Better season, lower OPS+

[digglahhh]

Okay, let's take a look at the polarizing Adam Dunn's 2004 season in comparison so the 2004 season of Bobby Abreu, whose numbers please both the SABR and traditional crowds.

Dunn: .266/.388/.569 OPS: .956 OPS+ 152*

Abreu: .301/.428/.544 OPS: .971 OPS+ 149**

*Great American PF: 92
**Citizen's Bank PF: 101

Abreu had 713 PA and Dunn had 681

In 32 more PA, Abreu had 22 more hits and actually had 19 more walks.

I understand that Dunn coming out above Abreu in the OPS+ is because of the Park Factor (which I don't really agree with because G.A.BP is good place to hit, especially for Dunn- the PF one year later was 106, Citizen's Bank jumps 7 points too, which really shows how the teams who play at the park determine the park factor, but that's another debate.)

Here's my question though, why is it even close?

Abreu beats him in just about every raw stat, including OBP, BA, RC...everything except for SLG (-25) and HR (-16), though he was +13 in doubles and all that should be taken into account in the SLG anyway.

Abreu also stole 40 bases at an 88% success rate and is a better defender, this is just about the stats.

I thought about this because of the debate that arose in the Gehrig/Thomas thread about whether you could be a better hitter with lower numbers. Now Abreu was certainly better and the numbers back that up. But I don't understand why it is even close here, could this be a case of the numbers not bearing out how much better one guy was than another?

[Ubiquitous]

You already answered why its even close and that is the park factor. Adam Dunn league average with PF is .330/.421 and Abreu's is .343/.439. If you simply compare the two players against league average hitters (pitchers excluded) it comes out to Abreu 150 OPS+ and Dunn 143 OPS+.

[misterdirt]

Besides the problem of Park Factor, using OPS is not nearly as good a measure of rate of performance as using a modified OPS which weights OBP at a higher value than SLG. I use (1.8 times OBP) + SLG but good arguments have been made for multipliers ranging from 1.5 to 2. Using 1.8, Abreu's modified OPS of 1.314 is appropriately ahead of Dunn's modified OPS of 1.267, although they both had very good seasons. I never adjust these numbers for PF's for multiple reasons that I have discussed in other posts. This raw number can be converted to a normalized scale like OPS+, but if I do so I use a system based on the number of standard deviations above the mean rather than a straight percentage.

[digglahhh]

You already answered why its even close and that is the park factor. Adam Dunn league average with PF is .330/.421 and Abreu's is .343/.439. If you simply compare the two players against league average hitters (pitchers excluded) it comes out to Abreu 150 OPS+ and Dunn 143 OPS+.

Right, but just viscerally, that even seems to close. Abreu seems to be far better. Seems like a lot of difference for a mere 7 points. If Abreu had the 143 and Dunn had the 150, I think Abreu's other skills would be enough to put him ahead of Dunn as an overall player.

I have a hard time making adjustments like that seeing as how I don't subscribe to any of the total player metrics. They are flawed, IMO, but at least they are consistent. The numbers may "want" certain types of players to be better than others but they don't have biases toward individual players like I might.

[Ubiquitous]

To me 7 points seems like a lot, as to other stuff well OPS doesn't measure baserunning and defense. It simply measures what one does in the batters box.

[digglahhh]

To me 7 points seems like a lot,

It is only 15 points in raw OPS, only a 1.5% difference, and that translates to 7 points of unadjusted relative OPS...that doesn't seem like 7 points mean a whole lot when you look at it like that.

I mean once you've already distanced yourself a lot, it gets harder to distance yourself even further, I get that, but still...

[Ubiquitous]

I finally figured out a quick and easy way to use lg averages in a database (well guicker then before)and I looked at all players with at least 550 PA from 1954 to now and guess who has the lowest OPS+ in that group? Now I should mention these are not park factored.





Who did you say?


Did you guess perhaps a light hitting SS?


Well you would be wrong, the answer is Billy Ripken in 1988 had the lowest OPS in 52 seasons of play at an incredible 46 OPS+.

Ozzie Smith and Mike Caruso are close behind at 48 OPS+, Neifi Perez comes in 4th with a 50 OPS+

[lllllllllllllllllll]

Abreu had more winshares than dunn in 2004.

[RuthMayBond]

I finally figured out a quick and easy way to use lg averages in a database (well guicker then before)and I looked at all players with at least 550 PA from 1954 to now and guess who has the lowest OPS+ in that group? Now I should mention these are not park factored.





Who did you say?


