Semi-Famous Baseball Geek Looking for Info on OOTP

[Raidergoo]

Originally Posted by OldFatGuy

But reading some of that stuff about DIPS, boy, I'm not exactly sure I agree with ALL of it.

It's heresy until it becomes gospel.

[OldFatGuy]

It is very interesting indeed. Good stuff, I'm a baseball nut and love to read stuff like this and learn new things too. But I'm also old and if there is ONE thing I believe (and have NO data whatsover to back it up, just years of playing/watching the game) the MOST important skill for any pitcher is control. The ability to throw strikes (and throw them in specific quadrants of the stike zone), I believe, can also influence the chances of a batter's contact resulting in contact that produces a hit or an out. Any hard hit ball is much more likely to produce a hit than a squibbler off the end of the bat because of timing or a jam job due to location. And a pitcher is able to much more frequently get that result with an 0-2 count than with a 3-1 count.

[AfWapo]

Originally Posted by OldFatGuy

It is very interesting indeed. Good stuff, I'm a baseball nut and love to read stuff like this and learn new things too. But I'm also old and if there is ONE thing I believe (and have NO data whatsover to back it up, just years of playing/watching the game) the MOST important skill for any pitcher is control. The ability to throw strikes (and throw them in specific quadrants of the stike zone), I believe, can also influence the chances of a batter's contact resulting in contact that produces a hit or an out. Any hard hit ball is much more likely to produce a hit than a squibbler off the end of the bat because of timing or a jam job due to location. And a pitcher is able to much more frequently get that result with an 0-2 count than with a 3-1 count.

I completely agree. While I do appreciate the ideas behind DIPS, and do think it is a very valuable tool, I also think it simply "sounds" like it takes things like the above for granted. Guys like Greg Maddux, Tom Glavine, Mark Mulder, and Chris Carpenter -- while they all CAN get a strikeout when they absolutely have to, they largely rely upon placement and inducing weak contact.

You said it: any hard hit ball is much more likely to produce a hit. I don't think DIPS really takes this into account. However, it may not NEED to in order to have predictive value for the vast majority of major league pitchers.

I'm not trying to downplay the significance of Mr. McCracken's contributions, but I think DIPS is far from being able to give us a complete picture of pitcher performance.

[Voros McCracken]

Originally Posted by AfWapo

I completely agree. While I do appreciate the ideas behind DIPS, and do think it is a very valuable tool, I also think it simply "sounds" like it takes things like the above for granted. Guys like Greg Maddux, Tom Glavine, Mark Mulder, and Chris Carpenter -- while they all CAN get a strikeout when they absolutely have to, they largely rely upon placement and inducing weak contact.

Actually Maddux does exceptionally well under DIPS. A little mini-test: I calculated Greg Maddux's career hits per balls in play as of right now and the number I get is .280.

As a mini-test I kept hitting the "random" button at Baseball Reference" until it popped up a pitcher who was a contemporary of Maddux. What I got was Allen McDill. McDill pitched in parts of four major league seasons, finishing with a career major league ERA of 7.79 in 34.2 innings pitched. His hits per balls in play? .277.

Greg Maddux = .280
Tom Glavine = .280
Mark Mulder = .292
Chris Carpenter = .302

Carpenter actually has a higher career hits per balls in play rate than Jose Lima. And that's really kind of the point: you use Chris Carpenter as an example of this kind of pitcher and the numbers tell a different story. Do you do that with any other important stat? Do you talk about a guy being a big power hitter and then look up the stats and see he hits 5 homers a year in full seasons? Do you say a player is perennial batting title challenger and check and see that he's a career .245 hitter?

It isn't that the skill doesn't exist, it's that it's small enough in importance to usually be thoroughly trumped by other more important skills, and that often who we think are the best in the stat turn out to be pretty unremarkable in it (and succeed in other ways like Maddux did and Carpenter does).

[RonCo]

I'm admittedly a bit wierd. As soon as I read the DIPS concept I was struck with one of those "well-duh" moments. It made total sense. Looking at the data was just a satisfying confirmation.

[Voros McCracken]

What's funny was that I was stunned as anybody else. The initial project that motivated the study in the first place was more of a "separating out fielding from pitching" sort of project. I was very surprised to see how little pitchers really differ in the stat, but after crunching the data 100 different ways it was tough to escape it.

