Why OPS+ matters

If you’ve been reading me for a while, you know that OPS+ is my go-to stat for a quick player evaluation. Here’s why.

Here are some actual player stats:

HR    BA    OBP    SLG    OPS
22   .259  .296   .452   .749
22   .261  .321   .428   .749
22   .250  .329   .416   .745
22   .249  .338   .411   .753

These are 4 seasons by 4 different guys in 4 different years. But they are quite similar, right?

Well, let’s take a look at 4 other seasons:

Player           Year   OPS+   PA   R   RBI
Craig Pauqette   1996    87    462  61  67
Aubrey Huff      2005    99    636  70  92
John Mayberry    1978   108    587  51  70
Boog Powell      1968   127    634  60  85

Big differences here.

Huff and Mayberry had relatively league-average years. Pauquette was well below-average, and Powell had a very nice season.

As you’ve probably guessed, these are the same 4 players as the 1st list above. They all had 22 homers and all had raw OPS right around .750. It’s just so damned difficult to judge raw numbers without some sort of context. OPS+ does a really nice job of smoothing over the differences in ballparks and offensive eras.

0 0 votes

Article Rating

46 Comments

Oldest

Newest Most Voted

Inline Feedbacks

View all comments

ERolfPleiss

12 years ago

Interesting that the players respective OBPs ended up being ordered the same as their OPS+. Which I might expect as OBP is more valuable than SLG in terms of player performance? Maybe I got that wrong.
Anyways, neat little piece here.

GrandyMan

12 years ago

Reply to ERolfPleiss

Interesting observation, E, but I’m afraid it’s just a coincedence. OPS+ is calculated by comparing OBP and SLG to league averages for that year, and then adding in park factors. When this is done, OBP and SLG are weighted equally, so that neither takes precedence over the other. The phenomenon you see here is because of differences in years and park factors. Putting up a .750 OPS at the cavernous Oakland County Coliseum in 1968, a historically bad year for hitters, would have been quite a feat. On the contrary, putting up a .750 OPS at Fenway Park anytime during… Read more »

Bryan

12 years ago

Reply to GrandyMan

Actually, GrandyMan, there might be more to ERolf’s point than just coincidence. The players are initially listed in ascending order by OBP and descending order by SLG. If I may oversimplify park and era effects, they tend to have a small impact on OBP and a bigger impact on SLG (with several exceptions, of course). If we read a higher OBP as “better offensive player” and higher SLG as “beneficiary of hitter’s park/era”, the adjustments that add the + may scale Paquette’s and Huff’s SLG back to where they belong, while keeping Powell’s and Mayberry’s OBP intact, thereby rewarding the… Read more »

GrandyMan

12 years ago

Reply to Bryan

Never stopped to think about that, Bryan. You’ve made an excellent point.

MrDave

12 years ago

Always loved how the player not named Barry Bonds who had the highest OPS+ for a season was Fred Dunlap back in 1884. It took 107 years for that record to be broken.

e pluribus munu

12 years ago

Reply to MrDave

Dunlap was greatly aided by the miracle of the Union Association’s resurrection to major league status 85 years after its untimely and asynchronous dissolution. For Bonds it was the miracle of modern medicine.

Dunlap’s nickname was “Sure Shot” – yet another bond between them!

Admin

Adam Darowski

12 years ago

My go-to is WAR batting runs. OPS+ is a rate stat, so here’s how they translate with everyone evened out to 700 PA:

Player           Year  OPS+     PA  Rbat  Rbat/700
Craig Pauqette   1996    87    462    -7       -11
Aubrey Huff      2005    99    636    -5        -6
John Mayberry    1978   108    587     1         1
Boog Powell      1968   127    634    16        18

Tmckelv

12 years ago

Reply to Adam Darowski

It is interesting to me that Mayberry was closer to “average” than Huff for Rbat/700. But then Huff was closer to “average” than Mayberry for OPS+. (Assuming 0 is “average” for Rbat/700, and 100 is “average” for OPS+) To me it makes a difference if a player is above, below, or just average. After thinking about it for a while AND not knowing how Rbat is calculated, I am realizing that 3 players probably could all have 108 OPS+ with 3 different Rbat values. So I guess you need to look at numerous stats (which we have) to evaluate average,… Read more »

Editor

John Autin

12 years ago

Reply to Adam Darowski

Adam, which is your A-number-one — Rbat, or Rbat/700?

