Part two on the debate between traditional statistics and sabermetrics.
Note: While I was typing this up, I had some inspiration on some additional posts that could be included in this series. Therefore, you may notice that I have changed the header at the top to reflect that this series will have an indefinite number of posts. The same has been done for the ERA/FIP Fistfight post. Also, I’m going to keep the whole series available at all times on the right of this page, under the title “The Statistics Debate Series.” I honestly feel that this series is one of the most interesting pieces that I’ve written since I created this blog.
Well I’ve been mulling this over for several weeks now. I posted a while back the first part of this series, which involved a disagreement over whether ERA or FIP was more useful in judging a pitcher’s worth. I used an extreme example of an infinitely large ballpark and a perfect defense versus a completely average ballpark and defense to show that FIP was a better predictor of a pitcher’s future ERA. However, an infinitely large field with a perfect defense doesn’t exist, so it’s arguable that my example only holds true in this single case (a point that many of my math teachers enjoy telling me when I try to write out some proofs), but I’ll save my talk about FIP and ERA for later.
Many people treat the traditional statistics and sabermetrics as being a black and white issue, which frustrates me to no end. It’s why I don’t like getting involved in a political discussion or leaving comments on a blog written by someone like La Velle E. Neal III: Either point A is true, or point B is true. If A is true, then B is clearly false, and vice versa. Oh, and if you don’t agree with someone’s argument, be prepared to be lambasted for having a different opinion. As I mentioned before, I often believe in gray areas when it comes to choosing a side for an argument. Yes, a person normally wears shorts when it’s hot outside and doesn’t when it’s cold. But what if it’s mild? If it’s fall, I’m probably wearing pants because we’re transitioning from hot to cold. If it’s spring, I’m probably wearing shorts because the temperature is warming up. See? Gray areas.
The same is true with these two viewpoints on baseball statistics. The traditionalists view sabermetrics as being complicated and made-up; sabermetricians aren’t as passionate about the sport* and don’t even watch games on TV, and they’re only out to prove the common opinion as being wrong. “Albert Pujols had a 2009 isolated power of .331? What does that mean? I understand that he had 47 home runs, so clearly he’s a great power hitter, but isolated power?” The sabermetricians view the traditional stats as being outdated and misguiding; traditionalists are more likely to display subjective opinions, and they supposedly ignore or are incapable of understanding why a player’s performance has taken an unexpected turn. “Emilio Bonifacio sure did start out hot at the beginning of the season, but his batting average on balls in play is through the roof. Once it normalizes, people will see Bonifacio is only good for a .250 batting average with no power.”
* In Jake Depue’s defense (he was guest writing for Howard), he was merely questioning if “Stat-Heads” were as passionate about baseball as “Dreamers.” He wasn’t accusing anyone of not caring as much about the sport.
I feel that both sides of the argument have made mistakes in judging people on the opposite side. Traditionalists seem to think that sabermetricians sit behind computers all day, never watch the games, and are only interested in the data and proving the casual fan wrong. But I can attest that this is simply not true. Aaron Gleeman, who I certainly would call a sabermetrician, is a big fan of the Twins and definitely baseball in general. The website FanGraphs is very similar, except their group of writers focus on everything that has to do with baseball (and some other topics if you’re Carson Cistulli. He seems to be easing up on the non-baseball talk lately, however). Additionally, these two often report on the happenings of baseball during the season. If anything, I feel that they watch more baseball than the casual fan.
Another issue is that traditionalists certainly seemed threatened by sabermetrics. If it doesn’t make sense, why use it? This does seem like a good point, but I feel that it could be alleviated with some help from the “Stat-Heads” themselves. It’s a bit tougher on a website like FanGraphs because the majority of the readers believe in these stats, so giving constant reminders and explanations of certain stats like BABIP (batting average on balls in play) is de-emphasized. However, I tend to give explanations every now and then because I don’t know which statistics you believe are useful, or because I myself do not know what they are when I first find them (in the case of sOPS+ and tOPS+) and I feel that if I don’t understand why we can use these statistics, then I certainly shouldn’t assume that you do understand them. A thought that I constantly have is from a comment that I read on Nick’s Twins Blog. I’m paraphrasing like I never have before in my life, but this was basically what the comment read:
“Hi Nick, I love reading your blog and what you write about, but I’m an older fan and thus do not understand some of these statistics that you sometimes reference. Could you please explain what this means so I can have a better understand of what you’re talking about? Thanks.”
Basically every time I include something like ISO or UZR, I’m reminded of this comment. Now if only everyone was this polite when they leave a comment!
Additionally, the traditionalists do not like seeing other statistics that seem to be taking over what they previously believed. Batting average is getting replaced by batting average on balls in play, fielding percentage and errors by UZR and Dewan’s +/-, ERA by FIP, etc. What many people don’t understand is that sabermetricians have run statistical tests on many of these to determine that they are indeed more reliable than the older statistics. I mentioned how FIP is more reliable than ERA in proving a pitcher’s future ERA in the first post of this series, and I hinted that I used statistical tests in the Starting Pitching Hospital post to see if each pitcher’s stats were above, at, or below the major league average. These tests shown that many of these sabermetrics correlate better to future traditional statistics than comparing traditional statistics from the past to the future, which is why they have been gaining so much popularity lately.
