Morris or Bumgarner?

Can anyone who values OPS+ as a measure of a hitter, actually tell me how OPS+ is calculated -- so that someone like me can even try to evaluate whether it tells me anything meaningful about the value of a player? I don't mean a written conceptual explanation, the way I have gotten in the past when I have asked this, but an actual mathematical formula with a demonstration for any player.

I'm particularly interested in the "+" part of the calculation. ... I can sort of conceptualize how you would adjust for the league (although, I would love to see the actual adjustment you made, done numerically). But the thing I am really interested in. ... I challenge anyone who throws around these numbers to demonstrate how to adjust simple divisible numbers by their "ballpark factor." Take any player. I'll throw out Robinson Cano. Can someone who actually puts any value on an "OPS+" number as a measure, please tell me what his career OPS+ number is, and then show how you calculated that number -- as in, give me the actual variables that went into the calculation, including the numbers that went into calculating his ballpark adjustment -- i.e. -- show me the actual math, please? I'd love to try to even evaluate the number as measure of something, but I find it impossible, because I can't really understand what went into it. Maybe that is my shortcoming. Does anyone understand it well enough to explain?

I'll state up front that I am always dubious about measures that make "adjustments" to simple divisible numbers, and then people who try to tell me they are meaningful for evaluation purposes. The actual adjustments you chose to make (which are subjective), as opposed to the ones you didn't choose to make (again, subjectively), are VERY meaningful -- and depending on the adjustments you chose to make, can make any particular player look all world or like a scrub. And then the methodology you used to make that adjustment is really meaningful, too. My piddling understanding (and perhaps my understanding is wrong) of OPS+, for example, is that the park adjustment number (at least the one based on what I think the Total Baseball guys invented) used a formula based on a subjectively chosen three-year average. But obviously, if a player changes parks or it is a stadium that hadn't been around for three years, that isn't possible, so it then gets changed to a one or two year average -- immediately making it into something that isn't apples to apples, even before they get into the mind-numbing calculation I have seen to come up with that ballpark adjustment factor. Maybe I am wrong about that. But I have trouble knowing for sure, because I see these numbers, but I can never get good explanations of where they come from.

Can anyone actually SHOW me an example of how you would calculate the thing?

 

I don't find it all that simple. Your park factor adjustment -- if you were to do it the simple way you suggested -- would be skewed right off the bat. I am sort of racking my brain for why that isn't apple for apples and my gut tells me there are multiple reasons I am not thinking of. ... But the big thing I came up with is. ... If you play for a good team that scores way more runs than everyone else in the league -- and you are ahead a lot late in games -- you are not going to play the bottom half of the 9th inning in many home games. But you will on the road. So many of that players' home games are only going to include 8 innings of the superior team's (in terms of run-producing ability, which is how you are subjectively measuring how friendly his home park is) runs, while all of your away games are going to include 9 innings of that team's ability to produce runs. I hope that makes sense.

Also, in the kind of attempt at an adjustment that you gave me, the park factor you told me you'd compute may (or may not, because gain you are assuming runs scored in a stadium are a reasonable way to adjust the number. ...) tell you about the boost a player gets at his own stadium. But your calculation multiplied your adjustment factor across all of his OBP and SLG performance -- from both his home and away games. Shouldn't that adjustment just be to his home game OBP and SLG numbers, and if you did that "adjustment", how would you get around the the fact that not every player plays in the same away ballparks the same percentage of the time?

 

Sorry. Maybe I am not following. Can you explain that?

The formula you gave me was 100 times an on base percentage that is divided by a league average on base percentage. .... plus a slugging percentage that is divided by a league average slugging percentage. ... all multiplied by a factor you created by dividing the total runs scored in his home ballpark by the total runs scored on the road. I have that right?

I don't want to bog this down, so I won't address how that league adjustment will effect players' stats differently (because teams don't all play the same exact schedule).

What I really don't get is that multiplier you created and why anyone would think that is a meaningful adjustment. It's not so much the number itself, but the way you used it. You said the simple way you'd create your adjustment would be to divide total runs in home games by runs in away games (i.e. divide home by away; 81 games of each).

What you did will give you something that tells you how the number of total runs scored in games on the road compared to the total number of runs scored in games played at home. Simple enough. I don't know how meaningful that is as an adjustment to someone's OPS and SLG (that is a separate question), but you certainly can use that as a multiplier of some sort. For example, you might find that a players' games on the road produced 7 percent more runs than games played at home. So the division you did would have given you a multiplier of .93.

I am having two problems with what you then did. I am assuming you are creating this adjustment to tell me that you are adjusting for how friendly or unfriendly his home ballpark is (relative to the other places he played) to his OBS and SLG numbers, right? But in the formula you gave me, you then multiplied ALL of his numbers by your multiplier -- not just the numbers he produced on the road. i.e. -- you just discounted his home stats by 7 percent, AS WELL as his away stats. So I am confused.

Secondly, if I am right about why you are doing this, shouldn't you have to create individual multipliers (or divisors) for EACH ballpark the player played in, and then weight his already league-adjusted OBP and already league-adjusted SLG by the actual stats produced in each of those ballparks?

None of that even addresses how meaningful these adjustments are, or the other issues I'd point out about why your "adjustment" may not be treating home and away games as apples to apples propositions when you do it that way. ... but hopefully it shows you why I don't find it very straightforward of simple.