Saturday, October 29, 2011

A Beautiful But Flawed Game



The English football match that everyone was paying attention to last weekend, likely the most highly anticipated contest of the season thus far, was the Manchester derby. And, yes, it was glorious watching the oil-fueled nouveau riche from Etihad punish and embarrass the always and ever loathsome paragons of English football, the flagbearers of success and excess, tradition, and mythologizing aka the Yankees of the old country, but even more annoying. (FN1) Nevertheless, what I want to discuss is the Queens Park Rangers v. Chelsea affair of later that day. It was a match that sort of perfectly illustrates what’s wrong with the sport, or at least what’s wrong with a certain aspect of the sport.

FN1 - I walked by the neighborhood Manchester City pub – I will concede that one of the nicer things about living in New York City is that there’s a designated Man City bar (and basically any other type of bar) just a few blocks away from me that is packed to the gills on weekend mornings during City games – and it was pretty great to see and hear pissed Mancunians singing loudly in unison and off-key, spilling out into the street as Dzeko put up a six spot on United.

In the 8th minute, the referee awarded a penalty to QPR from a David Luiz foul in the box against Heidar Helguson. While the call was fairly marginal – the announcers said it could have gone either way and that they’ve seen more egregious fouls gone uncalled this season; I think that's a pretty fair reading – the problem lies not specifically with a failure of officiating but rather the consequences of such a potential failure. Penalty kicks during the course of play (as opposed to settling a tie) seem to be converted at around 80% (though I can’t seem to find conclusive numbers for this in the EPL). That means the referee’s decision on a marginal call essentially awarded QPR .8 goals. The average number of goals scored by each team in the EPL last season was 1.4 (the highest number in league history). That means that the single call on Luiz represented more than half the scoring for an average team in an average game. That makes a single officiating decision a disproportionately high impact event, which is, in this case, exacerbated by the marginalness of the call.


I'll equate it to the sport that is structurally most similar to soccer, ice hockey, where there are a number of minor penalties called in each game, and where some are clearly marginal or even "phantom," meaning that the referee did'nt really see what happened but called something anyway based on what he thinks might have happened. But the effect of such calls are fairly limited in that NHL teams only convert around 18% of their power plays (this includes 5on3 and 4on3 advantages, so the number is a bit lower for 5on4's, but for our purposes, the difference is not important). This means that awarding a penalty results in the advantaged team receiving .18 goals (it's actually lower because shortanded teams score some of the time, but it's fine for our purposes). Last year in the NHL teams scored an average of 2.8 goals per game. To have the same effect on a team's scoring output as a penalty in the EPL, an NHL team would need to be awarded 1.6 goals for a penalty. To score 1.6 goals, an NHL team would need 9 power plays. So, to have roughly the same effect on the game as when a penalty is awarded in the EPL, a penalty in the NHL would have to yield an 18 minute major power play. Clearly, such a crime and punishment system would be absurd (see, e.g., drug possession laws in the U.S.). Penalties in soccer have a disproportionately strong effect on the outcome of the game. If a mistake is made and the referee makes the incorrect call, the game is essentially ruined in many cases.


