Explaining Analytics: wOBA

Explaining Analytics: wOBA

Credit: Getty Images

You don't need wOBA to tell you Adrian Beltre is the Rangers' best offensive player. But it does help - and it will tell you a lot more about a player than batting average does. (Photo by Al Messerschmidt/Getty Images)

Print
Email
|

by JOSEPH URSERY

WFAA Sports

Posted on December 24, 2013 at 10:27 AM

Updated Tuesday, Dec 24 at 10:29 AM

Whoa... I know wOBA...

 
What is wOBA?
 
WOBA stands for Weighted On-Base Average. The theory behind wOBA is simple; weigh each hit outcome so you can use one number to accurately display a player's offensive output, add 'em all together, and use one neat, tidy number to represent a player's total offensive production. Other metrics try to do this (and, arguably, some do it better), but wOBA is simple, easy to understand, and does a pretty good job of wrapping up all the juicy details in one friendly, batting average-like number.
 
What should I use wOBA to replace?
 
Put simply, anytime you want to use batting average (BA), use wOBA. It works a lot like batting average, except BA treats each hit the same way. Imagine you had one player who hit .300 but everything was a single; on the other hand, you have a guy who only hits .200, but everything's a home run. Obviously, weird players like this don't exist (although Adam Dunn and Joey Gallo are tryin'!), but BA treats the first player like a much better player than the second (although, in reality, you would rather have the contributions of the second player than the first, everything else equal). 
 
As another fan-friendliness portion, wOBA is scaled to look like OBP. This means that anything below .330 is typically undesirable, while over .330 is good. It also scales itself year-by-year; this is good if you want to compare players across eras. It's also one of many, many stats you can use to highlight how good Babe Ruth was; his career wOBA is a full twenty (TWENTY, GEORGE!) points higher than the second best all time mark.

What does wOBA say about the 2013 Rangers?
 
I hate to spoil things for people who haven't caught up on DVR, but Adrian Beltre had the best pretty much everything offensively on the 2013 Rangers. It follows, then, that he had the best qualified wOBA on the 2013 Rangers. His mark of .379 was twenty (TWENTY, GEORGE!) points higher than the second place man, Nelson Cruz. Why is this? Well, both he and Cruz had fantastic slugging seasons (Beltre slugged .509, Cruz .506) but Beltre did it with a far superior batting average (.315 to .266) and OBP (.371 to .327).
 
In other words, wOBA takes an otherwise unwieldy triple slash line of batting average, on-base percentage, and slugging percentage (in Beltre's case, .315/.371/.509) and condenses it to one easy to use number (.379) that also includes data about how his number compares to the rest of the league. In one handy metric, you convey about eight times the information that batting average gives you. Seriously, if wOBA was a something we sold on an infomercial, we wouldn't even have to hire the ShamWOW guy to sell it, because it just sells itself.    
 
So what does wOBA not do?
 
Well, for one, say you were to head over to the fine folks at Fangraphs.com and look at the 2013 Rangers, sorted by wOBA, you would notice that Adrian Beltre isn't actually top on that list; Joey Butler is. Joey Butler had 15 plate appearances for the Rangers, in which he struck out in 40%.  Heck, even Jeff Baker is second on the list, despite only 175 plate appearances due to a freak high-five injury incident and a focus on only hitting lefties. Do those 190 plate appearances make those two players better than Adrian Beltre? No of course not. I had to refrain from punching myself just for typing that. This displays one of the weaknesses of wOBA, which is it doesn't reward a player's durability or ability to play every day.  Thus, absent using a qualifying number of plate appearances, it makes Joey Butler's .426 (again, over 15 PAs) look just a wee bit better than Mike Trout's .423. Joey Butler was rewarded for his season by being diesgnated for assignment by the Rangers and claimed through waivers by the Cardinals; Mike Trout is about eleven months away from shattering the previous arbitration payday record. 
 
As well, wOBA doesn't adjust for park factors. In other words, a player who plays his home games in a park that inflates offense (such as the Ballpark in Arlington) should be expected to have a higher overall wOBA than one who plays in a run-supressing ballpark (Petco Park in San Diego, for instance). In other words, wOBA isn't the best route to compare Adrian Beltre and Chase Headley, as it makes Beltre look worlds better than Headley (when in reality he's only a lot better, rather than worlds better). 
 
As well, wOBA doesn't look at a player's position. It treats Mike Trout's .426 in centerfield the same way it treats Chris Davis' .421 at first base; although offense is easier to find at first than it is in centerfield. If ever you doubt this, look at Prince Fielder standing next to Leonys Martin sometime.  Actually, do that even if you don't doubt it, because that would be a pretty amazing picture.
 
Of course, we can use other metrics to normalize for durability, position, and ballpark, but that's an article for another day.
 
Still, given it's weaknesses, wOBA is a great stat for its ease of use, all-around measure, and fun pronouncability (seriously, say “wo-BAH” out loud to yourself right now. It's ok. If anyone looks at you cross eyed, give 'em my twitter handle and tell them I said I'd square them out). It's not perfect (no stat is, that's key to remember, that's why we use multiple stats!), but, anytime you feel compelled to use batting average, try it instead.  You're guaranteed to either win arguments or friends (or, at least, a few black eyes). 
 
WFAA does not condone you referring people to @thejoeursery on Twitter because you said wo-BAH out loud. WFAA does recommend you follow him on twitter, because he's generally knowledgeable about sports and food and general fatherhoodery.

 

Print
Email
|