and off the ball, runs into space, defending, attacking in the respective third of the field, etc. Now, as we've extremely rapidly progressed in the age of uber-fast computerization, programming, etc., I find it somewhat mindless for coaches at any level, to engage and use soccer-related metrics, and as is usual the case, IMO, it is taking away a coach's ability to adjust to the game accordingly, and position the players that can respond. Interestingly, as far as "sports metrics" is concerned, today's LA Times sports section has an article about a "new" computer program that can much more accurately determine whether a pitched ball is in the strike zone or not, drawing skepticism of some players, while favored by others, including the plate umpire. But know what, pilgrims, as PG says, until FIFA and IB wakes up from its doldrums, it will continue to be PLAY ON!!!

  12. aaron dutch, August 10, 2015 at 5:19 p.m.

    Why reinvent the wheel, The Dutch for 40 years & Germans for 30 years and English for 20 years and everybody else in Europe for the last 10 have successfully developed tactical/technical metrics BASED on the system of play not just collecting metrics. In the US we always think we are doing something new and special instead of looking at best practices which have a strong research base. They are 100's of real academies & 1000's of people who have received Masters & PhD in the study of football and applied research. Just google the papers or send me an email and i can give you a dump of hundreds of papers.

  13. Allan Lindh, August 10, 2015 at 7:47 p.m.

    Lou is right. Until ref's stop the goons from brutalizing the skill players, "the ugly game" will continue, get uglier. And the Gunners are sometimes pretty to watch, altho didn't work out so well on Sunday.

  14. Mark Konty, August 10, 2015 at 11:01 p.m.

    This is the same nonsense spouted by baseball fans, then American football fans, then basketball fans, and so on. Every fan believes his sport is a pinnacle of human creativity, comprised of intangible elements that make it, and it alone, the greets sport ever. Stuff and nonsense. And the notion that use of such information makes the game ugly, more "mechanical, fails to take into account that unattractive soccer was played, and played world wide, long before the term metrics meant anything more than a European system of measurement. I'll buy the notion that computers will never replace human managers, but it doesn't follow from this that information technology cannot produce better soccer.

  15. R2 Dad, August 10, 2015 at 11:21 p.m.

    These systems provide data, but not useful information. MLS teams want an objective filter they can mindlessly apply to players to determine if they play well, their positioning is correct, etc. Why? Because they don't really know how world-class players play, think, react. (Ironically, MLS coaches and scouts know what real players "look like" and they're all 6'+). We can see excellent dribbling skills, but that doesn't tell us the Moneyball answers: is this an undervalued quality player? However, there is value in measuring performance, because what you can't measure you can't improve. Aaron, I'm still waiting for some pHd to qualify Vision or Movement Off The Ball--can you find that paper?

  16. Kent James, August 11, 2015 at 8:33 a.m.

    I think many commenters have set up a straw man--coaching decisions determined by computer generated stats--and then destroyed it. I doubt even the creators of the system PG described would suggest their program picks players and determines tactics. So the question such critics need to answer is "are there no metrics that would help a team play better?" I think metrics, kept in context, can help a coach see things they were unaware of previously, providing more information on which to base their coaching decisions (presumably allowing them to make more informed, better, decisions). Some people have suggested that possession is a meaningless statistic; yes, possession for possession's sake is meaningless (no goals are awarded for possession), but it is the beginning. If you want your team to possess the ball, so your team controls the tempo of the game, even the existing stat can tell you whether or not you were successful. But a more nuanced meaning (possession in the final third, e.g.) can further describe what your team is doing. And one could envision some very useful metrics (penetrating passes vs. square or back passes) that might even allow you to tweak your team's performance (we score more goals when we attempt more penetrating passes, even though we lose a higher percentage of penetrating passes than square passes, therefore players should be encouraged to try more risky penetrating passes). So yes, statistics used poorly can lead to dull soccer. And yes, early attempts to quantify the game will likely be laughable, but the prospect of seeing something you didn't see before makes it worth the effort to try.