Adam Darowski

12 years ago

Reply to John Autin

Oh hey, John. Missed this until now. I’d call #1 Rbat. I have to factor in PAs as well, but Rbat is what I look at first when determining how good an offensive performance was.

12 years ago

I tend to use it as a quick and dirty as well. That being said, OPS+ has no clear meaning or interpretation. I’m not sure why B-R continues to use it whereas Fangraphs transitioned to the superior wRC+ in 2009. For those unfamiliar with it, wRC stands for weighted runs created and wRC+ is wRC compared to league average. Average is 100 and each point above or below 100 is a percentage point above or below league average. So a wRC+ of 120 means the player created 20% more runs than league average. wRC+ is also adjusted for park and… Read more »

bstar

12 years ago

Reply to Ed

I’m a little dismayed also that B-Ref didn’t include wRC+ over OPS+ in their recent changes to their WAR calculation. They’re still way behind Fangraphs in that regard, although I think fWAR for pitchers is bad, especially for relievers. There are two reasons that I continue to use OPS+: 1. B-Ref uses it and not wRC+. Kind of hard to do a search on the P-I involving wRC+ when it’s not available anywhere on the site. 2. Almost amazingly, there usually isn’t a big difference between OPS+ and wRC+. It’s very often less than 5%. Perhaps this is why Sean… Read more »

12 years ago

Reply to bstar

Bstar – I agree that OPS+ gets used a lot cause it’s “there” and easy to access. It’s why I use it. And Baseball Reference is simply 100x better than fangraphs for extracting data. Still, I think people should be clear on the limitations of the data which are at least 4 in my mind: 1) SLG% and OBP are weighted equally when we know they shouldn’t be. I think Tom Tango said it should be 1.2 for OBP and 0.8 for SLG%. 2) SLG% itself is a problematic statistic. It assumes that a double is worth twice as much… Read more »

bstar

12 years ago

Reply to Ed

You touched on a lot of important things, Ed, and I strongly agree with everything you’ve pointed out. I would add the following to your 4 points: 1) Actually, the proper weighting of OBP and SLG, according to Tango and Fangraphs, is 1.8*OBP + SLG=OPS(fangraphs) 1.7*OBP + SLG=OPS(Tango) Here’s an article from Tango’s blog that argues why this is the proper weight. I wouldn’t bother getting bogged down with all the math: http://www.insidethebook.com/ee/index.php/site/comments/why_does_17obpslg_make_sense/ 2) Not only does SLG improperly weight hits, it also excludes walks entirely. 3) OBP not only has the problem you listed, but it also improperly weights… Read more »

12 years ago

Reply to bstar

Thanks for the comments/additions. A few followups:

1) I wasn’t clear. The weights of 1.2 and 0.8 apply to the calculation of OPS+. Tango says using these weights would greatly improve OPS+.

3) I’ll have to take your word for that cause my brain doesn’t want to think that hard right now. 🙂

4a) Tango describes the general correspondence as “luck”.

4b) Another quote from Tango: “As you can see, OPS+ is biased toward power hitters, notably HR, and biased against guys who walk.”

Dr. Doom

12 years ago

Reply to bstar

Yup. You sure got it right, bstar. Usually, they’re about the same. Actually, I’m surprised you’ve seen it off by as much as 5% (though maybe with some of the bigger seasons, it is). Since most guys hang right around average, it’s usually the same, or 1-3 points off, if at all. Besides stats, though, different park factors MAY be influencing the differences, as well – so they may be even CLOSER than it looks! As for the reason b-ref doesn’t switch? I’m pretty sure it’s because OPS+ is one of Sean’s own inventions, so he has a personal interest… Read more »

bstar

12 years ago

Reply to Dr. Doom

Gotta agree with your comments, Dr. D. Babe Ruth’s career OPS+ is 4.5% higher than his wRC+, while Bonds’ is 4% higher. I think you’re right about the big seasons showing the most difference, and those numbers are why I felt confident enough to throw out 5% as about the highest difference you’re going to find. I also agree about ERA-. When a pitcher has an ERA+ of 130(and since ERA not lgERA is in the denominator), the only thing we can say about 130 is that the league average ERA is 30% worse than this pitcher’s ERA. But ERA-… Read more »

bstar

12 years ago

Reply to Dr. Doom

Doom, in the article Ed linked to @13 Tango says that Pete Palmer invented OPS+.

