Scorecasting: The Hidden Influences Behind How Sports Are Played and Games Are Won
Page 15
Then, in the bottom of the tenth inning, Jack’s idyllic afternoon was ruined. The Brewers had summoned Mark DiFelice, a right-handed pitcher who had recently won his first Major League game at age 32. When the Cubs loaded the bases, DiFelice faced Chicago’s third baseman, Jake Fox, a utility man who’d ricocheted between the majors and the minors. With a full count, two outs, and the decibel level soaring at Wrigley Field, DiFelice threw four consecutive pitches that Fox fouled off. On the next pitch of the at-bat, DiFelice reared back and fired a cutter that froze Fox and shot past him. After an awkward pause, home plate umpire Bill Welke popped up from his crouch and … stood idly. Ball four. The winning run had been walked home: Cubs 2, Brewers 1.
The crowd goes wild. Jack Moore of Trempealeau, Wisconsin, goes ballistic. “For five minutes, I just screamed words you can’t print,” he says. “Anyone who knows baseball knew that was a strike.” For years, fans in Jack’s position would bitch and moan and dispute balls and strikes until last call. But this was 2009, and Jack wasn’t interested in an argument; he was interested in a straight, objective answer. He fired up his Internet browser, logged on to MLB.com, and clicked on Pitch f/x. Sure enough, DiFelice’s pitch was gut-high and clearly within the upper-inside part of the strike zone. Minutes after the game had ended, right there in his parents’ home in small-town Wisconsin, a 19-year-old was able to confirm his suspicions. The ump had blown the call, permitting the home team to win.
What sports fan doesn’t harbor a belief that the officials are making bad calls against his or her team? It’s a home crowd that voices this displeasure the loudest. The criticism ranges from passably clever (“Ref, if you had one more eye, you’d be a Cyclops!”) to the crass (“Ref, you might as well get on your knees because you’re blowing this game!”) to the troglodytic (“You suck!”). Dissatisfaction is voiced individually and also collectively, often in a stereo chant of “Bullshit! Bullshit!” In Europe—quaint, civilized Europe—there are even various soccer websites that enable fans to download antireferee chants as ringtones.
What we’ve found is that officials are biased, confirming years of fans’ conspiracy theories. But they’re biased not against the louts screaming unprintable epithets at them. They’re biased for them, and the bigger the crowd, the worse the bias. In fact, “officials’ bias” is the most significant contributor to home field advantage. “Home cooking,” as it’s called, is very much on the menu at sporting events.
A statement like that had better have some backing, and we’re prepared to provide it. Warning: An assault of numbers awaits. But stick with us and we’ll walk you through it. We think the payoff is worth it.
Let’s start by determining how to measure ref bias. You could examine the accuracy of calls made by the officials and whether that accuracy differs for calls favoring the home team versus the away team. But doing that is a challenge because it requires a great deal of subjectivity as well as a deep knowledge of the circumstances of the game. Was it really a foul? Was it really pass interference? What else was happening during the game at that time? In light of the speed of the game and the reactions of players within the game, it is nearly impossible to control for all the potential factors that could lead to differing calls for the home and away teams.
Suppose we find that more fouls are called against road teams than against home teams—which, by the way, is often the case. Does this indicate a referee bias in favor of the home team? Maybe, but not necessarily. What if teams play more aggressively on the road? After all, road teams know that statistically, they are already more likely to lose. Or what if the road team, exhausted from those back-to-back games, lacks the energy for proper defense and clutches and grabs instead? They might be inclined to commit more fouls regardless of any referee bias, and so it’s difficult to identify the causal factor. Are referees causing more road team fouls because of bias against the road team? Or are players causing referees to call more fouls because of more sloppy or aggressive play? Or is there a third factor causing both?
We looked for a component of the game the refs control that isn’t influenced or affected by players. We found it in a sport for which we have not had much success in explaining its sizable home advantage—soccer. It also turns out that had it not been for a diligent grandmother from Spain religiously watching and recording years’ worth of Sunday evening matches, we might not have discovered this bias at all.
In soccer, the referee has discretion over the addition of extra time, referred to as “injury time,” at the end of the game to make up for lost time resulting from unusual stoppages of play for injuries, penalties, substitutions, and the like. This extra time is rationed at the discretion of the head referee and is not recorded or monitored anywhere else in the stadium.
As best he can, the referee is supposed to determine the accumulated time from unusual stoppages—itself a subjective measure—and add that time at the end of regulation. So does the referee’s discretion favor the home team? If so, he would lengthen this time when the home team is behind at the end of the game and reduce it when the home team is ahead, extending or shortening the game to increase the home team’s chances of winning.
