clock menu more-arrow no yes

Filed under:

Pitching, clutch situations, and other sabermetric musings...

New, comments

Numbers, stats, baseball...these go together like wine and cheese or peanut butter and jelly.  Maybe that's a controversial view, though, since some people think it's more like putting sardines in their hot fudge sundae.  (And, while I'm talking food and numbers, how many calories do you think this meal gives you?  I have to admit I have been tempted to try it, but so far have resisted.  I wonder if IHOP afficiando Hunter Pence has tried it?)  Knowing that some people around here like to talk numbers and baseball, I will provide a smorgasbord of statistical thoughts today (Note: I haven't eaten yet, if you hadn't guessed.)

As for today's photograph...that's just to remind you that the Astros have been on the same field with the icons of our culture.  (Not.)

Astros & GIDPs

The Hardball Times has a short piece, "Escape Artist or Rally Killer," which explores double play propensity in 2009.  The analysis is based on hitters' "double play opportunities" divided by "double plays," (or the ground into double play rate).  You may or may not be surprised (I'm not) that the Astros are the absolute worst team in the majors in double play rate. The Astros are -12.9 runs worse than average (as measured by the ML average rate).  The Astros are almost twice as bad as the next worse team (Mets) in GIDP rate.  The Astros are 2 wins worse than the Phillies (the best at avoiding the DP) due soley to the team's GIDP propensity.  The Astros lost more than 1 win because they are below average in GIDP rate.  (It seemed like more than just 1 or 2 wins when you watched it over the course of the season!) The Astros had one player (Bourn) on the top ten best GIDP rate ranking, and two players (Pence and Tejada) on the top ten worst GIDP rate list.  Will the 2010 Astros rank better on the GIDP list?  Probably somewhat, but I wouldn't be surprised if the Astros continue to be worse than average.  Tejada is no longer on the team, which probably helps the GIDP rate.  But Lee, Pence, Quintero, and Keppinger are high rate GIDP guys, and they remain.  I would expect Feliz to be somewhat below average and Manzella perhaps to be better than average, but adding those two players, removing the Blum-Keppinger platoon, and dropping Tejada, should be a nice net positive on the GIDP front.  Interestingly all but one of the top ten players at avoiding the GIDP are lefthanded.  And the Astros remain predominantly RH-heavy.  If Matsui and Towles/Castro play more this year, that may help the Astros with their GIDP problem (or "situation" as I look at the photograph to the left)..

Team Management and Sabermetrics

The Houston Astros are not a team noted for sabermetric influence.  Some of us wonder if the Astros will ever join the set of teams which are more oriented toward statistical analysis.  This is an interesting article by Phil Birnbaum asking why more teams don't hire sabermetric analysts, and if they do hire them, why the statisticians don't have more influence over the team's decision.   He points out that this should make financial sense for MLB teams.  He lists six possible reasons why MLB teams' don't spend more money on sabermetric analysis.  I think he has hit on some realistic points.  And it's not just saying they are stupid.  In the real world, impediments exist to changing the way you run your business.  For a GM who has experience with his own way of constructing a ballclub, and therefore has some natural skepticism of the statisticians, it's hard to prove that the sabermetric analysis is worth the money.  Some of the insights might have a relatively small impact on team production, which (from our sabermetric viewpoint) is cost-effectivve in an absolute sense; but that positive impact probably becomes muddled and difficult for management to extract and appropriately credit as a sabermetric contribution.  For this to change, my guess is that teams have to be run by GMs who, by their background, have confidence in using quantitative analysis for baseball decision.


A number of advanced stats and related data would suggest that Felipe Paulino was unlucky last year.  x-FIP, tRA, and hit tracker are among the statistical sources we've discussed with respect to Paulino's 2009 record and possible 2010 improvement.  Then, my next thought: is it possible that clutch hitting stats might tell us something about Paulino's luck or lack thereof?  After all, some people look at a team's clutch hitting stats to explain why an offense has over or under performed relative to its overall offensive measures.  The implication, for some analysts, is that a team just got lucky or unlucky in converting base runners to runs--since clutch hitting is not viewed as a repeatable skill by many saberists.  (Though that's not a unanimous verdict.)  So, here are Paulino's clutch hitting-against stats:

RISP .357 .430 .491 .921 .463


Bases Empty .297 .347 .532 .879 .333
Men On .341 .418 .523 .941 .419     

We could look at other situations (e.g., bases loaded, two out), but the base states above are the largest samples (and the sample sizes are a problem, even with these splits).  Now I come to a question I can't answer.  If clutch hitting isn't a skill, does this mean that Paulino just got unlucky with the distribution of his hits and walks?  Or, does this instead reflect on Paulino's lack of clutch pitching skill?  Or maybe, put more delicately, is Paulino's inexperience with tense run scoring situations causing the worse results when runners are on base?  My impression of the clutch hitting statistical work is that more studies have been performed on hitting than pitching.  If we don't know enough to be certain about clutch hitting skills, we probably know even less about whether clutch pitching is a repeatable skill. 

When we think about it, a lot of factors come into play, and it's unknown how those issue interact to affect the clutch situations.  Young pitchers frequently have difficulty maintaining their mechanics when they pitch from the stretch.  Some pitchers are better at picking off runners or holding runners.  The defensive alignment changes, frequently for the worse (from the pitcher's standpoint), when runners are on base.  First baseman and third baseman may not cover the hole in those situations.  Apparently, the Astros' new pitching coach has been particularly focused on pitchers' mechanics in the stretch.  The Astros' broadcasters mentioned that Arnsberg had all the pitchers on practice fields pitching with different base runner situations recently.  For what it's worth, Paulino's OPS-against is worse with the runner on 1st (.975) and runners on 1st and 3d (1.207) then with the runner on 2d (.782) or 1st and 2d (.671) or 2d and 3d (.400).  Does this mean Paulino has mechanical problems in the stretch? Maybe. But don't read too much into it, because the sample sizes become small when the  base runner situations are sliced up that much.

If we think that clutch pitching exists, we would probably guess that Roy Oswalt fits the clutch pitcher mold.  Below, look at Wandy's and Roy's clutch hitting against stats from 2009.  Oswalt had better results with runners in scoring postion than with the bases empty.  Wandy basically has he same result whether runners are in scoring position or not.  Roy's career clutch results are similar to 2009, though the difference between RISP and bases empty isn't quite as high.  And now I ask you: is the pitcher, Oswalt, clutch or is the batter non-clutch vs. Roy....or do we know?  In previous comments, I have mentioned this very type of question as an issue I have with the studies which conclude that clutch hitting is not a repeatable skill.  We can't rule out the possibility that either hitters or pitchers (or perhaps both at the same time) could be clutch specialists. And, if that's possible, how can we take that into account in reaching a conclusion about either clutch hitting or clutch pitching?

Roy RISP .239 .339 .359 .698 .294
Roy Base Empty .274 .307 .433 .740 .317
Wandy RISP .248 .322 .382 .704 .289
Wandy Base Empty .248 .302 .402 .703 .309