Author

Andy

12 years ago

Reply to Ed

Ed, I really appreciate your description of wRC+. My problem is that the Fangraphs site just blows. For most websites, there is a certain learning curve to figure out where stuff is, now to find it, read it, parse it, etc. (Remember when B-R revamped its layout about 3 years ago and moved all the stats around? It was really shocking at first but took only about a week to get used to). But Fangraphs I never get used to. Stuff is so damned hard to find. Explanations are hard to find. Data is hard to parse. Great content, god… Read more »

12 years ago

Reply to Andy

Oh I agree Andy. As I noted in comment #13, Baseball Reference is 100x more user friendly than Fangraphs. My hope is that Sean will start including wRC+ in Baseball Reference. I’m not really sure what the barrier is. I’m sure he knows wRC+ exists and that it’s superior to OPS+. I doubt there are any proprietary issues re: wRC+. So I’m a bit stumped.

Paul E

12 years ago

Reply to Ed

Ed:
Is there anything wrong with using baseball-reference’s RC/27 and dividing it by baseball-reference’s AIR to get a better picture of a player’s producticn within the scope of all environments?

BTW, I still believe dWar stinks. Darwin Barney is currently the 6th best position player by WAR (4.4) strengthened entirely by a dWAR of 3.3…Perhaps compared to peer NL 2B he is a superior defensive player, but that dWAR value is so over-the-top, it’s ridiculous, absurd, and strains all level and manner of credulity

12 years ago

Reply to Paul E

Paul E: No idea, not my area of expertise but one of the other really smart people on this site might be able to answer your question. Regardless though it would be a fair amount of work if you had to do it on a regular basis.

Yeah dWAR is probably still a work in progress. BTW, you may know this already but dWAR and oWAR both contain the position adjustment which is why you can no longer add them together to get total WAR.

nightfly

12 years ago

Reply to Ed

That drives me crazy. It would be much more intuitive to strip the positional adjustment out of BOTH oWAR and dWAR…. say, something like this for a player with six WAR in one year: oWAR: 3.0 dWAR: 1.0 pos: 2.0 (translated from runs) Rather a good season. It’s very easy to add it all up and know where the player stands relative to his peers. You can say to yourself, “Self, would I like my team to keep this six WAR player, or take the four WAR player at a discount? Because, Self, that four WAR guy got 2.5 oWAR… Read more »

Admin

Adam Darowski

12 years ago

Reply to nightfly

If you break it down to 3.0 for offense, 1.0 for defense, and 2.0 for position, you might as well just use the run components. They already break it down for you. They never should have called it oWAR and dWAR. They should have had two numbers: 1. WAR 2. diWAR, or Defense Indpendent WAR. This is a flavor of WAR for those who don’t trust the TZ, UZR, or DRS numbers. It keeps the positional value but removes the +/- defensive numbers. This is what we know now as oWAR. FWIW, I never ever ever look at oWAR or… Read more »

12 years ago

Reply to nightfly

Adam – When BR originally retooled their WAR calculations this year, they renamed oWAR as ndWAR (no defense WAR). People found that confusing, so later the same day, they renamed it afWAR (Average Fielding WAR). Apparently this was confusing as well because a few days later they went back to calling it oWAR. I realize this isn’t exactly what you were suggesting since they never dropped dWAR, but they did try renaming oWAR.

tag

12 years ago

Reply to Paul E

Paul E, I personally think defensive metrics, while not quite jokes, are so much in their infancy as to be virtually useless. The Barney rating is only one reason why. In fact, I’ve been watching Cub games to get an idea of why Barney is rated so highly, and one thing has become clear. Barney is a very good fielder; he could probably become the equal of Sandberg defensively if he hits enough to stay in the league. But Sveum or the Cub defensive positioners are exceptional. They make it easy for Barney. It’s almost uncanny how well positioned he… Read more »

Artie Z.

