A danger of OPS?

[backup catchers]

In my effort to learn I'm putting this out there as an opinion. If I am far off or behind the times please excuse my ignorance.

 

OPS is one of those stats that many people look at for a quick referance look at a players performance. The problem I've seen is one that plagues many "combined" stats. Ones that combine two or more other stats. Usually those stats have a flaw in that it's components are usually treated as equal when they are combined. As Russ has shown with ops is the components aren't equals at all. That leads to false conclusions about players abilities. I use ops as an example because it's the one that gets used most. Many times it's the only stat used for an arguement for or against a player. To me I would rather see obp and slugging over ops because it tells a more complete story about the type of player. I personally think ops is the thing that makes many non stat people cringe. Quite frankly I don't believe real stat people would ever use one stat for anything. They are usually smart enough to know better. If someone has a defense for using it please educate me. I am willing to learn.

[BillScottCanRake]

As someone who doesn't evaluation players for a living, and as someone who's not nearly as die-hard as most posters here, I will be the first to admit that I tend to use simple stats more than any other tool "evaluating" players. I simply don't watch as much baseball as a lot here or have the time to devote to statistical analysis that others, and certainly professional scout, do. OPS is about as simple a stat as you can find, and it does a reasonably giving you a general idea about a player. However, I try to not form strong opinions about a player based on that single stat.

 

If I am going to make a post about a player, I try to use everything at my disposal to form an educated opinion- observations (personal and second hand) and a range of different stats. I don't believe any single stat or single observation should be used to generalize or reach conclusions, even tho I am often guilty of such at times.

 

I will often use OPS as a "gateway stat" to determin how much interest I have in a player. If someone has got an OPS that peaks my interest, I investigate further. If I see a couple of players both have an OPS of 950, I will definately check their raw statistics as well as other metrics before trying to make any sort of comparison. I'll also make a point, if I am really interested, to watch them play- partially because I appreciate watching good players do their job, and partially because I can use those observations to make any opinions I form a bit more well rounded, even if they aren't any less amature.

[dadofandrew]

OXS treats OBP and SLG as equals, and is a better stat. But, few folks can multiply .350 and .475 quickly in thier heads. I can sure add 'em together though, and the difference is minimal.

[pogokat]

I think the current danger of ops is that if you write the acronym "OPS" in your post...the next 3 posts will be about how you are a fool/jerk

[endaround]

Al, OXS does not treat OBP and SLG the same. Its basically like a geometric mean which has very different properties than an additive mean (which is similar to OPS). Look at an example where you have two familar OBP and SLG scores, .350 and .450. Gvien than You have an OPS of 800 and a OXS of .1575. Now increase the OBP by 10 points. OPS goes up to 810 and the OXS goes up to .162. Now increase the SLG by 10 points. OPS is again 810 but OXS is .161. OXS rewards values close together and given teh scale that OBP and SLG operate on, it weights OBP more heavily.

[1992casey]

OPS has a .93 correlation to runs scored (1 is perfect). There are stats that do slightly better, OXS being among them. Expanding on what Al says, these other stats are often less convenient: they're generally less available, harder to compute, or both. And in the case of Runs Created, for instance, Bill James modifies the formula practically every year.

 

That makes OPS the best of the 'mainstream' stats. (OPS probably isn't cracking 'mainstream' for the general public, but they're starting to hear of it now. It would certainly be 'mainstream' for message board audiences.)

 

If we're simply trying to choose between two players and their OPSs are reasonably far apart, I don't see a lot of reason to take a lot of time looking beyond it. If two OPSs are close, then I'd certainly want to break them down into their components. If an OPS were low, I'd generally want to look into things more?the OPS might be low, but either the OBP or SLG portion of a player's performance might be useful somewhere. If I were making out a lineup card, I'd want to break it down, too.

 

I guess what it boils down to is that when a stat ranks as the best or about the best, there isn't a lot of need to discuss its flaws every time its used. Discussion of what makes up that .07 difference between .93 and 1 can generally be confined to threads like this.

 

On the other side of the coin, if I felt I wanted to make a point using stats that generally don't stack up well (batting average or range factor, for instance), I'd certainly want to acknowledge their flaws and make a case for why I'm using them under that narrow circumstance.