Using handwritten notes that his elderly mother had gathered logging matches she’d watched from her living room in Spain, Natxo Palacios-Huerta, a London School of Economics professor, joined with two colleagues from the University of Chicago, Luis Garicano and Canice Prendergast—all soccer fanatics—to study the officials’ conduct during games. The researchers were, quite justifiably, struck by what they found. Examining 750 matches from Spain’s premier league, La Liga, they determined that in close matches with the home team ahead, the referees ritually shortened the game by reducing the extra time significantly. In close games in which the home team was behind, the referees lengthened the game with extra injury time. If the home team was ahead by a goal at the end of regulation, the average injury time given was barely two minutes, but if the home team was behind by a goal, the average injury time awarded was four minutes—twice as much time. Sure enough, when the score was tied and it wasn’t clear whether to increase or decrease the time for the home team, the average injury time was right around three minutes.
What happened when the home team was significantly ahead or behind? In games that were not close, there was no bias at all. The extra time added was roughly the same whether the home team was ahead by two goals or more or behind by two goals or more. This makes sense. A referee has to balance the benefit of any favoritism he might apply with the costs of favoritism—harm to his reputation, media scrutiny, and potential reprimands. Adding additional injury time when the score was so lopsided was unlikely to change the outcome and therefore accrue much benefit, so why do it and risk the potential cost?
The study also looked at what happened when, in 1998, the league altered its point structure from awarding teams two points in the standings for a win (and one for a draw and zero for a loss) to three points for a win. That change meant that a win was suddenly worth a lot more than it had been before and the difference between winning and tying doubled. What did this do to the referee injury time bias? It increased it significantly. In particular, preserving a win against the possibility of a tie now meant a lot more to the home team, and so the referees adjusted the extra time accordingly to reflect those greater benefits.
This wasn’t unique to Spain. Researchers began looking for the same referee biases in other leagues—not hard given the global popularity of soccer. They found that the exact same injury time bias in favor of the home team exists in the English Premier League, the Italian Serie A league, the German Bundesliga, the Scottish league, and even MLS in the United States.
If referees are willing to alter the injury time in favor of the home team, what else might they be doing to help ensure that the home crowd leaves happy? We found that referees also award more penalties in favor of the home team. Disputed penalty shots and goals tend disproportionally to go the home team’s way as well. Looking at m
ore than 15,000 European soccer matches in the English Premier League, Spanish La Liga, and Italian Serie A, we found that home teams receive many fewer red and yellow cards even after controlling for the number of penalties or fouls on both teams. The dispensing of red and yellow cards has a large impact on a game’s outcome. A red card, which sends the offending player off the field, reduces a team’s chances of winning by more than 7 percent. A yellow card, which precedes a red card as a stern warning for a foul and may therefore cause its recipient to play more cautiously, reduces the chances of winning by more than 2 percent. These are large effects. When a single yellow card, followed by a red card, is given to a visiting player, it means the home team’s chance of winning, absent any other effects, jumps to 59 percent. Add the injury time, fouls, free kicks … and it suddenly isn’t so surprising that the home team in soccer wins nearly 63 percent of its games.
But could this be limited to the idiosyncratic world of European soccer? Surely, American sports wouldn’t be subject to the same referee bias … would they?
Remember how, despite a significant home team advantage, athletes do not hit or pitch, shoot free throws, slap goals, or pass the football appreciably better at home than they do on the road? This prompts the question: What do home teams do better that allows them to achieve a higher winning percentage?
In baseball, it turns out that the most significant difference between home and away teams is that home teams strike out less and walk more—a lot more—per plate appearance than do away teams. This could be for lots of reasons. One interpretation: Home team batters see the ball better or away team pitchers exhibit less control. But this contradicts our earlier results for batters and pitchers—in controlled, isolated environments, they hit and pitch the same at home as they do on the road. And as we’ve seen, road players in MLB aren’t performing worse because they’re exhausted from the travel.
Balls and strikes are the domain of the head umpire. Could the umpire be biased toward the home team? This would explain the differences in strikeouts and walks despite the lack of any difference in hitting and pitching.
But strikeouts and walks are not the right statistic to measure, because many strikes occur when a batter swings and misses or fouls off a ball. In such cases, there is no umpire discretion. A better metric to look at is called balls and strikes.* In other words, look only at pitches that do not involve swinging by the batter. It turns out that home batters receive far fewer called strikes per called pitch than away batters do.
It’s even more apparent when we look at called strikes and balls at different points in the game. Certain situations have a much bigger impact on the game’s outcome than others. Fortunately for us, Sabermetrics, an analysis of baseball through objective evidence, provides another useful tool to gauge the importance of a particular situation. A stats wizard, Tom Tango, devised a metric called the Leverage Index to measure the relative importance of any game situation. The idea is to take every game situation and consider every possible scenario that could occur in that situation, the likelihood of each scenario playing out from that point, and what effect each of those scenarios would have on the ultimate outcome of the game. Add up all these possibilities, their likelihood of occurring, and their potential impact on the game and you have a measure of how crucial the current situation is. A Leverage Index of 1 is the average situation; an index of 2 means the situation is twice as crucial. Here are two extreme examples: Down by four runs with two outs and nobody on base in the bottom of the ninth, where the game isn’t in much doubt, translates into a Leverage Index of 0.1—the situation is one-tenth as crucial as the average situation. Down by one run in the bottom of the ninth with two outs and the bases loaded, where the game is on the line, gives a Leverage Index of 10.9. It is almost 11 times more crucial than the average situation.