[AfWapo]

Originally Posted by Voros McCracken

Actually Maddux does exceptionally well under DIPS. A little mini-test: I calculated Greg Maddux's career hits per balls in play as of right now and the number I get is .280.

As a mini-test I kept hitting the "random" button at Baseball Reference" until it popped up a pitcher who was a contemporary of Maddux. What I got was Allen McDill. McDill pitched in parts of four major league seasons, finishing with a career major league ERA of 7.79 in 34.2 innings pitched. His hits per balls in play? .277.

Greg Maddux = .280
Tom Glavine = .280
Mark Mulder = .292
Chris Carpenter = .302

Carpenter actually has a higher career hits per balls in play rate than Jose Lima. And that's really kind of the point: you use Chris Carpenter as an example of this kind of pitcher and the numbers tell a different story. Do you do that with any other important stat? Do you talk about a guy being a big power hitter and then look up the stats and see he hits 5 homers a year in full seasons? Do you say a player is perennial batting title challenger and check and see that he's a career .245 hitter?

It isn't that the skill doesn't exist, it's that it's small enough in importance to usually be thoroughly trumped by other more important skills, and that often who we think are the best in the stat turn out to be pretty unremarkable in it (and succeed in other ways like Maddux did and Carpenter does).

All pretty cool stuff, I admit. Just by chance, do you happen to know the #'s for Carpenter since he's been a Cardinal?

In the last paragraph of your post you say that this skill is one that is usually trumped by other "more important skills." But I guess maybe I just don't understand what "skill" BABIP shows us. Maybe it doesn't show us anything -- and it is just randomness and luck which bears out the final numbers. Maybe that's what you're saying.

It just does seem very counter-intuitive.

Looking at Maddux and Carpenter, for instance -- they seem to be very different types of pitchers. When I watch both of them pitch they both seem to do a remarkable job of inducing weak contact and producing a lot of outs from contact. But your numbers suggest that Carpenter isn't actually very good at doing this. The numbers don't seem to match my recollections of actually watching the guy pitch. It just doesn't make much sense to me.

But something else these two seem to have in common is a very good K/BB rate. This is something which Carpenter has excelled at since coming to St. Louis -- and Maddux has ALWAYS been good at it.

Perhaps this illustrates what you mean when you speak of other skills which outweigh the ability to get outs from balls hit in play. ??

Either way -- getting back to your original point in starting this thread -- I think you'll enjoy OOTP. I still haven't bought the latest version, but am very satisfied with the old version 6.5. There are problems with certain things you mention in that version -- such as very random left/right splits, and those guys who can hit 35 bombs in AAA but can't do anything in the major leagues... Perhaps the new version is better at this.

[RonCo]

There are nuances, but basically if you assume that a pitcher can control HR/BF, K/BF, and BB/BF, you're in the right ballpark. If you assume a pitcher has control of anything that happens on a ball in play, you're generally not in the right.

The nuances have to do with GB/LD/FB% and whatnot, where wrinkles exist that indicate pitcher skills exist in great enough quantity to influence the game. But bottom line, if you start with the statement that all a pitcher can really influence are strikeouts, walks, and homeruns, you're most of the way there.

This is not counter-intuitive to me at all. But I've already admited I'm weird.

[Voros McCracken]

Originally Posted by AfWapo

But something else these two seem to have in common is a very good K/BB rate. This is something which Carpenter has excelled at since coming to St. Louis -- and Maddux has ALWAYS been good at it.

Perhaps this illustrates what you mean when you speak of other skills which outweigh the ability to get outs from balls hit in play. ??

Exactly. Something like K/BB rate really does mean a lot toward how well pitcher's succeed, whereas HBIP is a minor issue. However, there's a dangerous subset of control pitchers who have good K/BB rates but high HR totals so a SO/BB isn't perfect either, but it's on the way there. Maddux at his peak of course combined an outstanding K/BB rate with low home run totals, a combination that's a recipe for multiple Cy Young Awards.

[RchW]

Originally Posted by Voros McCracken

Exactly. Something like K/BB rate really does mean a lot toward how well pitcher's succeed, whereas HBIP is a minor issue. However, there's a dangerous subset of control pitchers who have good K/BB rates but high HR totals so a SO/BB isn't perfect either, but it's on the way there. Maddux at his peak of course combined an outstanding K/BB rate with low home run totals, a combination that's a recipe for multiple Cy Young Awards.

