Is OPS a Slightly Flawed Statistic?

[Navafan29]

I thought of something. Since OPS is an increasingly popular statistic, I'll bring this up. OPS is on base percentage plus slugging percentage, as in:

OBP * 1 + SLG * 1

or just:

OBP + SLG

 

If OPS is on base percentage plus slugging percentage, and there's no multipliers to weight OBP or SLG in the equation, this is saying that OBP and SLG are exactly as valuable as each other. I find it unlikely that this is true. Perhaps the ability to get on base is about as valuable as the ability to hit for power, but these are two different stats that are calculated in different ways. There's some redundancy in adding them together, since SLG is already composed of hits, which are part of OBP; and a big chunk of OBP is hitting for power, which is a part of SLG. Does the redundancy in each cancel out exactly even when they are added together? I doubt it. I feel like there is a better equation to be made that combines OBP and SLG, other than one which says that two differently calculated stats are exactly equal in their numbers' value. OPS might be a slightly flawed statistic. I can think more in-depth about this when I am not tired, like I am now.

[Bellhorn04]

It's been argued that it should be something like 60% OBP and 40% SLG.

 

I've wondered why they don't give the average of the two instead of adding them together. It seems weird that a seemingly big number like .550 is a HORRIBLE OPS.

[mvp 78]

How is "a big chunk of OBP is hitting for power?" Power has nothing to do with OBP. If Holt goes 2-4 with 2 singles and Papi goes 2-4 with 2 homers, don't they have the same OBP?

[a700hitter]

All stats are imperfect and all stats are reflective/indicative of ability and performance (even RBI). They are all necessary supplements to good scouting and they are great for fans who can't possibly watch 30 teams regularly. As for this question as to OPS being "imperfect", I am not sure what you mean by "perfection". It is a nice little tool that is easy to use-- add 2 numbers. You don't even need it to be listed to figure it out. If you start including multipliers, it is not such an easy tool to use.

[Jack Flap]

I admit that my knowledge of sabermetrics is pretty rudimentary, but there are stats like wOBA and wRC+ that attempt to measure a player's overall offensive production more precisely. I end up using OPS a lot because it's easy to find and calculate and, these days, more widely understood.

[sk7326]

More than slightly flawed - it's actually not very good at all. But - it is a good training wheels sort of entry into better stats - if your world was RBIs, batting average and pitcher wins ... OPS at least is headed in the right direction. The flaws?

 

1. OBP and SLG have different denominators (OBP uses plate appearances, SLG uses official at-bats). They aren't additive.

2. OBP is significantly more valuable than SLG relative to scoring runs. Something like 1.7*OBP + SLG would be more accurate. And by "accurate" - we are talking about how well the stat can predict run scoring.

3. Not all outcomes are equal. SLG's use of total bases means that homeruns are 4x as valuable as a walk - which is not true from a run scoring perspective. And it ignores walks and hbp altogether.

4. OBP includes intentional walks - which is arguably not something a player should really be credited with - if you want to be totally precise.

 

wOBA is a cleaner way to get to the information (wRC+ normalizes it, but wOBA works on an OBP type scale so it should be easy enough).

Edited May 10, 2016 by sk7326

[Spudboy]

How is "a big chunk of OBP is hitting for power?" Power has nothing to do with OBP. If Holt goes 2-4 with 2 singles and Papi goes 2-4 with 2 homers, don't they have the same OBP?

 

My immediate thought. How is a walk a display of power?

[Slasher9]

OBP = (H + BB + HBP) / (AB + H + BB + HBP + SF)

SLG = total bases / AB or (1b + 2*2b + 3*3b + 4*hr) / AB

 

OPS = the 2 results above added together (OBP + SLG)

[Northern Star]

My immediate thought. How is a walk a display of power?