  17. Soccer Madness, August 11, 2015 at 9:18 a.m.

    Kent, at the moment metrics will do more harm to USA soccer than good. Thats 100%. Overall we are niether knowledgable nor have the common sense to make a metrics system useful. Look at the youth game and you will see this. Many top tier Youth coaches still cant figure out posession that doesnt result in goal scoring opportunites is useless and boring and anti development. Same coaches force their teams to play out of the back even when its not an option. Makes the team look like they have no common sense or smart enough to know when to play it long. Not only boring, predictable but dumb as well. Same coaches are anti creativity and dont allow their best to dribble. Lets fix that first then we can look into these metrics if needed. My guess is its not really needed. Common sense can take you a long way.

  18. Kent James replied, August 12, 2015 at 12:43 a.m.

    SM, I don't think you can blame the very real problems you cited on metrics (though as I said (and you fear), it is certainly possible to use metrics to create boring soccer). But the systems PG was describing would only be available at the top levels (due to their high cost), so I don't think it would be a concern for the youth. I'm really curious to see what sorts of things they measure, and how they might contribute to understanding the game, which is so superficially simple (kick the ball in the goal) but is clearly so complex in reality. So yes, they could be misused, but I'd like to see them tried before I dismissed them.

  19. Soccer Madness, August 11, 2015 at 9:34 a.m.

    In USA metrics system will result in more boring play. Most coaches already play to not lose instead of to win, no matter what players they have. I am tired of seeing very skilled players tied up or benched because of coaching agenda. DA Coaches, especially, should be there to enhance top player's skills and strengths. They are not there to force the same style of play to everyone equally. Contrary to what many believe, I know for a fact that we have very skilled players but our overall focus is not to develop them into tomorrow's stars. If I am in charge of metrics I bury it on the opposite side of the world where we cant even smell it.

  20. Gus Keri, August 11, 2015 at 9:45 a.m.

    metrics is here to stay, so, get used to it. But it's naïve to think that managers and coaches will rely solely on it. This is soccer, not athletics. Coaches and managers will still look into the players' technical ability and tactical awareness before looking into his physical conditioning. Metrics is a helping tools to add to the armamentarium that soccer management relies on. If any of Freddi Adu's managers used metrics system on him, none of them would have signed him. no doubt he has the technical ability but his physical conditioning is a big problem and Metrics would have established that.

  21. Soccer Madness, August 11, 2015 at 9:50 a.m.

    Gus, Adu is your argument?? If Brazilian or Argentinian coaches had overseen him they would have told everyone he was not even close to Pele at any point and would have instead turned him into a good USA National Team Player at least. No need for metrics for them to evaluate and establish Adu's case. The bigger issue would always be who we deem as qualified personel. You just gave a perfect example as to why we need to stay far away from these type of systems until we get everything else close to right.

  22. Lou vulovich, August 11, 2015 at 9:53 a.m.

    Instead of focusing on the different development stages of players from 14-16 16-19 and 19-22 and finding out what is the best and most effective ,technical, physical, and physiological approach to developing maximum potential of each age and player,individually they want to have computers do the work for them.
    As for professional teams it is a group of coaches not one and they should be well informed as to what each players strengths and weaknesses are.
    All coaches know that at high level the difference between an average players and good players is between the ears, and no computer or statistics can help that.
    There are enough statistics in sports and most of the time they lie. That is what WALL STREET IS

  23. Soccer Madness, August 11, 2015 at 10:31 a.m.

    Lou, right with you. Instead of admitting that we are just picking the wrong coaches at all levels we make the problem even beigger with this nonsense. Thx to Gus, we have a prefect example in Adu, who always had potential but was just brought up the wrong way by people who did not know any better that hyped him up thinking he was next Pele instead of developing him into an effective pro. Had Adu recieved proper evaluation by knowledgable coaches and therefore proper development and realistic expectations, I strongly believe he would be on Klins's National Team and we would be praising him somewhat instead of irresponsably bashing him every chance we get. ANd by we I am speaking in general terms and not including myself.

  24. Scot Sutherland, August 11, 2015 at 5:56 p.m.

    As a learning scientist who regularly uses real-time analytics along with qualitative methods, I don't see metrics replacing humans. In fact research shows that technology actually increases the need for human labor.

    Analytics tools can bring fresh insights but only when humans apply them in context. Only coaches can use the results of analytics to build a team. I agree with Gardner that the focus of miCoach advertising was individual player performance. In soccer individual performance only matters if it increases team performance.

    Analytics can optimize existing tactics and structures, but it does not predict what might happen due to a tactical change or instructions that change the role of a player.