Not only did the Luiz penalty have such an overly significant effect on the QPR-Chelsea match - Helguson buried the penalty of course - but Chelsea was the victim of soccer's other extra-punitive rule, the ejection of a player following a red card. Now, I don't have a problem with the general rule itself. Chelsea's Didier Drogba actually received a red card late in the first half, Chelsea's second straight red of the match, and I think everyone would agree that the call was justified. He went in wildly with two feet for a late challenge. The punishment is harsh so as to dissuade dangerous tackles that could lead to serious injury. Consequently, you rarely see such challenges, and I'm not sure what Drogba was doing or thinking there. But Chelsea's first red card of the half (what a fucking miserable first half they had), was a very different kind of play. Jose Bosingwa and Shaun Wright-Philips raced for a loose ball deep in Chelsea's zone. The players seemed to be relatively even, arriving at the ball just outside of the box when Bosingwa appears to have wrapped his arm around Wight-Phillips, possibly tugged on his jersey, and both players went down. The alleged tug was imperceptible in real time and very subtle on the replay. It certainly wasn't obvious or gratuitous. The referee found that a goal scoring chance was denied and handed out a straight red. Chelsea would be down to 10 men for the remainder of the match, that is, until they subsequently went down to 9 men. My problem lies not with the questionable call, I certainly disagreed with it (it looked to me that Bosingwa had the angle on Wright-Phillips and no great scoring chance was going to come), but it's a judgment and the ref has to make a decision one way or other. The problem is that the the consequences of a difficult 50/50 judgment call, that the player is sent off and the team is down to 10 men, are massive: shooting rates go way down, shots allowed go way up, and the impact when the visitors lose a man for half the game is approximately 0.75 goals - roughly the same as giving up a penalty kick. That is an incredibly harsh result for a fairly borderline call that could have gone either way. Not every ref awards a red card there, and the effects of such a decision are just too significant.


Perhaps the great irony of all this is that despite being down 11 men to 9, Chelsea dominated the second half and had by far the better of the chances. In what was so infuriating to me as a casual disinterested observer that I cannot image what Andre Villas-Boas was going through, there were two instances where QPR players engaged in shirt pulling in the box to deny or hinder Chelsea scoring chances (one on Frank Lampard and another on Luiz) that went uncalled. (FN2) Neither infraction was particularly egregious or obvious, but they were both at least just as bad as the Luiz bump or the Bosingwa tug. The non calls were particularly disappointing because of Chelsea's fight and battle showed in defending and attacking relentlessly with just 9 men. A harsh result for Chelsea, needlessly so only because of draconian and irrational penalties associated with certain rule violations.

FN2 - Luiz, by the way, is a stud. He's one of the few soccer players that I can confidently say would have made a great hockey player. He had a physically dominating and exhausting second half, going up an down with the pitch with controlled abandon. Beauty player.

Friday, October 7, 2011

Old Wine, New Bottles

There's been a recent spate of articles detailing the "Moneypuck" revolution in ice hockey. I've used advanced statistics before, e.g. in defending the misunderstood and oppressed, and find them useful in distilling large amounts of varied data into digestible pieces; they aid in the understanding of certain hockey events. We know more about the game today than we did 5 years ago, and every NHL team should be utilizing advanced statistics in making personnel decisions. That said, I'm with Tom Benjamin: "Hockey statistics will never do what we want them to do, which is to [comprehensively] evaluate individual hockey players." The reasons for this, I think, are complicated and not well understood: the team context pollutes individual results, the game is not comprised of binary or even recordable events, and "[h]ockey statistics do not add up to goals." In any case, NHL teams are starting to look at data more critically and are finally beginning to eschew traditional mythmaking as an explanation for results, and that's a good thing.

I imagine, however, that within this movement both outside and inside organizations there are more than a few snake oil salesmen (cough*puckstopshere*cough) who don't really know what they're doing and are going to try to hoodwink and bamboozle some teams into buying what they're peddling. I don't know if John Weisbrod is such a fellow, but he apparently went to Harvard and reads Thoreau, meaning he's like super smart and well edumacated, and he's involved in the Calgary Flames' efforts to use analytical software programs to evaluate talent. As a scout in the Boston Bruins organization, he "broke it down and redefined what we wanted a Boston Bruin to be," which sounds very, um, new and, er, scientific. But Weisbrod's money quote, after a desultory discussion of the sophisticated ways in which the Flames are using these tools, is regarding his time as GM of the NBA Orlando Magic:

[McGrady] was one of the most talented players in the league, very popular, but I came to the conclusion he didn't have the internal fortitude to win a championship.

The heart/60 metric is one of the great successes of the advanced stat discipline, no doubt.