 

[Northern Star]

I thought of something. Since OPS is an increasingly popular statistic, I'll bring this up. OPS is on base percentage plus slugging percentage, as in:

OBP * 1 + SLG * 1

or just:

OBP + SLG

 

If OPS is on base percentage plus slugging percentage, and there's no multipliers to weight OBP or SLG in the equation, this is saying that OBP and SLG are exactly as valuable as each other. I find it unlikely that this is true. Perhaps the ability to get on base is about as valuable as the ability to hit for power, but these are two different stats that are calculated in different ways. There's some redundancy in adding them together, since SLG is already composed of hits, which are part of OBP; and a big chunk of OBP is hitting for power, which is a part of SLG. Does the redundancy in each cancel out exactly even when they are added together? I doubt it. I feel like there is a better equation to be made that combines OBP and SLG, other than one which says that two differently calculated stats are exactly equal in their numbers' value. OPS might be a slightly flawed statistic. I can think more in-depth about this when I am not tired, like I am now.

 

Isn't the answer to your question OPS+?

 

I love OPS. I don't take stats as seriously as some out there, because you can drive yourself crazy with this stuff. But as a quick guide to a player's usefulness, I like it a lot.

[Kimmi]

More than slightly flawed - it's actually not very good at all. But - it is a good training wheels sort of entry into better stats - if your world was RBIs, batting average and pitcher wins ... OPS at least is headed in the right direction. The flaws?

 

1. OBP and SLG have different denominators (OBP uses plate appearances, SLG uses official at-bats). They aren't additive.

2. OBP is significantly more valuable than SLG relative to scoring runs. Something like 1.7*OBP + SLG would be more accurate. And by "accurate" - we are talking about how well the stat can predict run scoring.

3. Not all outcomes are equal. SLG's use of total bases means that homeruns are 4x as valuable as a walk - which is not true from a run scoring perspective. And it ignores walks and hbp altogether.

4. OBP includes intentional walks - which is arguably not something a player should really be credited with - if you want to be totally precise.

 

wOBA is a cleaner way to get to the information (wRC+ normalizes it, but wOBA works on an OBP type scale so it should be easy enough).

 

The Hardball Times uses (or used to use) a stat called GPA, Gross Production Average. It is similar to OPS, but weighs OBP more heavily than SLG, 1.8 times to be exact. The formula also divides the sum by 4 to put the value on a scale similar to BA, which makes it a little easier to understand in relative terms. And it's adjusted for park factors.

 

GPA = (OBP*1.8 + SLG) / 4

 

That said, OPS has a strong correlation to runs scored, so it works reasonably well. That said, IMO, wOBA is a better stat.

[Bellhorn04]

OPS is as much as the average fan can handle. Once the stats start looking like chemistry equations, I'm outa there.

[a700hitter]

OPS is as much as the average fan can handle. Once the stats start looking like chemistry equations, I'm outa there.

You would need to watch the game with a scientific calculator.

[Navafan29]

How is "a big chunk of OBP is hitting for power?" Power has nothing to do with OBP. If Holt goes 2-4 with 2 singles and Papi goes 2-4 with 2 homers, don't they have the same OBP?

 

Oh, all I was saying is that the hits that are hitting for power (extra base hits, home runs) are part of OBP, which adds to the overlap of the two stats.

 

Really good input on this topic so far. I like statistics that I can understand the math behind, as long as it isn't a crazy formula. GPA sounds very promising.

 

Not sure if this is possible (or it might already exist), but what if there was a stat that proportioned everything perfectly according to the situation etc.? It would treat a bases empty walk as equal in value to a bases empty single, since they accomplish the same thing. A walk off sac fly would be equal to a walk off home run. Even if you don't go that far, maybe you could take SLG and merge OBP into it, to make it a "how good is every at-bat on average" stat. Have walks and hit by pitches each have a multiplier like homers are x4, etc.. I will be looking into the stats you guys mentioned here.

[Hugh2]

How is "a big chunk of OBP is hitting for power?" Power has nothing to do with OBP. If Holt goes 2-4 with 2 singles and Papi goes 2-4 with 2 homers, don't they have the same OBP?

 

SLG is more of a component of OBP than OBP is a function of SLG. Every possible way you can slug a double, single, triple or HR contributes to your OBP but not every way you get on base contributes to your SLG. Maybe a "big chunk" is a bit of an exaggeration but that statement itself would be open to interpretation...no?