12 years ago

Reply to tag

tag – but positioning, whether by the team or by the player, is an immensely useful skill. I don’t watch many Cubs games, but if the team’s defensive positioning is that good and Castro is NOT listening, doesn’t that signal that he is in some way hurting the club and costing them wins? And doesn’t Barney, by listening to their advice, help the team win? Positioning (and a strong arm) is what made Cal Ripken into the defensive shortstop that he was. Ripken didn’t have the flash of someone like Omar Vizquel, but he was a great defensive SS because… Read more »

Admin

Adam Darowski

12 years ago

Reply to tag

I personally think defensive metrics, while not quite jokes, are so much in their infancy as to be virtually useless. I dunno. These are the all time leaders in defensive runs: 1. Brooks Robinson+ (23) 293 R 2. Andruw Jones (17, 35) 243 R 3. Mark Belanger (18) 241 R 4. Ozzie Smith+ (19) 239 R 5. Roberto Clemente+ (18) 205 R 6. Barry Bonds (22) 191 L 7. Carl Yastrzemski+ (23) 184 R 8. Cal Ripken+ (21) 181 R 9. Buddy Bell (18) 174 R Paul Blair (17) 174 R Looks pretty good to me, no? I think you… Read more »

bstar

12 years ago

Reply to Adam Darowski

That table you displayed above is why I ultimately think these metrics are doing a credible job of calculating defense. Doesn’t that list pass the eye test with flying colors? Aren’t our current defensive metrics, while still flawed, way better than assuming everyone is a league-average fielder? Haven’t we made improvements to catcher defense just in the last year by implementing pitch framing and pitch blocking into the equation?

Hopefully first baseman defense will be next on the list of improvements.

e pluribus munu

12 years ago

Reply to tag

Piling on a bit, tag, since Artie and Adam have already made the essential points: If we followed your line of reasoning, wouldn’t we have to assess pitching success relative to catchers’ abilities to call a good game? I’ve always seen this as a huge gap in a primary category of assessment, but one we really will never have any way to measure outside of occasional anecdotal comments from players. There are many areas of the game where I think we have some confusion about what we value: outcomes or skills (perhaps “earned accomplishments” would be better, but too long)… Read more »

tag

12 years ago

Reply to tag

My imprecise wording is doubtless the reason for it, but I think there’s been some slight misapprehension of what I meant. I think defensive positioning is enormously useful and is, for instance, what is responsible for Barney’s high rating, along with (probably) the small sample size. I’m saying that Barney definitely deserves his high defensive rating, but it’s not necessarily because he has unheard-of range / otherworldly skills. It’s because he’s almost uncannily in the right place at the right time, either due to Sveum / the Cub staff’s positioning statistics or Barney’s own study and instincts. And there is… Read more »

bstar

12 years ago

Reply to tag

tag, one of the reasons you see blips in Brooks Robinson’s career fielding numbers is because they take longer than HR to stabilize. And by stabilize I don’t mean in the sense that we can make a general statement about a player’s abilities, I mean stabilize in the sense that we can take some sort of meaning from it at all. It’s well known that Fangraphs’ UZR takes 3 years to stabilize; I’m not sure about B-Ref’s DRS. So consider the possibility that looking at Brooks’ one-year d numbers is like looking at 1/3 of a season’s worth of HR… Read more »

bstar

12 years ago

In fact, raw OPS has absolutely nothing to do with OPS+. It is not used in the calculation whatsoever. The four variables that determine OPS+ are OBP, lgOBP, SLG, and lgSLG, where: lgOBP = the OBP a league avg player would have in the same ballpark lgSLG = the SLG% a league avg player would have in the same ballpark A lot of people, including a lot of well-respected authors, make the mistake of saying that when a player has an OPS+ of 150, he has a raw OPS that is 50% better than league average. This is wrong. It… Read more »

bstar

12 years ago

Reply to bstar

You can find lgOBP, lgSLG, and lgOPS for individual players by clicking on the More Stats tab next to Standard Batting. From there, scroll down to Advanced Batting and you can calculate OPS+ yourself to see the individual OBP and SLG components of the final number.

bstar

12 years ago

Reply to bstar

Amendment to comment #5:

A lot of people think OPS+ = 100 x (OPS/lgOPS),

instead of a lot of people think OPS+ = 100 x (OPS-lgOPS)/lgOPS

I wish we could edit our posts.