Coincidently I'm just reading the chapter "When Does a Pitcher Earn an Earned Run" in Baseball Between the Numbers. At the end of it they discuss DIPS and some of the pitchers who apparently do limit hits on balls in play such as knuckle ball pitchers (Tim Wakefield), soft tossers (Jamie Moyer) and control pitchers like Ferguson Jenkins, who fits the good K/BB high HR totals you mention above.

Just what did someone like Jenkins do to limit HBIP?

[OldFatGuy]

Originally Posted by RchW

Just what did someone like Jenkins do to limit HBIP?

Pitch ahead in the count.

[RchW]

Originally Posted by OldFatGuy

Pitch ahead in the count.

Doh. Obvious Eh?

[OldFatGuy]

Originally Posted by RchW

Doh. Obvious Eh?

Well, to me yeah, but if I'm reading this DIPS theory correctly, no. I can't pull numbers up to show it, but I've watched/played the game pretty much my whole life and I believe, but again can't pull out the stats, that when you're ahead in the count, it keeps the batter guessing and off-balance, and thus, any contact made would be less likely to be a hit.

However, IF, and I emphasize, IF, I'm reading that DIPS stuff correctly, it pretty much says it doesn't matter what pitchers do, its just really random luck once the batter makes contact and its NOT a home run. At least that's the way it reads to me. And, as I stated above, I do agree with most of it, it makes sense to me.

After all, if a batter makes good contact and hits the ball hard, it really is just dumb luck whether its going to be a hit or not. But, at least in my opinion, if a batter doesn't make solid contact, the chances of that ball being a hit is less. And, again my opinion, I think the pitcher has a better chance of producing that type of contact by being a good control pitcher who is ahead of hitters instead of pitching behind. After all, a home run is the ultimate in contact, right??? Well, do you think more home runs have been hit 3-1 counts or 0-2 counts?? I'd be willing to be the percentage is MUCH higher in 3-1 counts.

[RonCo]

This has always been one of my favorite articles on the value of the count. It exposes a small fallacy about the value of being ahead in the count--that being that these studies usually only take into account counts where an at-bat actually ends. This is a problem because no hitter can strike out on a 2-1 count (for example), but they can and do on a 2-2 pitch. This means there are fewer ways an at-bat can end on a count with less than 2 strikes. When you remove this bias what you might find is that hitters do better when they see more pitches--regardless, to a degree--of count.

Ultimately it says that pitchers who get ahead in the count do better primarily because they increase the likelihood of a strikeout, not necessarily because the batter makes any worse contact.

http://ise.tamu.edu/people/faculty/B...0by%20count%22

[Syd Thrift]

Man... can't believe I didn't notice this before. One of the things I thought of when v6 moved to the DIPS engine was "oh boy, Voros McCracken would LOVE this game". FWIW, Markus has noted in the past that really good pitchers just didn't dominate enough in the game so there is a little bit of variability in BABIP (aside from the variability you get when you have a great defense playing in a strong pitcher's park). Also, knuckleball pitchers also get a slight boost to their BABIP. I don't think one-pitch specialists like Mariano Rivera do (IRL this could be a big part of why knuckleballers do a little better than normal pitchers), but then again I don't think the game currently generates one-pitch specialists.

As one of many asses in rec.sports.baseball from the late 90s who thought DIPS was silly until you and others explained it to me, welcome aboard (hopefully)! To my knowledge Roger Maynard does not play this game so you are safe.

[Jestre]

Hey Voros, welcome, two quick questions; are you still working for the Red Sox and did you have anything to do with bringing in Hideki Okajima

[Kelric]

Originally Posted by Jestre

did you have anything to do with bringing in Hideki Okajima

I would think that as lucky as they got with Oki, anyone on the club would like to take credit for bringing him aboard.