Did you guess perhaps a light hitting SS?


Well you would be wrong, the answer is Billy Ripken in 1988 had the lowest OPS in 52 seasons of play at an incredible 46 OPS+.

Ozzie Smith and Mike Caruso are close behind at 48 OPS+, Neifi Perez comes in 4th with a 50 OPS+
What about Rafael Belliard '92, Mario Mendoza '79, Jim Levey '33, Hal Lanier '68, Tommy Thevenow '31, Bob Lillis ...

[BoSox Rule]

OPS+ just isn't a very good stat.

Abreu: .325 EqA
Dunn: .315 EqA

[Ubiquitous]

What about Rafael Belliard '92, Mario Mendoza '79, Jim Levey '33, Hal Lanier '68, Tommy Thevenow '31, Bob Lillis ...


Somebody didn't read what I said fully.

[RuthMayBond]

*Great American PF: 92

I understand that Dunn coming out above Abreu in the OPS+ is because of the Park Factor (which I don't really agree with because G.A.BP is good place to hit)And therein lies the problem. Who figured the park factor for that bandbox launching pad?

[Ubiquitous]

OPS+ just isn't a very good stat.

Abreu: .325 EqA
Dunn: .315 EqA


EqA is basically OPS/3 with a park factor and stolen base metric built in. If OPS+ isn't that good of a stat then neither is EqA.

.326*3 is .978. Abreu had a .971 OPS
.316*3 is .948. Dunn had a .956 OPS

[Ubiquitous]

And therein lies the problem. Who figured the park factor for that bandbox launching pad?

The first year it was open homers were flying out of there, but since then the last two years they have been flying out of there at a more moderate rate.

First year 22%
next Two: 5%

[digglahhh]

My real question is, am I the only one who doesn't think any of these stats really do justice to how much better Abreu was than Dunn?

If you agree that the gap between them was greater than the stats indicate, what is it about the figuring of the numbers that is responsible for the discrepency? Either something Abreu does is undervalued or something Dunn does was overvalued. Forget about the PFs for a second and use the raw numbers.

Can somebody post their win shares, please?

If Matt is around let's see their PCA

OPS:

Abreu: .971
Dunn .956

Eqa:

Abreu: .325
Dunn: .315*

* and if there's an SB component here, how does Abreu not blow Dunn away?


I mean, if you determine who is good and how good they are strcitly by these numbers then this is going to be an exercise in circular logic. But doesn't some of you guys' baseball instincts tell you that Abreu was far better than Dunn?

[dl4060]

Right, but just viscerally, that even seems to close. Abreu seems to be far better. Seems like a lot of difference for a mere 7 points. If Abreu had the 143 and Dunn had the 150, I think Abreu's other skills would be enough to put him ahead of Dunn as an overall player.

I have a hard time making adjustments like that seeing as how I don't subscribe to any of the total player metrics. They are flawed, IMO, but at least they are consistent. The numbers may "want" certain types of players to be better than others but they don't have biases toward individual players like I might.

7 points seems just fine to me. Dunn is a far better player than alot of people realize. Many on this board have a irrational prejudice against people who do not hit for a high average. You can hold that opinion if you want to, but it is a subjective viewpoint, unless you can back it up. Mark Mcgwire will win alot more games for you than Ichiro will, and until I see some objective evidence that they are even close that is what I will continue to believe.

[SABR Matt]

Actually...when you combine Ichiro's offense (usually produces about 8 wins a year by PCA) with his defense (2.5 to 4 wins in a year by PCA), Ichiro wins just about as many games in his average season as McGwire did in his...

But offensively...a line-up of McGwires will score more runs than a line-up of Ichiros...meaning no offense to Ichiro. :)

[dl4060]

My real question is, am I the only one who doesn't think any of these stats really do justice to how much better Abreu was than Dunn?

If you agree that the gap between them was greater than the stats indicate, what is it about the figuring of the numbers that is responsible for the discrepency? Either something Abreu does is undervalued or something Dunn does was overvalued. Forget about the PFs for a second and use the raw numbers.

Can somebody post their win shares, please?

If Matt is around let's see their PCA

OPS:

Abreu: .971
Dunn .956

Eqa:

Abreu: .325
Dunn: .315*

* and if there's an SB component here, how does Abreu not blow Dunn away?


I mean, if you determine who is good and how good they are strcitly by these numbers then this is going to be an exercise in circular logic. But doesn't some of you guys' baseball instincts tell you that Abreu was far better than Dunn?


