The Derrick Rose MVP debate brings to a head an issue that has been simmering in the sports community for the better part of a decade. Sports have undergone a statistical revolution in recent years. Words like VORP (value over replacement player), WAR (wins above replacement) and PER (Player Efficiency Rating) have become a casual part of conversation amongst sports fans. Each one of these stats aids in providing a greater understanding of the game; Sabermetrics add a great deal of context to any statistical debate and their importance should not be understated. Unfortunately, the proponents of advanced statistical metrics thumb their noses at the traditional stats thereby undervaluing stats like wins and RBIs. Two camps have formed on the statistical landscape. In one camp, you have the traditionalists. They see advanced statistical metrics as confusing simple games. Their thinking is simple: as a pitcher, the ultimate goal is to win the game. Hence, wins are an important stat. The ultimate goal of the offense is to score runs. Hence, RBIs are an important stat. They are easy to calculate and easy to understand. Advanced metrics complicate a simple game. In the second camp, you have the stat heads. The people that come up with these advanced metrics aim to give a better sense of what is actually happening on the field. They are correct that arbitrary factors outside of a pitcher’s control lead to wins (bullpen blows the game, offense is shutout, defense commits errors), just like errors are not indicative of a shortstop’s defensive value (Jeter committed fewer errors last year than Alexi Ramirez, but Jeter’s limited range - as calculated by range factor - shows that he reached far fewer balls than Ramirez and was a weaker defensive shortstop). But the proponents of Sabermetrics tend to look down at the traditionalists as simple minded people unable to understand a complex game. I am here today to say that there is value in both approaches and each leads to a greater understanding and enjoyment of the game. As baseball has the most developed advanced stats, in addition to being the sport that holds statistics most dear, this analysis will be limited to that topic.
I would argue that traditional statistics offer a fairly complete picture of an athlete’s career over time in the context of pitching and hitting, but fail to adequately address fielding. The advanced metrics are correct in that a lot of variables determine stats like wins and RBIs. The problem with wins is that a 10-6 win counts the same for a pitcher as does a 1-0 win and for more than a 1-0 loss. Bullpen, run support and defense are external factors beyond the pitcher’s control. Stats like WHIP (walks and hits per inning pitched) more accurately measure a pitcher’s effectiveness because it takes the majority of those variables out of play. Felix Hernandez’s Cy Young season last year is example A. Hernandez had a record of 13-12, but with an ERA of 2.27 and a WHIP of 1.057, he was far and away the best pitcher in the American League, out pitching 21 game winner CC Sabbathia and 19 game winner David Price. Wins fail to account for the fact that the Mariners had a truly putrid offense and only won 61 games while the Rays and the Yankees won 96 and 95 respectively. Now, over the course of his career, many of those variables will even out. Hernandez should eventually have seasons with a better offense providing him run support and a better bullpen to maintain leads. Thus, if Hernandez pitches like he did last year for the next several years, his wins should, over time, correlate to his value as a pitcher. As a metric to determine how well a pitcher pitched in a single season, however, a stat like WHIP is more indicative of a pitcher's performance.
The same holds true with RBIs, a stat that is clearly influenced by opportunity. Adam Dunn, toiling away with the Nationals, had far fewer RBI opportunities than he will for the White Sox this year. If Dunn knocks in more runs this year, does that mean that he is a better player? Not necessarily. And that is where stats like OPS (On Base Plus Slugging) come into play.
Where advanced metrics really help, however, is in analyzing fielding. Errors provide such a limited view of defensive prowess because they start from the premise that each player is exactly the same defensively. As noted above, Derek Jeter can no longer make a play deep in the hole like he used to. Alexi Ramirez, on the other hand, gets to far more ground balls than Jeter. Does he make more errors? Yes. But errors fail to quantify the number of additional hits conceded by the Yankees because of Jeter’s lack of range and the number of additional hits Ramirez saves. That is not to undermine errors as a stat. Errors are relevant (anytime you give a team an extra out in baseball, it is relevant), but stats like range factor put errors in the proper context, as one of several metrics to quantify how many additional outs you give the opposing team defensively over the life of a season.
The ultimate problem with advanced metrics, however, are twofold. The first is that, while advanced metrics provide a greater picture of what is happening on the field, they are much more difficult to calculate. When I was a kid, falling in love with baseball, I loved stats. I kept track of my favorite players’ batting averages, I scoured box scores to see how Frank Thomas’ rivals were doing and what Frank needed to do to win the homerun, RBI, or batting average titles (I have a vivid memory of Frank Thomas, down a single RBI to Albert Belle in 1993 on the final day of the season, being intentionally walked with a runner on first against Cleveland - and swinging at two of the pitches - and how livid I was that Cleveland would resort to that tomfoolery to get Albert Belle the RBI title). They were easy stats for me to track and follow. Ask me to calculate VORP and I look at you blankly. VORP is an extremely valuable stat, but 10 year old me would not have been able to follow VORP with the same zeal that I followed batting averages and ERAs (note, I probably would have loved VORP, I just would not have been able to calculate it).
The second is that the real die hard stat heads seek to remove the human element from baseball entirely. They add all these metrics that really do provide a more accurate tool to compare players, but they are not all encompassing the way many proponents of advanced statistical metrics would have you believe. Two examples of this: (1) Tadahito Iguchi played second base for the White Sox in 2005. Iguchi was the consummate professional that season. He gave up outs to move runners over and was, arguably the most valuable offensive player on the Sox that year. Now, if you look at his OPS, you are going to be underwhelmed. To my knowledge, there is no stat that takes into account sacrifice for the team (Iguchi did lead the league in productive outs that year, but you are digging pretty deep to reach that stat). (2) Javy Vazquez is the type of pitcher who defies stats. If you need someone to throw a two hitter against the Royals in July, he’s your guy. But to my knowledge, there is not stat that can show how terribly Vazquez performs in big games. The point in all this is that, even the most advanced metrics fail to account for the space between an athlete’s ears. Iguchi was willing to give up his numbers to benefit the team. Vazquez pitches terribly when the chips are on the line. Stats do not tell me that.
The ultimate point here is that there is a place in baseball for both traditional stats and advanced stats. But for some reason, both sides have a tendency to shout each other down. The goal of statistics is to make the game easier to understand and more fun to follow. I love a good argument about the relative merits of Player A vs. Player B. I will call on advanced stats when it matters (see discussing how good Adam Dunn is offensively) and I will fall back on stats like batting average when necessary to make a point (see my famous argument with one Tom Smith over the merits of the Mets giving Carlos Beltran the contract they gave him - a truly epic argument over beers in a Breckinridge, Colorado bar our Senior year of undergrad). They both tell a story, and they both are important when it comes to enhancing interest in the game. If only the two sides could see eye to eye.
No comments:
Post a Comment