wOBA
The statistic wOBA (weight on base average) is now available in the player pages, leaderboards, team pages, my team pages, and the projections.
wOBA, created by Tom Tango, is a version of linear weights that has been weighted to fit an OBP scale. The weights have been properly adjusted by season and for the minor leagues by season and by league. For more information about wOBA, check out the links below:
Weighted On Base Average or wOBA
wOBA year-by-year calculations
Getting to Know wOBA
David Appelman is the creator of FanGraphs.
Excellent addition
David Appleman can read my mind. I was JUST thinking about why fangraphs didn’t include wOBA.
Appelman, sorry.
It would be nice to know what the league average for 2008 was, along with what would be considered replacement level.
It’s my understanding that league average is always .330.
Scappy, a .330 league-average wOBA would put replacement level right around the .295 mark. Position-wise, using offensive averages based on my own faulty multi-year calculations:
Catcher: .312 average, .282 replacement
First base: .352 average, .311 replacement
Second base: .323 average, .291 replacement
Third base: .336 average, .299 replacement
Shortstop: .316 average, .286 replacement
Corner outfield: .341 average, .303 replacement
Centerfield: .325 average, .292 replacement
David, what’s the basis for the wOBA weights? Do they include the impact of the park?
All the weights were derived using this method: http://www.insidethebook.com/ee/index.php/site/article/woba_year_by_year_calculations/
No park adjustments, yet.
Thanks, David. Your site is really humming these days. Congrats!
Great addition!!
Firstly, awesome. The fangraphs leaderboards UI is so much nicer and faster than BPs. It’s great to have wOBA up there. Now if only you listed PA on the leaderboards…
Second, a question on wOBA. Forgive my ignorance, but I’ve never understood the need to put wOBA or EqR on the scale of another statistic. Why not simply including run value of negative events in the equation and make it a per PA measure of run production?
Other than familiarity of a scale, what’s gained? Isn’t the value of that familiarity harmed by the inability to easily interpret.
A very large reason why batting average, on base, and slugging have the traction they have is because one can easily look at that .293 average and say “that guy gets a hit 29.3% of the time”. Sure, there are some caveats given the “AB” denominator, but it still sticks. We can easily interpret the meaning of that number, even outside of a specific league context. We may not know immediately if .293 is good, but we know what it means.
It seems that the additional machinations to make wOBA on a familiar scale are only done for the purposes of easy consumption. However, in fitting the metric to the familiar scale for purposes of valuation, you complicate the extremely important “unit of measure”, obscuring the interpretation of any given value.
Why not just leave it as .357 runs per PA? Or is this just runs created, base runs, what have you by a different name?
Rick, I talked about the scale ad nauseum on my blog when I first rolled it out. The short of it is that I needed a binomial-like metric for The Book, since I needed to establish statistical significance. Putting it on the OBP scale helps accomplish that.
Also, if you know the player’s OBP and his wOBA, and you see it’s a match, then you know he’s got a proportionate SLG. The more his wOBA exceeds his OBP, the more his value is tied to his SLG. And vice versa.
Finally, the sum of the weights*frequency for each event is identical to the sum of the frequency themselves. So, it’s got a built-in scaling mechanism to make sure all the weights (walk, hit, HR) is properly weighted.
David, Just curious…. Whats the process like to add a stat to the fangraphs site? What things do you consider? Who is involved in the decision making process. Thanks in advance
Thanks Tango. I saw mgl just posted a wOBA thread over in your hood — I’ll go educate myself a bit. Probably should have picked up my copy of the The Book first in any case.
What are you including in your calculation of wOBA? How are you treating IBB and HBP? I am wondering specifically if SB are included in this calculation of wOBA.
Reggie: SB and CS are included in the calculation. HBP are also included and IBB are excluded for years where IBBs are available so post 1954.
Thank you. I was just wondering since once you get to stuff like SB and IBB they wouldn’t have much impact on the final number except for extreme cases. Also I don’t think SB/CS were usually included in any of the wOBA calculations in “The Book.” I just like to know what goes into a stat.
No problem Reggie. There’s an interesting thread on the value of the IBB over at insidethebook.com. It starts here at comment #21 and continues for a bit: http://www.insidethebook.com/ee/index.php/site/comments/rob_neyer_touts_woba/#21
So, I am getting a spreadsheet together with wOBA to try to rank some players in context. I have calculated wOBA using Tango’s original formula ((0.72xNIBB + 0.75xHBP + 0.90x1B + 0.92xRBOE + 1.24x2B + 1.56x3B + 1.95xHR) / PA), done it w/ just BBs instead of NIBB, and used SB/CS (+.22 for SB and -.38 for CS). Yet somehow I still end up w/ different values. I’d like to use the same formula so I can refer readers over here, but I can’t seem to get the exact formula down…
HELP!
Ryan, the actual weights used on FanGraphs are based off these calculations: http://www.insidethebook.com/ee/index.php/site/article/woba_year_by_year_calculations/
And the final results of those calculations are here: http://tangotiger.net/bdb/lwts_woba_for_bdb.txt
The final weights that are calculated at FanGraphs are essentially the same as the ones presented in that file, though there might be extremely small differences. We don’t round numbers behind the scenes.
Second, we are not using RBOE because we just don’t have that.
Third, we only use NIBB when available. IBBs are not available before 1954 (or something like that), so before that we just use all BBs.
Finally, our career wOBA lines reflect the different weights for different seasons, so you would need to individually calculate each season’s wOBA and then weight each wOBA appropriately in a full season calculation.
Also, if you are doing a quick calculation for wOBA, I believe .+0.25 and -0.50 are better for the SB and CS. If you look at that file, most of the recent CS weights are near -0.50.
As you can see, there’s a bit of heavy lifting done behind the scenes mainly because we have to account for over 100 years of baseball, but if you use Tango’s standard calculation I think your numbers will come out very close to those on FanGraphs, assuming you change the weights on your CS and SB, mainly the CS one.
Thanks
In the irony of all ironies, I actually had those weights on the next tab and didn’t even think to use them. Thanks again.
Why is the value of the out subtracted from the weights? I never understood this.
When calculating wOBA w/o ROE stats, do you just leave that out of the calculation altogether, or use an assumed number for each player?
excellent areoter with fagud sunsune for treave edunbe. cojunt have carial that ceasing it corry and precoting, scimived proven.
Amazing blog! Is your theme custom made or did you download it from somewhere? A design like yours with a few simple adjustements would really make my blog shine. Please let me know where you got your theme. Thanks
I as well as my guys ended up studying the great secrets and techniques found on your web site while instantly came up with a horrible feeling I had not thanked the web site owner for those techniques. All the ladies happened to be totally thrilled to see all of them and have pretty much been taking advantage of them. Many thanks for indeed being so thoughtful and also for choosing certain helpful subject matter millions of individuals are really eager to know about. Our own sincere apologies for not expressing gratitude to you sooner.