Roughly three months ago, Ken Rosenthal wrote an article about how sabermetricians should stop trying to force their ideas down everyone’s throats because their statistics were now firmly established in the baseball world. In my opinion, Rosenthal did give some reasonable explanations as to why he believed that Kevin Youkilis deserved the MVP award as much or more than Joe Mauer, but my issue was that he seemed to completely ignore the fact that sabermetrics are not in the mainstream.* Yes, baseball teams are hiring sabermetricians like Paul DePodesta (San Diego) or Jack Zduriencik (Seattle) in response to Billy Beane and Moneyball, but they haven’t been as well publicized as Beane. Plus, just because a major league team is putting less emphasis on the older statistics** doesn’t mean that the entire baseball world is doing the same. Baseball executives are engaged in the sport year round. The casual fan might own a jersey or two, some t-shirts of his/her favorite team, and watch a few games during the year. Hell, even a moderate fan probably only watches around half of the games during the season and can’t even recognize who Jamie Moyer is, or even that he’s the oldest current major leaguer right now. Comparing these people as equal is not a wise decision. Lastly, I do admit that www.mlb.com and ESPN including OPS on a player’s main profile page and stat line is an improvement, but I barely (if even) count OPS as a sabermetric.
* Just ask Patrick Reusse.
** You can also include players like Zack Greinke with FIP and Brian Bannister and Max Scherzer with pitchFX here.
As for Moneyball, few things have annoyed me more than the reactions of many people after reading or hearing about this book. The reason I say “hearing of” is because many people don’t realize that Billy Beane did not write the book about himself. The author was actually Michael Lewis, but so many critics are bothered by the premise of the book that they don’t even bother reading it. Another issue I’ve commonly heard is if Beane was such an expert on building a baseball team at the beginning of this decade, then how come the A’s haven’t been a good team for the past few years? I think this is another case of simply not being informed. Beane acquired Jack Cust from San Diego and freed him from Triple-A purgatory in 2007 because of the home run potential that Cust had displayed in the minors. He made eventual AL Rookie of the Year winner Andrew Bailey the closer at the beginning of the season when he was unproven. He did stray from his typical gameplan when he traded for Matt Holliday, but quickly returned to it midseason when he shipped Holliday away to St. Louis for prospects. In 2009, the A’s were right in the middle of the pack in runs scored, ERA, and UZR. Their Pythagorean win expectation was .499, so they should have had an 81-81 record in ’09, compared to their actual 75-87, which was last in the AL West. Besides, it’s not fair to use the Athletics as the poster child for sabermetrics when other teams (notably the Seattle Mariners with defense* and the Boston Red Sox with just about everything) have been implementing their own emphasis on sabermetrics into their teams.The problem is that these teams haven’t been advertised as much, so maybe someone else should start writing a new book or make a movie to get the news out to the rest of the world.**
* I’m assuming the Mariners used something like UZR to realize that they should get Franklin Gutierrez from the Cleveland Indians in that three-team trade last year. Also, they’re noticeably building a great defensive team, which I cannot remember ever happening.
** By the way, I call dibs on the Mariners.
One of the most discouraging thoughts that I had relating to this post was that although I claim to find both sides of this argument as being useful, it’s much easier for me to defend sabermetrics than the traditional stats. After all, I am planning on becoming a high school math teacher. Anyway, here’s my attempt to argue in favor of the older statistics.
Basically any statistic that’s been around before this relatively recent sabermetric wave is pretty easy to understand. Batting average = probability the player has gotten a hit. ERA = number of earned runs the pitcher has given up per 9 innings. With something like FIP, you need either a more vague (eliminate defensive influences…well alright, how?) or longer description (not worth typing, look forward to Part 3 for this) to get the main point across of what it is trying to measure. Due to the elementary level of comprehension, it’s understandable as to why these statistics are the most popular, generally speaking. It’s sort of like comparing high school to graduate school: The traditional statistics are something that every baseball fan has seen and can understand pretty well, while sabermetrics are the ones that only a handful of people actually study.
Another thing is that while the traditional statistics don’t necessarily paint the best picture of a player, they still do a fairly good job at giving you a rough idea of how a player is performing. It’s obvious that a pitcher with a 6.00 ERA isn’t pitching well or a hitter batting .300 is having a good season. Sabermetrics might tell you if the pitcher has been unlucky or the hitter has been lucky, but you still can’t argue against how that player has been performing up to that point. Unfortunately, I don’t have much more to add in favor of the older statistics, which is why I was hoping to get someone like TT from Granny Baseball to help me with this post. If any of you have some ideas, I highly encourage you to leave them in the comment section.
Personally, I agree with both sides. In my post about the possible starting pitchers available to the Twins, I included a sabermetric in FIP, a couple stats that I don’t consider to be sabermetric but others might argue otherwise, like K/9, BB/9, and HR/9, and some of the more traditional stats like ERA, WHIP, and AVG. Despite my clear one-sidedness in this post, I was much more even when evaluating the pitchers a week and a half ago. When I talk about hitters, I almost always start with their AVG/OBP/SLG line, which I do consider to be an effect from the new prominence of sabermetrics (shying away from the AVG/HR/RBI that used to be popular) but OBP and SLG were not byproducts of Bill James or anyone else.
I really think that these can be used together. If they aren’t, then at the very least it’s preferred if a person remains level-headed when arguing for or against one side of the topic. I may not necessarily be like everyone else, but hey, I knew I was different even before I started seriously liking baseball.
The next post in this series will highlight some of the most common sabermetrics and why they are used. If you would like some additional reading on this topic, I recommend this link, which relates English grammar rules with the usage of baseball’s traditional statistics.