View Full Version : Issues and questions with OPS formula

Let me preface this post by saying that I appreciate OPS as a stat, and definitely see its value.

That said, I've been thinking about its construction, and I got to wondering if certain things, like hits, aren't overvalued in the formula. I'm sure this has been brought up before, but I've never seen it here. And I know there are people much more well versed than me who can wax philosophic on what I'm trying to discuss.

OPS Formula:

(h+2b+3b+hr)/ab + (h+bb+hbp)/(ab+bb+hbp+sf)

The formula for OPS counts hits twice. Once in the OBP formula, and again in the SLG formula. Then when those parts are added together, you've essentially double-counted hits. This takes away a little from the TB aspect of the formula, too, since a single is valued much closer to a double when it is counted twice.

Does that make sense? Does that give an accurate representation of what OPS is trying to measure?

To me, a better stat would be (BB+TB+HBP+SF)/PA. That way hits are only measured once and total bases are valued a bit higher. I also moved SF to be combined with BB, TB and HBP because I argue that sacrifice flies are a productive out.

So what do you all think? Does this make sense, or am I way off base here?

elfmanvt07

05-27-2012, 08:26 AM

Not that you're wrong, but aren't you equivocating hits and singles?

Slugging is actually

(1B+2(2B)+3(3B)+4(HR))/AB

elfmanvt07

05-27-2012, 08:29 AM

What I mean to say is that ALL hits are counted in OBP, and each type of hit separately in SLG. Not all hits twice in addition to doubles, triples, and taters.

elfmanvt07

05-27-2012, 08:34 AM

Third comment in a row, yeah!!

I was going to say that I do like that your stat gives a positive credit for sac flies.

_Sir_Charles_

05-27-2012, 09:53 AM

Let me preface this post by saying that I appreciate OPS as a stat, and definitely see its value.

That said, I've been thinking about its construction, and I got to wondering if certain things, like hits, aren't overvalued in the formula. I'm sure this has been brought up before, but I've never seen it here. And I know there are people much more well versed than me who can wax philosophic on what I'm trying to discuss.

OPS Formula:

The formula for OPS counts hits twice. Once in the OBP formula, and again in the SLG formula. Then when those parts are added together, you've essentially double-counted hits. This takes away a little from the TB aspect of the formula, too, since a single is valued much closer to a double when it is counted twice.

Does that make sense? Does that give an accurate representation of what OPS is trying to measure?

To me, a better stat would be (BB+TB+HBP+SF)/PA. That way hits are only measured once and total bases are valued a bit higher. I also moved SF to be combined with BB, TB and HBP because I argue that sacrifice flies are a productive out.

So what do you all think? Does this make sense, or am I way off base here?

I see that somebody already corrected the H = 1b thing. But in your suggested formula.... (bb+tb+hbp+sf)/pa aren't you also double counting several items? If you take a walk, doesn't that count towards your total bases? Same for a HBP. The way slugging is, at least it differentiates between the types of hits and how the total bases are acquired. Seems more informative and useful to me. Just my 2 cents.

Brutus

05-27-2012, 10:02 AM

OPS isn't really intuitive in its raw nature. It doesn't really jive with how runs are created. There are run estimators that correlate with run scoring a little better (wOBA, GPA, etc.), but because it's more easily calculated and has become mainstream, it gets more favor.

There are a couple of good articles discussing the shortcomings of OPS compared to wOBA here (http://www.insidethebook.com/woba.shtml) and here (http://www.fangraphs.com/blogs/index.php/a-visual-look-at-woba/).

Here's a nice graphic that sums up the differences in weighting between OPS and wOBA:

http://www.fangraphs.com/blogs/wp-content/uploads/2011/04/woba5.png

PuffyPig

05-27-2012, 10:20 AM

IIRC, the OBA component of OPS is worth more in run correlation, such that if you multipy OBA by 1.7 and add it to SL% it correlates to run scoring more accurately.

Not that you're wrong, but aren't you equivocating hits and singles?

Slugging is actually

(1B+2(2B)+3(3B)+4(HR))/AB

Well, with every hit you're getting at least 1 TB (with a double you get +1TB, a triple +2TB, and a home run +3TB). I think someone above posted a nice graphic showing the co-efficients, and that helps explain it.

RedsManRick

05-27-2012, 10:35 AM

The only reason to use OPS instead of wOBA is because we often have OBP and SLG handy. That's it. It not simply to calculate from it's component pieces and it's got some structural issues. If we moved to wOBA and used OPS as simple a rough approximation, we'd be doing a better job.

There are a couple of good articles discussing the shortcomings of OPS compared to wOBA here (http://www.insidethebook.com/woba.shtml) and here (http://www.fangraphs.com/blogs/index.php/a-visual-look-at-woba/).

Here's a nice graphic that sums up the differences in weighting between OPS and wOBA:

http://www.fangraphs.com/blogs/wp-content/uploads/2011/04/woba5.png

As usual, Nate hits a grand slam. But it's only worth 2 runs.... :p

elfmanvt07

05-27-2012, 11:04 AM

Well, with every hit you're getting at least 1 TB (with a double you get +1TB, a triple +2TB, and a home run +3TB). I think someone above posted a nice graphic showing the co-efficients, and that helps explain it.

Right. SLG is TB/AB. But, in your original post, your math indicated that it was far more than TB, with every XBH being counted more than once.

You typed that SLG was (H+2(2B)+3(3B)+4(HR))/AB

It's actually (1B+2(2B)+3(3B)+4(HR))/AB

By your math in the first post, if a player hit 2 singles, a double, a triple and a homer in 15 AB over a weekend series, your formula produces a slugging percentage of .933, versus his actual SLG of .733.

As usual, Nate hits a grand slam. But it's only worth 2 runs.... :p

I see what you did there!

:cool:

OPS is probably roughly OK. It's just a combination of double-counting and adding two fractions with different denominators.

I prefer wOBA if for no other reason than because I can say "WHOA BAH!"

Powered by vBulletin® Version 4.2.2 Copyright © 2014 vBulletin Solutions, Inc. All rights reserved.