wOBA = (0.691×uBB + 0.722×HBP + 0.884×1B + 1.257×2B + 1.593×3B + 2.058×HR) / (AB + BB – IBB + SF + HBP)
That above is the formula for wOBA (Weighted On Base Average). Looks complex, doesn’t it? It isn’t. I should point out that is was just last week when it hit me exactly what wOBA was measuring and gave me a good anecdote. When Tom, Rachael, and I had our pre-Spring Training meeting on the direction of Nationals coverage for the site Tom had the idea to do a fan Spring Training. To write a serious of articles about advanced stats. This scared the hell out of me, because there are a lot of people smarter than me and many of them happen to be Nationals fans and some of them may be reading this right now. The big fear I had was that in explaining advanced stats my tone would come across as if I was talking down to my audience.
So last week I was working on some things trying to figure out how many runs Denard Span at lead-off and Wilson Ramos at catcher would be worth. You can look at the wOBA formula and see that it assigns a run value to each batting event, and these values are based on the average run expectency of the 24 base out states. If we remove all context then a base is worth 0.25 runs, because you need four bases to score a run, but that would be in a line-up where no one ever hits anything but a single or walks and no batter ever advances more than one base. So more or less a base is only worth 0.25 runs in a world that doesn’t exist. I kept thinking about it and trying to figure how much to add to get a truer value based on the game of baseball that is actually played.
That is when it hit me. What wOBA is measuring is the average offensive value for times on base for a player. As soon as I got home I rushed to my computer to check the Fangraphs glossary to see if I missed anything. It said it in the third paragraph, but not in as plain of English as it is the average run value for times on base for a player. And really it isn’t even that it is the average run value for times a player reaches base in a plate appearance. Steals, taking the extra base, advancing on fly balls, passed balls, and wild pitches all add run value but wOBA is a batting stat and all of those are base running events.
Now that I have explained how it took me way too long after starting to use a stat to understand what exactly the stat was measuring I will say that all stats tell a story and the key to understanding, and using them effectively is to understand that story. wOBA uses the league average run expectancy and assigns weights to events that can occur during a plate appearance. It is better than both OBP and SLG because both those stats ignore certain events. OBP gives a percentage of time a batter will reach base per plate appearance and SLG tells the average number of bases a player will bat their way to per at bat. OBP completely ignores the number of bases reached and SLG ignores anything that isn’t a hit.
wOBA not only includes all those things but also gives them context. Why is a homerun worth 2.058 runs? Well it is that way because a homerun is worth between 1 and 4 runs depending on the base/out state and the proceeding base/out state is going to have a run expectancy value as well. Here is the run expectancy matrix for 2012. Now if we were to sit here and average the value of a homerun through all base out states we would come out to the wOBA value, but to give an idea a solo homerun to lead-off an inning would be worth 1.4886 because it is the value of the one run scored plus the remaining run expectancy value. The same would apply for singles, doubles, triples, walks, and outs. The value of the event is the resulting base/out state minus the proceeding base out state. So in 2012 a lead-off single raised the run expectancy from 0.4886 to 0.8557 and the value of the single is the difference between the two states.
wOBA is a fascinating stat as it gives the number of runs that can be expected per time a batter reaches base. Remembering the stat that way and knowing the story behind it makes it more useful than simply looking at the league average of around .330 and figuring a good hitter from there. Neither are bad methods and the former is what I did before my minor epiphany. wOBA is a valuable offensive stat as a catch all, but to many people it is still easier to be given the raw OBP and SLG numbers. When trying to tell the story of a batter and if they are good or not wOBA doesn’t separate the information. So while both Ian Desmond and Jayson Werth were .362 wOBA hitters in 2012 wOBA doesn’t inform us that Werth had a much better OBP and Desmond a much better SLG making Werth the more ideal top of the order batter and Desmond the more ideal middle to bottom of the order batter. Overall wOBA is the best single offensive stat, but no one stat tells the entire story.