-
notsellingjeans Great research, Sal. An interesting and important question to study.
I'm excited to see the answer next week to your final question: "...we should check if certain types of hitters do show significant differences from league average in the length of their plate appearances. We'll get to that next time."
One guy I want to point you to, if you get a chance to analyze him:
Jorge Posada's statline has drastically changed over his career. He's gone from a 150K, 100 walk player (2000) to a 90K, 60BB player for the third year in a row. Interestingly, this year he's finally matched that career year of 2000 (.417 OB, .517 SLG), and yet he's done it an ENTIRELY different way - by cutting down his K's dramatically and raising his average 50 points.
No doubt we'd find he's swinging at pitches earlier in the count "nowadays" (last few years) than he was in 2000-02/03. As you and I have discussed with Dave re: Youkilis '06 vs '07, doing so would serve to a.) lower his walks b.) lower his k's c.) raise his average, by virtue of lowering his k's and thereby allowing him to show BA improvement without raising his BABIP.
Which version of Posada is the better hitter? Can we even come close to universalizing an optimal approach for all hitters? Is being aggressive early in the count more advantageous for crappy hitters, or for great ones? Power hitters, or slap hitters?
Intuitively, I'd say that if a hitter knows he doesn't hit sliders/curveballs well, or isn't a "two-strike hitter", or is a horrible guesser, he NEEDS to be more aggressive than the next guy earlier in the count.
-
salb918 Good questions, all of them. I guess my only answer is: I've got a lot of work to do.
I plan to break it down by hitter types, although any suggestions on "type" of hitter would be good. Doing it for individual batters will be very difficult given the small sample size. But it woud be awesome if we could, as you point out, analyze changes in approach from hitter to hitter.
The most important part of this kind of analysis is measuring against a baseline, and I hope that by showing the league-average numbers I've done that.
-
-
RedsManRick Perhaps I'm misunderstanding, but what causes pitch count variations between the control artists and the average pitcher is not pitch/event variation, but event/PA variation.
That is, it's not that the pitch count from an identifcal 6 IP, 6 H, 2 BB, 4 K will vary, it's that the line for the control artist will be 6 IP, 7 H, 0 BB, 4 K. And as we can see from the P/event data, trading a BB for a ball in play saves you 2.5 pitches and gives you a 70% chance for an out which saves you even more pitches.
It's interesting to note that Jake Peavy, he of the 3.72 K/BB ratio, has yet to go longer than 7 IP and 112 pitches this year.
Also interesting, if a pitcher was to throw a 0 BB, 27 K perfect game, it would take him on average 130 pitches. If a pitcher were to throw a 0 BB, 0 K complete game, allowing a ball in play every AB with 30% falling for hits, he would average 117 pitches, not accounting for double plays.
-
salb918 Perhaps I'm misunderstanding, but what causes pitch count variations between the control artists and the average pitcher is not pitch/event variation, but event/PA variation.
Exactly. Maybe I wasn't clear, but that is exactly the point of the conclusion, "If they last deep into games, it is because they turn plate appearances into outs, not that they end plate appearances quickly."
That is, it's not that the pitch count from an identifcal 6 IP, 6 H, 2 BB, 4 K will vary, it's that the line for the control artist will be 6 IP, 7 H, 0 BB, 4 K.
That's correct; I used identical lines to illustrate the point that the pitch count is a function of overall line and not of the "type" of pitcher. But you are right in saying that the the control freak is more likely to post the latter line than the former.
Re: your point about Peavy and perfect games, it would be interesting to see if so-called "workhorses" accumulate innings because they conserve their pitches (by getting lots of outs and not walking guys) or because they actually throw more pitches overall.
-
-
GuyM Sal: Why not use % of pitches thrown that are strikes as your measure of control of the strikezone? Seems like a good measure, with a general range of 58% (Rueter) to 69% (Schilling). Just eyeballing the ranking of pitchers, it looks like a pretty good predictor of pitches/PA. And pitchers who improved their control over their career, raising their strike%, also appear to reduce pitches per PA. For example, Randy Johnson was in the 58%-63% range through 1997, and averaged over 4.0 P/PA. Since 1998 his strike% has been consistently in the 65-69% range, and he's thrown more like 3.8 P/PA. It would be interesting to see if this pattern holds more generally.
While I'm sure both K/BB and BB/9 tend to correlate with strike%, I'd guess that BB/9 is a better proxy. Kazmir, for example, is only league-average in strike%.
-
salb918 With fangraphs and b-ref, these numbers are available, but not easily integrable (by me) into my database. But the point is well taken - rather than use proxies such as K/BB and BB/9, it would be better to simply use the actual strike %. I'll definitely put it on the to-do list.-
thrower25usr After reading your article earlier, I decided to try and see what happened if I used strike percentage. So far, I have 40 pitchers so it is still early, but it is beginning to look like the higher the strike percentage, the lower the p/pa. I also added a comparison of strike percentage and k/bb and there is an extremely obvious correlation.-
salb918 Remember that we expect to see differences in p/pa for different classes of pitchers. What we're interested in is p/pa broken down by result (K, BB, BIP out, BIP hit are the most important categories).
-
-
-
-
tangotiger Also note that the reason the high K/BB pitchers have more pitches per walk is that they throw strikes! That is, if I said: given that a walk was thrown, who was more likely to throw a 4-2 walk: Pedro or Jose Lima? Given that a walk was thrown, who was more likely to throw a 4-0 walk?
-
salb918 Indeed. Good point; that may apply to the control freak group as well.
-
-
matts I would guess that the reason that "control freaks" throw fewer pitches per ball in play in the data is not that hitters feel compelled to swing earlier in the count. It is probably because hitters are more likely to swing at balls over the plate, these pitchers will put more pitches over the plate earlier in the count, and therefore those pitches are more likely to be hit in play earlier in the count.
I don't think this is evidence that hitters change their approach from pitcher to pitcher. They probably have the same approach-- swing only if it's in the strike zone.
-
David Gassko
Sal, that would make for a fascinating analysis. Do batters swing on earlier strikes in the count for any specific group of pitcher? That is, look only at strikes and see how many strikes a pitcher has to throw before the batter swings for the first time in the plate appearances. Is that number smaller for some group of pitchers? I would imagine it would be for high-strikeout (especially if combined with low walk totals) pitchers like Johan Santana -- if a batter sees a pitch he thinks he can hit, he's going to be swinging at it, whereas with a worse pitcher, he might wait for something he thinks he can crush. -
salb918 Perhaps "compel" is the wrong word, but your explanation is basically what I was shooting for. The question is: can pitchers conserve pitches based on the way they pitch?
While your second paragraph makes intuitive sense, the groups we've looked at so far do not appear to have a lower P/PA for PA resulting in BIP. I guess that's what's so fascinating, that we haven't found anything yet.
-
In search of efficient pitchers
The Hardball Times —
... , broken down by result (walk, strikeout, hit, or out). The last article in this series looked at two different types of pitchers to understand how strike-throwing correlated with pitching efficiency. We left off by deciding to look at different classes of hitters. I still intend to do that, but I want to first wrap up some loose ends with pitchers. Most of the feedback (on ...