Maybe because not everyone agrees with you on how much better Abreu was then Dunn. You also seem to be implying that everyone has the same 'baseball instincts' as you do, when that is not necessarily the case. Your argument also has a circular ring to it. Something along the lines of 'I just know Abreu was much better because of my baseball instincts therefore any argument that does not agree with mine must be wrong', I mean if you determine how good someone is just by subjective opinion is that also not an excercise in circular logic?

[SABR Matt]

Which year? if you're talking about 2005, I don't have PCA values for that year yet because certain people haven't updated their damned parks databases yet...LOL And certain other people refuse to put out 2005 PBP data (*cough*Retro*cough*)

[Ubiquitous]

My real question is, am I the only one who doesn't think any of these stats really do justice to how much better Abreu was than Dunn?

If you agree that the gap between them was greater than the stats indicate, what is it about the figuring of the numbers that is responsible for the discrepency? Either something Abreu does is undervalued or something Dunn does was overvalued. Forget about the PFs for a second and use the raw numbers.


* and if there's an SB component here, how does Abreu not blow Dunn away?


I mean, if you determine who is good and how good they are strcitly by these numbers then this is going to be an exercise in circular logic. But doesn't some of you guys' baseball instincts tell you that Abreu was far better than Dunn?

Well if you were a Phillies fan or at least a lot the ones I know you would say Abreu is a lazy bum.

I would say that you are viewing the two players through prejudiced eyes. In effect you are deciding the conclusion and now looking for the data make it appear so. I don't mean it as an insult just that is how it appears to me. To me OPS+ has it about right in terms of hitting. Abreu in non PF ops+ is clearly the better hitter and isn't really close. That is what OPS+ is telling me. It appears to me that you think OPS+ should be something like 150 to 110. I don't see that. Dunn had a good OBP and had a very powerful SLG. Dunn did have more total bases.

AS for SB, well seasonal metrics just don't value them all that much. Even if you were to look at it through PBP and run or win expectancy one extra base through stealing is not all that much.

[Ubiquitous]

Which year? if you're talking about 2005, I don't have PCA values for that year yet because certain people haven't updated their damned parks databases yet...LOL And certain other people refuse to put out 2005 PBP data (*cough*Retro*cough*)


The 2005 PBP data is out.

[digglahhh]

Matt,

I'm referring to the 2004 campaigns of each player.

DL,

I respect your viewpoint, but please don't continue to state that Win Shares or any other complex metric has objectivity, that is just not true at all- there is no such thing. What those metrics do have is consistency and stability. They evaluate all players the same way and have no particular attachment to an individual player.

I have seen no proof of the objectivity of any statistic, and there is none to be found- believing in the objectivity of numbers is closed-minded. The decisions about what to count in the formulas, what to leave out and how much leverage to grant each component of the metric is subjective in the first place. Second of all, the most elemental atoms of any metric come from rulings by official scorers, and that is more of an art than a science, believe me, I oversee the scoring of games for MLB. If you begin to think about it from a deconstructionist perspective any whiff of objectivity disappears in the blink of an eye. What about umpires? What about their calls of balls and strikes that dictate the situations for the hitters- plenty of subjectivity there. You may dismiss all these points as insignificant in the long run, but these phenomena are the gravity that allows a snowball to induce an avalanche.

The advantage of using the metrics is consistency, not objectivity, any flaws in them will be systematic and penalize players or reward them for the same things. Formulas may favor certain types of players, though they don't favor individuals for any sort of personal reasons.


And Abreu is a lazy bum most of the time. He has lots of skills on defense but only chooses to use them some of the time.

[Ubiquitous]

But I bet you wouldn't have minded him patrolling the outfield at Shea for the last 6 years.

[SABR Matt]

The 2005 PBP data is out.

Ah...finally.

I'll check and see what they have...but I don't use it directly...KJOK needs to run has program to suck in the PBP data and output defense against info for all of the years missing from his parks DB before I can rerun PCA. We have info back to at least 1959 now that should be included.

[Ubiquitous]

Actually its been out since the first week of January.

[SABR Matt]

Well then I'll have to get on KJOK's case...because he promised that as soon as the 2005 data came out...he'd be working on the DB update...LOL

[Ubiquitous]

Well if you only look for it on retrosheet you'll never find it.

[SABR Matt]

what?

Explain.

[Ubiquitous]

Technically its been out since last year but the latest version came out in January.

There are stat groups on Yahoo that have the PBP data.

[SABR Matt]

Please...link me.

Perhaps one of them has data released I can use...KJOK's been slow lately.