Fangraphs, now with wOBA

Fangraphs has added wOBA (weighted On-Base Average), one of the stats that improves on OPS by essentially weighing the relative value of on-base and slugging more accurately and scaling it to league average OBA (.335 is average, .400 is awesome, .300 is virtually replacement value). The link above explains the basics in more detail. The Fangraphs version of wOBA does incorporate SB/CS.

The Royals last year were 25th in MLB with a team wOBA of .314. Mike Aviles led the Royals (100 PA min.) with a .360 wOBA, with only David DeJesus (.355) and Alex Gordon (.344) also above average.

12 months ago Gopherballs 24 comments 0 recs |

Story-email Email Printer Print

Comments

yay, I noticed

is it the one with custom linear weights by year or just the standard model?

At stat corner they h ad Callaspo and Grudz also slightly above average

OMG Banny. FWIW I am only crdtng u w/3 runs allwd bc of DDJ OMFG

by devil_fingers on Nov 26, 2008 1:39 AM EST reply actions 0 recs

Langerhans

.353

That would be 3rd on our team

Relive Royals History at royalsretro.blogspot.com

by RoyalsRetro on Nov 26, 2008 9:43 AM EST up reply actions 0 recs

why acquire talent for practically free when you can trade for it?

OMG Banny. FWIW I am only crdtng u w/3 runs allwd bc of DDJ OMFG

by devil_fingers on Nov 26, 2008 11:32 AM EST up reply actions 0 recs

".300 is virtually replacement value"

I think that’s wrong, because Gload is at .290, and there is no way he is below replacement value!!!!

Relive Royals History at royalsretro.blogspot.com

by RoyalsRetro on Nov 26, 2008 9:41 AM EST reply actions 0 recs

I think I read in the comments that .298 is replacement value for 700 PA

But it’s more like .285 for 500 PA – so Gload is still safe. No worries.

by jsolo on Nov 26, 2008 10:59 AM EST up reply actions 0 recs

I wish they'd put wOBA+ on there

This is just my opinion. I could easily be wrong.

by NYRoyal on Nov 26, 2008 10:59 AM EST reply actions 0 recs

Does that adjust for ballpark and position factors?

by jsolo on Nov 26, 2008 11:01 AM EST up reply actions 0 recs

To tell you the truth, I’m not sure. In the very least it compares a player’s wOBA to the league average. I believe it (like OPS+) normalizes for park and league, but not for position.

This is just my opinion. I could easily be wrong.

by NYRoyal on Nov 26, 2008 11:05 AM EST up reply actions 0 recs

I suppose position factors might be too much of a derivative from actual performance

It would be nice to have a very accessible chart that is updated daily/weekly that has league average wOBA across positions. I mean, even for something like OPS, I’m never sure if say .750 OPS is good for a SS (assuming average defense) though I’m sure those metrics are out there if I take the time to look.

by jsolo on Nov 26, 2008 11:14 AM EST up reply actions 0 recs

Agreed

I’d love to know at any given point what the average wOBA is for each position. I have a general understanding of it for OPS, but of course it is somewhat different every season.

This is just my opinion. I could easily be wrong.

by NYRoyal on Nov 26, 2008 11:16 AM EST up reply actions 0 recs

wOBA+

is like OPS+ in that is the park-adjusted precentage of league average

wOBA* (which they have at Stat Corner, along with wOBA+) is park-adjusted

At stat corner they also have bRAA (batting runs above average).

Given Dave Cameron’s recent (and mostly correct, IMHO) critique of position-adjusted stast, I doubt we’ll be seeing anything like that at fangraphs.

The better way to do it would be to it for oneself is to go to stat corner (or so bRAA oneself). For a full season, here are the position adjustments from Tom Tango I like.

These is in wins, so the run value over a full season (and to do it completely right would require prorating) is these times 10.5, but times 10 is pretty close:

DH -1.7.5
1B -1.25
LF/RF -.75
2B/3B/CF +.25
SS +.75
C +1.25

OMG Banny. FWIW I am only crdtng u w/3 runs allwd bc of DDJ OMFG

by devil_fingers on Nov 26, 2008 11:37 AM EST up reply actions 0 recs

sorry, boys

I guess while I do think that’s a better way to do it, it isn’t straightforward, and it doesn’t really answer to what you’re looking for. It would be nice to have positional averages during the season, like b-r has.

OMG Banny. FWIW I am only crdtng u w/3 runs allwd bc of DDJ OMFG

by devil_fingers on Nov 26, 2008 11:42 AM EST up reply actions 0 recs

Wow. The same adjustment is used for 2B, 3B, and CF?

Is that implying that 2B is equally defensively valuable as a 3B?

by jsolo on Nov 26, 2008 12:00 PM EST up reply actions 0 recs

It's a more complicated argument

that I admit I don’t fully get… I’ll try to track down the link.. It’s a combination of defensive value, talent distrubtion, and relative offensive skill

OMG Banny. FWIW I am only crdtng u w/3 runs allwd bc of DDJ OMFG

by devil_fingers on Nov 26, 2008 12:13 PM EST up reply actions 0 recs

Complicated?