[OldFatGuy]

Originally Posted by RonCo

Ultimately it says that pitchers who get ahead in the count do better primarily because they increase the likelihood of a strikeout, not necessarily because the batter makes any worse contact.

http://ise.tamu.edu/people/faculty/B...0by%20count%22

Well, then if batter's "don't" make worse contact when behind in the count, I assume the probability of home runs (the ultimate in contact) is the same on 0-2, 1-2, counts as for 2-1, and 3-1 counts???? Sorry, I have dial-up and when I press some sights if takes over 20-30 seconds for me to see something I just close it, and this was the case with the sight you referenced, so I couldn't read it. But, again I'm not arguing because I don't know the stats, just from years of watching it seems to me batters have a higher probabiliy of better contact in hitter's counts, or even counts. Perhaps I am wrong. Did the article compare counts on when home runs are hit????

[Syd Thrift]

Another factor to consider, OFG, is that good pitchers are more likely to get 2-strike counts on their opponents than bad pitchers are. I'm not sure that demonstrating that there is a slightly lower chance to hit HRs on an 0-2 count compared with a 2-0 is necessarily a sign that it's actually easier, everything else taken into account, to hit a HR on the latter count.

[Syd Thrift]

Originally Posted by OldFatGuy

Well, to me yeah, but if I'm reading this DIPS theory correctly, no. I can't pull numbers up to show it, but I've watched/played the game pretty much my whole life and I believe, but again can't pull out the stats, that when you're ahead in the count, it keeps the batter guessing and off-balance, and thus, any contact made would be less likely to be a hit.

However, IF, and I emphasize, IF, I'm reading that DIPS stuff correctly, it pretty much says it doesn't matter what pitchers do, its just really random luck once the batter makes contact and its NOT a home run. At least that's the way it reads to me. And, as I stated above, I do agree with most of it, it makes sense to me.

After all, if a batter makes good contact and hits the ball hard, it really is just dumb luck whether its going to be a hit or not. But, at least in my opinion, if a batter doesn't make solid contact, the chances of that ball being a hit is less. And, again my opinion, I think the pitcher has a better chance of producing that type of contact by being a good control pitcher who is ahead of hitters instead of pitching behind. After all, a home run is the ultimate in contact, right??? Well, do you think more home runs have been hit 3-1 counts or 0-2 counts?? I'd be willing to be the percentage is MUCH higher in 3-1 counts.

Okay, here's the original post... a couple comments:

1. DIPS doesn't say that a pitcher has no control over HRs allowed. The control he has over them is less than he has over Ks and BBs, but Voros' original DIPS formula included HRA on the pitcher's side (the Aaron Sele jokes you saw in this thread were because the first essay he wrote on the subject showed that Sele was actually a better pitcher than Royals ace Jose Rosado).

2. While the core of DIPS is that *most* (not all, most) of a pitcher's BABIP is luck, not all of it is. Jim Palmer, for example, consistently had good BABIPs playing in front of Gold Glovers like Mark Belanger, Brooks Robinson, and Paul Blair. You can call that luck if you want, but it's a different kind of luck: the luck of having a great defense behind you. One-pitch specialists like Charlie Hough and Mariano Rivera also have statistically significant differences in BABIP. It's not *common*, but pitchers who can positively affect their BABIP do exist.

3. Although I was one of those who had to be converted initially, when you think about it DIPS really is not counter-intuitive. I think it was only counter-intuitive to those of us who grew up playing APBA and Strat-o-Matic baseball. Real-life pitchers have said for decades stuff along the lines of "once the ball is put into play, it's up to your fielders to do the job" or "boy, my at-em ball was really working tonight" (used by a guy who knew he didn't have his stuff but got lucky). Very rarely do you see a manager pull a pitcher after a string of seeing-eye hits. He might come out after giving up a couple of liners or a stretch of wildness, but managers have realized for a hundred years or more that the difference between a grounder in the hole and a grounder right to the shortstop is random chance, not pitcher ability.

Ironically, I think it was largely DIPS that caused a lot of the stathead community to go back and rethink some of those bromides touted by baseball insiders that before then had been dismissed as foolishness. DIPS angered some of the baseball elite because change always pisses someone off, but IMO by and large it shook the stathead community a lot harder than it shook the actual baseball community. Baseball Prospectus, for example, has done a... 90 degree turn, perhaps? from their saucy, anti-establishment tone of the late 90s to a "we aren't sure how these guys are successful, but instead of just dismissing their success as luck we're going to try and figure it out" attitude of today.