Hartvig

12 years ago

OPS+ is my go to stat was well. That said, I do think that it tends to undervalue leadoff hitters, particularly if they’re high-percentage, prolific basestealers (Raines, Henderson, Molitor and a few others) so like almost any stat it does require a little context.

Timmy Pea

12 years ago

I think something special is happening this year, and that is Juan Pierre might lead the NL in stolen bases. Juan is awesome as we all know but he has been playing only part time for the great Philadelphia Phillies this year. Despite that he has a real chance at leading the league in SB’s. I think he has been helped by Victorino being shipped to LA.

Editor

birtelcom

12 years ago

I put every hitter with at least 502 PAs in 2011 in a spreadsheet with their OPS+ and their wRC+. In 2011, the correlation (using Excel’s CORREL function) between the two stats for this group of hitters was 98.7%; an extremely high correlation. I suspect most of the remaining difference is explained simply by the fact that RC includes stolen base value and OPS doesn’t. In percentage terms the guy with the largest 2011 advantage of OPS+ over wRC+ was Miguel Olivo and the guy with the biggest advantage of wRC+ over OPS+ was Brett Gardner. I think that confirms… Read more »

kds

12 years ago

Reply to birtelcom

To me, wRC+ is much more intuitive than OPS+. What is RC? It is runs created,an estimate of the runs the particular batter we are looking at would be worth when inserted into an average team. Divide by PA and we have a rate stat. At today’s scoring levels, an average player who never missed a game would create about 80 runs in an average park situation. A really good hitter will create 120 runs. A great one maybe 150. wRC+ is a rate stat that is closely related to runs, and runs, (and wins) are central to sabermetrics. OPS+… Read more »

birtelcom

12 years ago

Reply to kds

At some level, this debate becomes more a matter of personal esthetic judgment. To me, once you start multiplying one number by .34, another by .72, another by 1.18, etc. to get your stat, you have slipped out of the realm of the pleasingly intuitive. Not that there is anything wrong with complicated fractional calculations — I love the WAR family of uber-stats and nothing is more complicated than they are. I just find OPS elegant in its use baseball’s own unfiltered numbers — plate apps, times on base, extra bases — to produce results that happen to come out… Read more »

Drew

12 years ago

What’s the best way to introduce something like this to people? Even the most basic, widely used, established, simple stuff like OPS+ seem to go over the heads of most baseball fans I talk to (aside from my wife, she gets it).

Author

Andy

12 years ago

Reply to Drew

I make sure people understand that OPS is OBP + SLG, and then I say it’s normalized to their ballpark and league average, and that over 100 is better.

I also call it “ops” as opposed to “oh pee ess”. I think the latter has too many syllables and causes confusion for a lot of people.

Zach

12 years ago

Wow! For a budding sabrmetrician like myself, this post and the many comments were quite helpful. I did not know how wRC+ was determined – it was always pretty opaque to me, and OPS+ was more helpful. Does anyone else think baseball-reference.com needs an Android app? How helpful would that be for on-the-go analysis!

Venetta

10 years ago

Assumptions have to be made about what is going on beneath the land surface between wells.
An important part of any plumbing system, the venting system pipes connect plumbing directly fixtures
to the outdoors, often through the roof. You may think you are
past the point of starting septic maintenance routine but the
truth is beginning to use the right septic tank treatment now,
may not only reverse current problems with your septic, but will prevent future issues.

Anthony

6 years ago

Sorry for not having read every post. My quick point: OPS+ (for all its referenced [minor] faults) is, in my opinion, far and away, the best single statistic to gauge a player’s offensive production, compared to his peers (although I would certainly separate close scores by considering Grounded-into double-plays [which often changes one-on and one-out, if you just strike out or fly-out or beat the throw to first base, to none-on and two out] and Stolen Base Percentage (both of which reward you for speed). My reasoning is simple – there is not another stat that starts with Ruth, Williams,… Read more »

Ess land ops UPDATED 2022

2 years ago

[…] http://www.highheatstats.com/2012/08/why-ops-matters/ […]