Using the Leverage Index to examine called strike and ball counts in different situations, we found, just as with the soccer referees, that in low-leverage situations, when the game is not in much doubt, the home team advantage in receiving fewer called strikes and more balls goes away. But as the following chart shows, the called-strike advantage for home teams grows considerably as the game situation gets more and more important. In noncrucial and average situations, the home team receives about the same strike calls as, or even a few more strike calls than, the away team per called pitch, but that changes dramatically when the game is on the line. In crucial situations, the home team receives far fewer called strikes per called pitch than does the away team.
DIFFERENCE IN PERCENTAGE OF CALLED PITCHES THAT ARE CALLED STRIKES ON HOME VS. AWAY BATTERS
This makes sense. If the umpire is going to show favoritism to the home team, he or she will do it when it is most valuable—when the outcome of the game is affected the most. You might even contend that in noncrucial situations the umpire may be biased against the home team to maintain an overall appearance of fairness.
Think back to that Jake Fox pitch in the Cubs-Brewers game, on a 3–2 count with the bases loaded and a tie game on the line in the bottom of the tenth inning. It was an astronomically high-leverage situation. Knowing the statistics, you would have bet the house that the pitch wouldn’t have been called a strike. And it wasn’t.
Let’s look at other calls that fall under the domain of the umpires, in particular, close calls that typically elicit a home crowd reaction. Two good examples would be stolen bases and double plays. We found that home teams are more likely to be successful when stealing a base and when turning a double play, yet the distance between the bases is identical in every stadium—stolen base success can’t be driven by home field idiosyncrasies. In addition, the success rates of home teams in scoring from second base on a single or scoring from third base on an out—typically close plays at the plate—are much higher than they are for their visitors in high-leverage/crucial situations. Yet they are no different or even slightly less successful in noncrucial situations. (Third-base coaches, take note: If it’s a close game and you’re playing at home, windmill your arms and send the runner!)
But the most damning evidence of umpire bias might be a function of a tool that was employed for the specific purpose of eliminating umpire bias. Remember the Pitch f/x system that tracks the characteristics of each pitch, including location? Well, its predecessor—a digital technology called Umpire Information System (UIS) from QuesTec—was installed five years earlier by Major League Baseball for the specific purpose of monitoring the accuracy of umpires. According to Major League Baseball, QuesTec was implemented in six ballparks in the first year; by the time it was discontinued in 2008, 11 ballparks had the technology.* With two cameras positioned at field level and two in the upper deck, QuesTec combined the four images to track where the ball crosses the plate, and it was used by baseball executives to determine how closely an umpire’s perception of the strike zone mirrored reality.
We also used the presence of QuesTec to evaluate umpire accuracy, but in a different way. We asked whether the same umpire behaved differently when he knew the cameras were monitoring him. If the home field advantage in called strikes disappears when the umpires know they’re being watched—while everything else stays constant—it’s pretty clear that official bias underlies it. Imagine you own a coffee shop and put out a jar in which patrons can donate or take loose change. You notice at the end of each day that the jar is empty. You deduce that either some customers are taking advantage by depleting the jar or your employees are stealing the coins. You tell only your employees that you are installing a hidden video camera. If the change jar is full at the end of each day, you’re pretty darn sure it was your employees, not customers, who were to blame.
To test our theory, we first compared all pitches, about 5.5 million of them, from 2002 to 2008 made in stadiums using QuesTec versus those without it. For example, we looked at all called pitches when the Astros visited the Cardinals (at their non-QuesTec stadium) and when the Cardinals visited the Astros (at their QuesTec-equipped stadium).
What did we find? Called strikes and balls went the home team’s way, but only in stadiums without QuesTec, that is, ballparks where umpires were not being monitored. This is consistent with an umpire bias toward the home team causing the strike-ball discrepancy. We also found something surprising. Not only did umpires not favor the home team on strike and ball calls when QuesTec was watching them, they actually gave more strikes and fewer balls to the home team. In short, when umpires knew they were being monitored, home field advantage on balls and strikes didn’t simply vanish; the advantage swung all the way over to the visiting team.
We then looked at the same pitch counts in low-leverage (not crucial) and high-leverage (crucial) points in the game. Again, when a plate appearance is expected to have little effect on the outcome of the game, there is no bias for or against the home team. Umpires call things evenly whether QuesTec is present or not. But when the at-bat can have an impact on the game, we found both biases to be even more extreme. That is, when the game is on the line, home teams in non-QuesTec stadiums get a big strike-ball call advantage and those in QuesTec stadiums get a huge strike-ball call disadvantage.
In practical terms, when the umpire is not being monitored by QuesTec, a home batter in crucial game situations will get a called strike only 32 percent of the time if he doesn’t swing. In the same situation, a batter from a visiting team gets a called strike 39 percent of the time. That’s a big difference. Now consider the same two situations when the umpire is being monitored by QuesTec. Here the home batter gets a called strike 43 percent of the time, and the away batter only 35 percent of the time.