Sabermetrics: A Science
In this series I will describe the relatively new phenomena in baseball called sabermetrics. In the 2000's, sabermetrics have really started to gain popularity. Front office decision makers and hardcore, modern fans have really started to pay attention to these stats.
Why aren't sabermetrics well known by casual fans? Well, old habits die hard. Everyone loves batting average and home runs and ERA. However, these simple stats can be influenced by teammates, the ballpark, and worst of all, luck. I am taking on the responsibility of teaching you about the improved statistical analysis called sabermetrics.
Installment I: Win Shares
In 2002, Bill James, sabermetrics extraordinaire, wrote a book called Win Shares. Yep, it took an entire book to describe this stat. In the book, James uses a very complicated formula to determine how much one player contributes to a win for their team. He also ranks players in various ways in the book. It sounds like a very good read.
Like many other metrics (mostly the incredibly complicated ones), win shares was created by Bill James, who was hired by Theo Epstein to work for the Red Sox organization. He is possibly the most respected statistician (he coined the term sabermetrics) in baseball. It took James 25 freaking years of research to develop win shares. It is viewed as the first stat ever that allows us to compare players of different teams and different eras.
You don't want to know, seriously. The biggest downside to win shares is the fact that is so difficult to calculate. It took a 728 page book for James to describe the formula in his book that focused solely on win shares. In fact, it is so hard to calculate, that I could not find the actual formula on the internet, but here is a website that has a link to a spreadsheet calculating approximate win shares. I have no idea how this guy did it.
What it tells us:
""The purpose of figuring Win Shares was not to wrap up any discussion; rather, the purpose was to open up many more issues, to make additional topics and additional issues more accessible to research."
- Bill James, Win Shares
So, as you can tell, win shares is NOT a be-all end-all, as no stat is. The creator intended for it to aid you in learning about a player and studying up on a player.
Having said this, I know for a fact that FanNation users argue all the time about who is the greatest player of all-time, who is the greatest batter or pitcher of all-time, or even who is overrated and who is underrated. Everyone argues about who had the better season, who had the better career. Win shares is meant to make it easier to compare players who played in different eras and on different teams.
It reduces all statistics to a single number, typically between 0 and 40, that tells us how many wins he contributed to his team. It basically describes how successful a player's season was, just more accurately than on-base percentage and batting average against. Win shares is often used to evaluate potential Hall of Famers, All-Stars, MVP's and more. Here is, generally, what the number tells us (I find this very useful as I am not always sure what is considered good for certain stats):
0 - 10: Bench Player
11 - 20: Good Starter
21 - 30: All-Star
31 - 40: MVP Caliber
Here are some players and their respective win share totals last season. The site I got these stats from will be listed below (Note: Win shares is a surprisingly difficult stat to find because it is so difficult to calculate and Bill James' site requires a subscription):
Top 5 in American League:
1.) Joe Mauer: 36
(.365 BA / .444 OBP / 28 HR / 96 RBI / .996 Fld%)
T2.) Felix Hernandez: 35
(19-5 W-L / 2.49 ERA / 238.2 IP / 1.135 WHIP / 217 K)
T2.) Justin Verlander: 35
(19-9 W-L / 3.45 ERA / 240 IP / 1.175 WHIP / 269 K)
4.) Roy Halladay: 34
(17-10 W-L / 2.79 ERA / 239 IP / 1.126 WHIP / 208 K)
5.) Zack Greinke: 33
(16-8 W-L / 2.16 ERA / 229.1 IP / 1.073 WHIP / 242 K)
Top 5 in National League:
1.) Albert Pujols: 43
(.327 BA / .443 OBP / 47 HR / 135 RBI / .992 Fld%)
T2.) Prince Fielder: 34
(.299 BA / .412 OBP / 46 HR / 141 RBI / .995 Fld%)
T2.) Ryan Braun: 34
(.320 BA / .386 OBP / 32 HR / 114 RBI / .994 Fld%)
T4.) Hanley Ramirez: 33
(.342 BA / .410 OBP / 24 HR / 106 RBI / .983 Fld%)
T4.) Dan Haren: 33
(14-10 W-L / 3.14 ERA / 229.1 IP / 1.003 WHIP / 223 K)
- AL MVP had the most win shares, but the Cy Young Award winner finished behind three other AL pitchers.
- On the surface, the AL is a pitching-dominated league according to win shares.
- The most interesting fact? The defending champs, the Yankees, were the team with the most players above 25 win shares with 6 (and A-Rod wasn't one of them).
- Holy crap, 43? What does that mean? Should Pujols be immediately put in the Hall of Fame?
- Once again, the MVP is the leader in win shares, but there is a much bigger gap between Pujols and the 2nd place finishers.
- The San Francisco Giants were the only team with three players of at least 30 win shares. A sign of things to come?
Note: I am not 100% sure on the legitimacy of these win share statistics. Consider them as rough estimates. I think they may be a bit higher than the actual win shares.
- This is extremely complicated to calculate, but, luckily, it is easy to understand what the final number means.
- There is no way to create a negative win shares, and I think we all know there are some players take away wins from their team.
- Many believe that win shares (in combination with loss shares) are basically just linear weights, which are easier to calculate.
Some find win shares awesome. Others? Not so much. I tend to agree with those others who are not fans on win shares. It is not that they aren't interesting and don't tell us about the players, but I believe that if you want win shares, you might as well look at linear weights. The calculation is complicated and, therefore, a lot of baseball sites (FanGraphs, Baseball-Reference) don't have it listed, even under advanced stats.
Overall, win shares is a difficult stat to calculate and it is hard to be completely certain of the stats authenticity unless you get it from Mr. James himself. Many other metrics are much easier to calculate, understand and get a hold of.
Useful sites on win shares:
- An in-depth article on win shares, written by Michael Hoban, a professor in mathematics. I got a lot of interesting information from here.
- This is the site that I got the win share stats (above) from, click here.
- Here is a site with an unbiased look at win shares that both explains it, but points out its issues.
- Alex Remington wrote a great article on win shares as well. You will probably find the best explanation of win shares, outside of James' book, here.