Sabermetrics help/questions

[Quiet Learner]

I am a longtime reader/first time poster, and while I have basic knowledge on Sabermetrics, I'm certainly not an expert. I first started reading this board because of the poster, Tango Tiger, who wrote "The Book" which is one of the most complete stats books I have ever purchased.

Anyways, I am working on some research and I've developed a few rudimentary pitching stats to aid my research and I want to know if these can be, or have been used already. The reason I am posting here is because I used Tango's linear weights to help create some of my stats.

Essentially, I am trying to research the value of Walk Rates, Groundball Rates, Homerun rates, and Strikeout rates and how they relate to each other. More clearly, I want to show how each stat helps a pitchers overall value. I'm basing most of my work around BABIP, XBHIP%, and the linear weights assigned to the offensive categories of singles, doubles, triples, HRs, and BBs and applying those to the aforementioned rate stats. I'm using strikeouts mainly under the idea that every 10 strikeouts prevents 2.9 hits (simply preventing 10 balls in play, which would equal about 2.9 hits). Combining all these factors against league averages I find just how much "damage" a pitcher has prevented. Ultimately, based on as many component stats as I can use and applying them to the adjusted laws of BABIP and XBHIP% and then applying those stats to the linear weights.

My questions:

1) Well, if I'm making sense to anyone, has this been done before?
2) Can you apply Tango's linear weights to pitchers? I assume you can, because I'm simply applying them to the offensive categories they would give up
3) Is there any research/data you guys know of that shows that groundball pitchers can "control" extra base-hits? In my research I have found that they are the one group of pitcher's that consistently limit extra base hits on balls in play.
4) Is there any concrete data about HR/9 rates? Most of what I have found shows they flucuate a great deal, although I haven't taken any park factors into consideration, which is especially important in HR rates


Sorry for the long-winded post, and thanks in advance for any help. Please ask if you want me to clarify anything, I realize I might not have explained myself that well

[Tango Tiger]

While you can generally use LWTS for pitchers, BaseRuns would be what you are really after. LWTS presumes a certain common environment to which each stat has a linearly value-added effect, while BaseRuns's environment is the stats themselves, and therefore has a more multiplicative effect.

To the extent that you want to use LWTS for pitchers (and there's nothing really that bad about that, since you will get close enough to what you want, and it's a heckavu lot easier to manipulate), then your adjustment for K is dependent on how you want to handle all balls in play (BIP): do you want to treat it as completely a fielder event or not?

You will find that FIP might already give you what you want, in terms of LWTS for pitchers that excludes the BIP.

As for GB and SLG: most GB are singles. If you go to http://danagonistes.blogspot.com , you will see that Dan shows you the BA, SLG on GB, FB, LD, etc.

[Quiet Learner]

Thanks for replying

I thought about using FIP (and still might in a modified manner), but one of the key things I wanted to add into the measure was how groundball pitchers affect balls in play (namely doubles and triples).

My main problem seems to be how I weight strikeouts, because as of now I mainly seem to view them as strictly events that produce less balls in play (every 10 K's saves 2.9 hits). The problem is K's have more value than just preventing balls in play, because even balls in play that don't go for hits can be productive (shown simply by a .73 LWT for Sacrifice Flies). I think I need to find a way to weight K's a little more heavily.

Currently in my Fielding Independent "damage prevented" stat the pitchers that prevented the most offensive damage above league average in 2007 (god that was a mouthful) were:

1. CC Sabathia
2. Peavy
3. Webb
4. Hudson
5. Joe Blanton

The top 3 are fine, but when Joe Blanton cracks a top 5 it is really troublesome. Granted, I'm just showing the overall "damage prevented" and not "damage prevented" per 9 innings (which Blanton falls to about 8 or 9 I believe), so Blanton's inning's pitched does help him here. Still, Blanton could not have been THAT good in 2007. Sure he had a solid groundball rate, was no. 12 in HR/9 (although I've been told I should switch to HR/FB, which is harder to find), and no. 7 in BB/9, but with his below average K rate, I couldn't see how he could be placed so high. I am definitely going back to the drawing board.

Then again, maybe it's not so crazy, since Blanton ranked 13th in FIP and when you turn my stat into "per 9" Blanton ranked 9th on my list for starting pitchers. He is conceivably helped on my list by the fact he is a groundball pitcher

[Tango Tiger]

The run value of GB is identical to the run value of a FB (if you exclude HR).

[Quiet Learner]

Is that so? I hadn't research it fully, but to the extent I had researched it, Groundball pitchers were prone to give up a smaller percentage of doubles and triples on balls in play. It seemed logical to me, because there are seemingly only two places on the diamond a groundball can become an extra base-hit, whereas a flyball (or line drive for that matter) can seemingly be an extra base hit in a lot more places.

[Tango Tiger]

Yes, the GB has a lower SLG... but a higher BA.

The FB has a higher SLG... but lower BA.

As luck would have it, the run value of both is around -.10 runs per PA.

Which works out great for FIP, since that means it is not biased toward either FB or GB pitchers.

Note: all this excludes the HR

[Quiet Learner]

Sounds good and makes perfect sense. I did find that groundball pitchers did have control over their extra base hits, but I didn't think to consider batting average.

I seem to find that K's aren't valued properly in my metric, what kind of weight do you think I should put on them beyond preventing balls in play? I know they have a value beyond just limiting balls in play, but to what extent?

[Tango Tiger]

FIP has the weights as 13 for HR, 3 for walks and 2 for strikeouts. It depends on how you handle hits or not, and if you use IP or PA in the denominator.

[Bothrops Atrox]

Yes, the GB has a lower SLG... but a higher BA.

The FB has a higher SLG... but lower BA.

As luck would have it, the run value of both is around -.10 runs per PA.

Which works out great for FIP, since that means it is not biased toward either FB or GB pitchers.

Note: all this excludes the HR

I was previously under the impression that the extra double plays induced by groundball pitchers swung the pendulum slightly in favor of GB pitchers, as long as they didn't have consistantly worse than average BABIP on grounders. I guess this ins't true.