I’m sure it’s a simple matter of taking the 3rd derivative of a regression analysis run using a combination of the fibonacci sequence and fractal geometry, all based upon recent sunspot activity.

by jsolo on Nov 26, 2008 12:21 PM EST up reply actions 0 recs

My pet peeve of the day

Different websites using different versions of a stat, but calling it the same thing. Fangraphs and THT both use FIP, but their formulations are different, so the numbers are different. Fangraphs and Stat Corner both use wOBA, but their formulations are different, so the numbers are different. If they calculate them differently, shouldn’t these stats have different names?

This is just my opinion. I could easily be wrong.

by NYRoyal on Nov 29, 2008 3:59 PM EST reply actions 0 recs

hmmm... just noticed that

I wonder what the differences in wOBA are? I don’t think either include stolen bases. On a quick glance, Fangraphs seem to be uniformly lower… I’m not sure why. Two guesses, each equally poorly-founded:

1) The Fangraphs version is lower because it doesn’t include one or more of the following that are in Tango’s basic formula: HBPs, iBBs, and/or Reaching base on error, or

2) The Fangraphs version is the “custom linear weights” version that Tango just released as a MySQL query, which gives different event values based on their particular contribution to run-scoring each year. he did mention something about Fangraphs starting to post wOBA in the thread following his posting fo that query…

OMG Banny. FWIW I am only crdtng u w/3 runs allwd bc of DDJ OMFG

by devil_fingers on Nov 29, 2008 10:03 PM EST up reply actions 0 recs

Could be

Regardless, if you find out which version of wOBA and/or FIP is better, please let me know. I’d hate to have to say that a player’s wOBA is .326-.333

This is just my opinion. I could easily be wrong.

by NYRoyal on Nov 29, 2008 11:24 PM EST up reply actions 0 recs

I think there are multiple formulas for both

I’m not sure who uses what for FIP. I’m not even sure which is “better.” You’re the pitching guy, you tell me! I guess if I had to choose, I’d use THT’s, because it has an xFIP that matches up nicely. Although I don’t know exactly how tRA works, I’d take it over all of them if forced to choose because it incorporates batted ball data and park adjustments. I also like that it goes by Runs Allowed rather than Earned Runs, but that’s just me. Of course, one doesn’t have to choose. I liked what you did in your post on rotation slots where you sort of combined ERA/FIP/xFIP/tRA.

I’m not sure who uses which formula for FIP, or how to determine which is better. The one I use when I need it (and I think it matched up with Fangraphs, but I don’t remember) is

((HRx13 + BBx3 + HBPx3 -SOx2)/IP) + 3.2

As for wOBA formulas, I’m not sure who one would judge which is better. Assuming that #1 is the issue, then it depends on who much you think HBPs, iBBs, and ROEs reflect hitter skill or not. If #2 is the case, then it depends on whether you want wOBA to measure something that has a relative constant formula from year-to-year in the way that OBP, BA, SLG, etc. have constant “weights” for each event, or something that precisely measures the batter’s value per PA in the run-scoring environment of the year.

For now, I like Stat Corner’s version because they have park adjustments, too. But ask me in a couple days and I probably will have changed my mind again.

OMG Banny. FWIW I am only crdtng u w/3 runs allwd bc of DDJ OMFG

by devil_fingers on Nov 30, 2008 12:03 AM EST up reply actions 0 recs

FIP formulas

The difference between the two might be the “3.2”. FIP uses a league-specific constant which changes from year to year. It is possible that Fangraphs and THT have different methodologies for coming up with that constant. I prefer tRA as well but, as you know, I like to go with multiple stats. I guess I’ll have to go with multiple FIP’s too.

This is just my opinion. I could easily be wrong.

by NYRoyal on Nov 30, 2008 2:47 AM EST up reply actions 0 recs

For FIP

THT adjusts the constant so that league average FIP equals that season’s league average ERA, while I think Fangraphs and others use the traditional 3.2 constant. The difference is usually pretty small (less than 0.10).

As noted above, the Fangraphs wOBA includes SB/CS, while StatCorner does not (at least last time I checked), which leads to a small difference as well.

When citing wOBA or FIP, the easiest thing to do is just use one version and indicate which site (THT, Fangraphs, etc.) you are using.

by Gopherballs on Nov 30, 2008 1:41 PM EST up reply actions 0 recs

I should have thought about that

But that would make Gordon’s wOBA higher, so there must be other differences.

GB, if you’re still reading this, do you know if they’re using the custom weights version or not?

OMG Banny. FWIW I am only crdtng u w/3 runs allwd bc of DDJ OMFG

by devil_fingers on Nov 30, 2008 4:10 PM EST up reply actions 0 recs

Yes, if I am reading the various links correctly

according to the other wOBA article on Fangraphs, the site is using this version of Tango’s run values. It apparently does not include HBP or reaching on errors, or as one the comments notes, pitcher’s hitting. StatCorner confirms that they use reaching on errors. While unclear, StatCorner also may scale its wOBA to AL and NL average, not MLB.

by Gopherballs on Nov 30, 2008 7:02 PM EST up reply actions 0 recs

I'm such a dork

I ran that MySQl already. I personally prefer Statcorner for the park adjustments, unless Fangraphs is already ddoing it. Add in SBs in my head isn’t that hard. The park adjustments also makes the bRAA more accurate.

OMG Banny. FWIW I am only crdtng u w/3 runs allwd bc of DDJ OMFG

by devil_fingers on Nov 30, 2008 11:47 PM EST up reply actions 0 recs