How do you calculate wOBA?

by Kirk on May 26, 2009 3:07 PM PDT

I have seen this stat often on this site and I understand what it is and its purpose but what I don't know that I want to know is how to calculate it. I would love to be able to calculate it myself and I'm sure other people on the site would as well.

I found this explanation on how to calculate it but the way it is explained is a little over my head as it looks like how you would do it on an Excel spread sheet and I do not know how to use excel. If anyone can explain to me how to do it that would be great.

Tweet Comment 65 comments | 0 recs |

Do you like this story?

Comments

wOBA is a similar calculation to tRA

there’s no real shorthand way to calculate either, since they are both the summations of individual results over the course of a season.

by seattlebruin on May 26, 2009 3:14 PM PDT reply actions

I understand that there is no shorthand way to do it but I'm willing to spend some time on it.

by Kirk on May 26, 2009 3:18 PM PDT up reply actions

wOBA (without SB/CS factor)

0.72*uBB + .75*HBP + .9*single + 1.24*double + 1.56*triple + 1.95*homerun + .92*rboe

by Matthew on May 26, 2009 3:17 PM PDT reply actions

You forgot to divide by (plate appearances - IBB)

by Graham MacAree on May 26, 2009 3:19 PM PDT up reply actions

Yeah, that. I only listed the numerator.

by Matthew on May 26, 2009 3:22 PM PDT up reply actions

Out of curiosity, how would you factor SB/CS into wOBA?

Is that the light at the end of the tunnel, or the headlights of an oncoming train?

by Benne on May 26, 2009 3:19 PM PDT up reply actions

I wouldn't, and don't.

If you wanted to however, you’d need to come up with weights in the same manner as done to come up with the weights for the above events.

by Matthew on May 26, 2009 3:24 PM PDT up reply actions

So wOBA treats all outs equally?

also, a little confused as to why your values differ from these linear weights

by seattlebruin on May 26, 2009 3:20 PM PDT up reply actions

Because you don't want 0.00 to be average

by Graham MacAree on May 26, 2009 3:21 PM PDT up reply actions

Gotcha

by seattlebruin on May 26, 2009 3:50 PM PDT up reply actions

Oh that seems to be very easy.

Thanks Matthew. I have just one question though, what is rboe?

by Kirk on May 26, 2009 3:25 PM PDT up reply actions

Thank you guys for the help.

I am sure you have answered this question many times before and I’m glad you answered it once more.

by Kirk on May 26, 2009 3:46 PM PDT up reply actions

Why is reaching base on error worth more than a single?

Is it because you can get more than one base on an error?

Determined, Jonesing Commentor | Proud proprietor of Wyomingroutes.org & Washingtonhighways.org

by Corco on May 26, 2009 11:02 PM PDT up reply actions

Pretty much.

But there’s more to it. I’m going to have to let others explain it, though.

Fans are typically idiots.

by The Typical Idiot Fan on May 27, 2009 2:12 AM PDT up reply actions

Reply fail.

by Kirk on May 26, 2009 3:26 PM PDT up reply actions

Is there an easy way to find a batter's RBOE?

I will smash your face into a jelly.

by Phil Hatzenbuehler on May 26, 2009 3:49 PM PDT up reply actions

As Jeff said above, B-Ref has it

by Graham MacAree on May 26, 2009 3:50 PM PDT up reply actions

This is also a helpful link

http://www.insidethebook.com/woba.shtml

by Jeff Sullivan on May 26, 2009 3:57 PM PDT reply actions

SB/CS

wOBA at FanGraphs includes stolen bases and caught stealing. The values for SB/CS in the formula are always pretty close to .5 and .25 respectively.

by davidcameron on May 26, 2009 5:14 PM PDT up reply actions

I assume you meant .25 and .5

by Jeff Sullivan on May 26, 2009 5:51 PM PDT up reply actions

Why is it that CS has a value?

I would think getting caught would have a negative value or none.

by Kirk on May 26, 2009 9:53 PM PDT up reply actions

Should be negative

A steal is about +0.25, while getting caught is about -0.50.

by Jeff Sullivan on May 26, 2009 10:21 PM PDT up reply actions

Has that changed?

I thought it was a 3 successful steal to 1 caught stealing ratio. Those numbers would make it a 2:1 ratio.

Fans are typically idiots.

by The Typical Idiot Fan on May 27, 2009 2:13 AM PDT up reply actions

2:1 would be neutral

you have to successfully swipe the third before you’ve positively contributed to run scoring.

by seattlebruin on May 27, 2009 8:07 AM PDT up reply actions

Or this is a really rough estimate, I've always heard 70% is the number at which you're making a positive contribution

by seattlebruin on May 27, 2009 8:08 AM PDT up reply actions

They're approximations

After doing some research, the true values as derived by Tango are closer to +0.2 and -0.47.

by Jeff Sullivan on May 27, 2009 9:38 AM PDT up reply actions

I have another StatCorner related question:

Why is Zack Greinke’s tRA+ 175, while his tRA is 1.29 and the league average is 5.06? 1.29/5.06 is 3.92, so why doesn’t he have a 392 tRA? Or am I way off base about how ERA is calculated?

by Decatur on May 27, 2009 12:32 PM PDT reply actions

*a 392 tRA+ I mean

by Decatur on May 27, 2009 12:33 PM PDT up reply actions

From the Glossary

tRA+
Like wOBA+. tRA+ is equal to [((lgTRA – tRA) / lgTRA) + 1] * 100. This formula puts tRA+ on the same scale as wOBA+. A tRA+ of 150 represents that the pitcher’s tRA was 50% better than league average. A tRA+ of 75 means the pitcher’s tRA was 25% worse than average.

by Jeff Sullivan on May 27, 2009 1:03 PM PDT up reply actions

Damn, I should've just looked there.

Thanks for answering my question in any case, Jeff.

by Decatur on May 27, 2009 3:00 PM PDT up reply actions

This might better be suited for the statistical analysis thread

but how do you convert tRA into runs?

St. Louis Cardinals... defying win expectancy since 2008

by vivaelpujols on May 27, 2009 8:19 PM PDT up reply actions

I mean

if you apply the coefficients on Graham’s primer to batted ball rates of players, how do you convert those to tRA?

St. Louis Cardinals... defying win expectancy since 2008

by vivaelpujols on May 27, 2009 8:20 PM PDT up reply actions

How far back have you gone in your tRA reading?

I’m pretty sure he explained this in a huge post sometime a while ago.

by Jeff Sullivan on May 27, 2009 8:49 PM PDT up reply actions

I read the tRA primer, but I couldn't find anything else on it on LL (besides the "no numbers" version)

It says the formula is tRA = expected_runs/expected_outs*27. I tried doing that and I got a “bad” number. Maybe I’m calculating xRuns and xOuts wrong. Is it just the out or run coefficients for each outcome, multiplied by the frequency, and then added all together?

St. Louis Cardinals... defying win expectancy since 2008

by vivaelpujols on May 27, 2009 9:27 PM PDT up reply actions