[mvp 78]

SLG is more of a component of OBP than OBP is a function of SLG. Every possible way you can slug a double, single, triple or HR contributes to your OBP but not every way you get on base contributes to your SLG. Maybe a "big chunk" is a bit of an exaggeration but that statement itself would be open to interpretation...no?

 

There is no power impact on OBP. A HR counts the same as a single. Everything contributes to OBP regardless of power. Power has absolutely nothing to do with OBP.

[Hugh2]

There is no power impact on OBP. A HR counts the same as a single. Everything contributes to OBP regardless of power. Power has absolutely nothing to do with OBP.

 

I agree, but let me rephrase my original point. If you hit a double your OBP and your SLG % goes up. If you walk your SLG is unaffected. Obviously the weights are different as you said and when you combined them the end result is OPS weighs HR's 2B's and 3B's less than you 'd expect by cutting because every hit weighed in slugging is weighed equally in OBP and added back in. I'm not sure how I feel to be honest with you as I don't have a hard stance but i would entertain the ideal that SLG% is undervalued in OPS. Call it 45/55 maybe 40/60

[mvp 78]

I can't argue with that. I don't think any stat is perfect.

[Hugh2]

I can't argue with that. I don't think any stat is perfect.

 

I agree even stats like WAR an OPS are far from perfect and flawed. I do think they are useful and by themselves are more useful than other stats if you HAD to just look at one.

 

I think it's become the new "quick" reference that OPS used to be.

[moonslav59]

I have always argued that OBP is more important than SLG, and the data supports this position. Maybe something like this might be better: OBP + OBP + SLG/ 3 would be best, but if the evidence supports a 60-40 ratio is more realistic, then work it out that way: (6 x OBP) + (4 x SLG) / 10.

 

Now that we are so used to using OPS, maybe going with OBP + OBP + SLG/2 would be better.

[sk7326]

OPS is as much as the average fan can handle. Once the stats start looking like chemistry equations, I'm outa there.

 

Probably - although wOBA is not very difficult to explain at all. It's really just correcting slugging percentage.

[a700hitter]

Probably - although wOBA is not very difficult to explain at all. It's really just correcting slugging percentage.

And you can't compute that from the information they give you on the scoreboard.

[Maxbialystock]

One nutty thread because OPS does not claim to be anything more than adding to numbers (percentages) together. It is almost impossible to see OPS without seeing, right next to it, the actual OBP and slugging percentage. It is far from being the ideal representation of a hitter's worth, but it's not bad and has the virtue of simplicity.

[devildavid]

What is flawed is the use of OPS, or WAR, or any single number as an important indicator of a player's performance. They are shortcuts which don't really describe the player in any strategically meaningful way. Look at all the numbers all the time if you want a true picture, and even then you still don't have it. Look also at the player who the numbers attempt to evaluate.

[Bellhorn04]

At first I thought OPS was quite silly. Then I looked at the list of the all-time OPS leaders and I thought, OK, I guess there's some merit there after all.

[User Name]

No stat is flawless, because there are a lot of things stats don't measure. But nothing beats OPS when you want a quick look of a guy's offensive profile. wOBA and wRC+ are great, but not necessarily as accessible or easy to understand as OPS and OPS+ .

[Spudboy]

As a reminder, f*** UZR.

[User Name]

UZR is great.

[Youk Of The Nation]

As a reminder, f*** UZR.

 

Hey, UZR is the statistic that allows us Red Sox fans to take comfort in knowing the truth: That Derek Jeter did not deserve the last few Gold Gloves he won.

[devildavid]

As a reminder, f*** UZR.

 

I agree. UZR is a relative stat. It measure players against the performance of other players. It is like scaling scores on an exam. If most of the class is stupid, the top score is less indicative of how much has been learned by the top student. Fielders are rated on a scale against all other fielders in their position, so how they are rated depends upon the overall quality of fielding at that position for that season. It is based on a subjective measure: deciding what plays a fielder should have made. It takes more than one season of data to come up with a reliable number. The zones themselves are somewhat arbitrarily drawn. Too many moving parts to get a very reliable number. Just because you use math and put a number to something does not mean it is an accurate description of reality.