Comments on: 2008 PrOPS-Stars, Overperformers, and Underperformers

By: JC

JC — Tue, 15 Jul 2008 16:23:42 +0000

I do not have the raw data anymore. What is presented is what is available through THT.

By: Frank

Frank — Tue, 15 Jul 2008 16:13:30 +0000

JC–do you have league averages (DH excluded)? Among the leaders, there are several positions for which the NL props is greater than the AL props. E.g., the third place NL catcher tops the first place AL catcher. I’m wondering if there might be the beginnings of a shift back toward the NL after several years of AL domination (something like 9-10 all star games, a majority of recent WS, etc.).

It might also be interesting to see props by position (for all players not just top 3) b/c one often hears debates over players such as “so and so is below avg overall but above average for his position.”

By: MattS

MattS — Tue, 15 Jul 2008 15:19:28 +0000

I would imagine that there is probably a good deal of persistence to which players over- and underperform their PrOPS. If you try to separate individual skills that you think would cause this (pull hitting, speed, etc.), you might not blur the results with an imperfectly measured dependent variable, resulting in attenuation bias.

What about including a variable in the regression for the previous year’s {OPS-PrOPS}? Wouldn’t that capture most of the omitted variables bias? After all, if there is a characteristic that leads to over- or underperformance, it would probably be present in prevoius years.

By: Victor

Victor — Tue, 15 Jul 2008 14:56:49 +0000

I was just about to comment on the shift affecting pull hitters. I think Howard is a great example of this. How many times have you see him line it hard, but straight into the 2B who’s playing a very shallow RF? And “big, slow guys” tend to correlate with pull hitters who require shifts so that makes sense too.

By: JC

JC — Tue, 15 Jul 2008 14:03:38 +0000

I just remembered that I received an e-mail from a reader who had looked into over/underperformance of PrOPS. He found some evidence that pull hitters underperformed, and he hypothesized that this was a result of defenders being able to shift to take away more hits on balls in play than hitters who spread the ball around the field.

JC—Is there any data that suggests that pull hitters PrOPS outperform their regular OPS because the defense is better aligned as to where they will put the ball in play….I looked at the data on Hardball Times at the five seasons with the highest and lowest PrOPS-OPS values….. I only used players who are active this season so I could see what percent of the time they pull the ball..check this out…..The numbers show with the noticeable exception of Curtis Granderson, that they the guys whose OPS outperform PrOPS hit to both sides of field, whereas the other five guys are all noticable pull hitters

FIVE PLAYERS WHOSE OPS OUTPERFORMED PrOPS
Ordonez 35-33
Granderson 50-22
Holliday 37-31
Ichiro 33-31
Hanley Ramirez 38-30

FIVE PLAYERS WHOSE OPS UNDERPERFORMED PrOPS
Adam Dunn 50-20
Cust 46-26
Frank Thomas 43-25
Giambi 50-23
Aaron Boone 49-22

By: JC

JC — Tue, 15 Jul 2008 13:56:53 +0000

A quick look shows the underperformers to underpeform much more than the overperformers overperform.

This is the result of offense being down this year. It is based on past data, so there is some bias in this direction.

At an earlier rendering of PrOPS I seem to remember you suggesting that “big, slow guys” might underperform at a higher clip than they should (and with above random frequency).

This was suggested by others, but I did not find much evidence of this.

I did look at qualified batters only, but I’m not sure how many that is off the top of my head.

By: Cliff

Cliff — Tue, 15 Jul 2008 13:44:49 +0000

A quick look shows the underperformers to underpeform much more than the overperformers overperform. First place is higher. Last place (25) is higher. At spot 10 the overpeformance approximately equals the under performance at spot 25.

At an earlier rendering of PrOPS I seem to remember you suggesting that “big, slow guys” might underperform at a higher clip than they should (and with above random frequency). Just wondering how the mean, mode and median sort out.

Also I would suppose that you looked at “qualifying” batters only. If so, wondering how many that is. I would expect 180 to 200. (AL teams at 6 to 7 per team and NL teams at 5 to 6)