    Research also shows that unless analytics information is managed carefully to ensure that visualizations are tightly tied to the context in which they are applied, they can actually hamper successful decision making.

    Most analytics tools are very new. We are just beginning to understand how they interact with existing decision making processes. The Oakland A's (Money Ball) are a good example of how analytics can be applied successfully in context. However, unlike soccer, baseball is for the most part a collection of individual performances.

    Successful soccer depends upon a team performing as a cohesive whole.

  25. Kent James replied, August 12, 2015 at 12:47 a.m.

    Well put. Things are not quite as black and white as PG makes them.

  26. Bob Ashpole replied, August 17, 2015 at 12:39 p.m.

    Scot, in my experience dealing with many business management problems over the years I find the capable managers using information to further business performance greatly outnumbered by managers who use imaginary "metrics" contrived to "prove" whatever point they goal they have, usually the promotion of their own career rather than to manage the business. To the unknowledgeable, the garbage metrics on the power point slides look genuine. Then you have well-intentioned, but unknowledgeable, people who think they can make valid conclusions about causal relationships merely by examining the magical "metrics." The IT companies market to them all.

  27. aaron dutch, August 11, 2015 at 10:06 p.m.

    Europe they are committed to the game and study it. we just wing it and think we are making knew approaches. Here are a few papers

    http://cogpsy.let.hokudai.ac.jp/~ogawa-lab/lec/semi2014a/67510670.pdf

    http://etheses.bham.ac.uk/3828/1/Lamche12MPhil.pdf

    Until the US federation has the leadership to really focus on building an environment that is development based not “pay to play” so its about the leadership to build a National Development Plan like Belgium in 2001 with Michel Sablon.

    http://www.theguardian.com/football/blog/2014/jun/06/belgium-blueprint-gave-birth-golden-generation-world-cup-

    It has to be fun, inclusive, technical/tactical/football iq based and organized. The whole federation leadership should be canned as we have not done anything really in 10+ years if we are honest. And given the resources which are $$billions spent at all levels of football in the US we should be far better.

  28. Raymond Weigand replied, August 12, 2015 at 11:54 a.m.

    Thank you.

  29. aaron dutch, August 11, 2015 at 10:09 p.m.

    sorry the text editor wrapped around

    This was for R2DAD

    Here are a few phd papers he asked for

    http://cogpsy.let.hokudai.ac.jp/~ogawa-lab/lec/semi2014a/67510670.pdf

    http://etheses.bham.ac.uk/3828/1/Lamche12MPhil.pdf

    and the link to the belgium blueprint

    http://www.theguardian.com/football/blog/2014/jun/06/belgium-blueprint-gave-birth-golden-generation-world-cup-

  30. beautiful game, August 12, 2015 at 10:29 a.m.

    Bottom line is that athleticism and a low soccer IQ brings nothing to the table. Take a peak at Fiorentina and Sevilla, i.e., in their recent matches; big athletic players with a hi-soccer IQ who provide speed, physicality, and above all EFFICACY. No one seems to mention that in the MLS too many players play into pressure and can't handle pressure; and add to that not understanding the simplicity of play and playing within ones own limits.

  31. Francois Gazzano, August 12, 2015 at 3:41 p.m.

    What do you think of this? Players GPS metrics streamed live to fans during games. http://www.afr.com/business/sport/game-data-to-stream-live-for-melbourne-city-football-fans-20150723-giit15?stb=twt

  32. aaron dutch, August 12, 2015 at 3:48 p.m.

    Our system builds athleticism first then tactical (ex- american football coaches who know nothing about world football) then technical then Football IQ we almost never get to those if we do they are when the player is U16 to U18 and its too late then they go to college to begin the death of their football craft.

    Its our development system which is all about parents thinking basketball as soccer with american football yelling which we have all heard as coaches with parents saying "long ball, send it".

  33. aaron dutch, August 12, 2015 at 4:31 p.m.

    R2DAD

    here are 2 more good papers on movement/positional/ systems of play

    paper 1--

    https://www.repository.utl.pt/bitstream/10400.5/6844/1/Tese%20vFinal.pdf

    paper 2--

    http://eprints.port.ac.uk/17194/1/Put_your_money_where_your_mouth_is.pdf

Next story loading loading..

Discover Our Publications