February 21, 2006
Strikeouts and Pitch Counts
I'd like to return to this article by Rich Lederer which I attempted to criticize last night. This is the line that caused me to criticize the article:
Just as striking out the side in order is preferred over getting all three outs via the K regardless of the number of batters faced, a pitcher who strikes out hitters on three pitches is more effective than those who take five or six to get the job done.
I got the feeling from this statement that Rick felt that strikeouts per 100 pitches was greatly influenced by the number of pitches per strikeout. That the pitchers at the top of the K per 100 pitches list would on average throw many fewer pitches on strikeouts than people at the bottom of the list. I didn't make that clear last night, and I'd like to make it clear with some research tonight.
For the 2005 season, I looked at all pitchers with at least 50 strikeouts and calculated their Pitches per K. Here are the results, ranked fewest to most Pitches per K.
Pitcher | Pitches Per K | Strikeouts | Pitches |
Brad Lidge | 4.194 | 103 | 432 |
Jamey Wright | 4.347 | 101 | 439 |
Joe Kennedy | 4.351 | 97 | 422 |
Mark Mulder | 4.387 | 111 | 487 |
Brandon Backe | 4.402 | 97 | 427 |
Bobby Jenks | 4.420 | 50 | 221 |
Miguel Batista | 4.426 | 54 | 239 |
Randy Wolf | 4.443 | 61 | 271 |
Nate Robertson | 4.451 | 122 | 543 |
Tyler Walker | 4.463 | 54 | 241 |
Todd Jones | 4.484 | 62 | 278 |
Mike Timlin | 4.492 | 59 | 265 |
Jason Jennings | 4.493 | 75 | 337 |
Chad Qualls | 4.500 | 60 | 270 |
Huston Street | 4.500 | 72 | 324 |
Dan Wheeler | 4.507 | 69 | 311 |
Kerry Wood | 4.519 | 77 | 348 |
Jon Lieber | 4.530 | 149 | 675 |
Jeremy Bonderman | 4.531 | 145 | 657 |
Scott Elarton | 4.534 | 103 | 467 |
Chris Carpenter | 4.535 | 213 | 966 |
Doug Waechter | 4.540 | 87 | 395 |
Derrick Turnbow | 4.547 | 64 | 291 |
Kelvim Escobar | 4.556 | 63 | 287 |
Paul Byrd | 4.569 | 102 | 466 |
Brian Bruney | 4.569 | 51 | 233 |
Jose Valverde | 4.573 | 75 | 343 |
Scott Downs | 4.573 | 75 | 343 |
Trevor Hoffman | 4.574 | 54 | 247 |
Horacio Ramirez | 4.575 | 80 | 366 |
Kiko Calero | 4.577 | 52 | 238 |
Jose Contreras | 4.578 | 154 | 705 |
Carl Pavano | 4.589 | 56 | 257 |
Rafael Betancourt | 4.603 | 73 | 336 |
Duaner Sanchez | 4.606 | 71 | 327 |
Sunny Kim | 4.607 | 56 | 258 |
Kris Benson | 4.621 | 95 | 439 |
Matt Morris | 4.624 | 117 | 541 |
Ben Sheets | 4.624 | 141 | 652 |
Ted Lilly | 4.625 | 96 | 444 |
Giovanni Carrara | 4.625 | 56 | 259 |
Matt Belisle | 4.627 | 59 | 273 |
Brad Halsey | 4.634 | 82 | 380 |
Salomon Torres | 4.636 | 55 | 255 |
Freddy Garcia | 4.637 | 146 | 677 |
Roy Oswalt | 4.641 | 184 | 854 |
A.J. Burnett | 4.641 | 198 | 919 |
Rudy Seanez | 4.643 | 84 | 390 |
David Weathers | 4.645 | 62 | 288 |
Odalis Perez | 4.649 | 74 | 344 |
Lance Cormier | 4.651 | 63 | 293 |
Byung-Hyun Kim | 4.661 | 115 | 536 |
Brett Tomko | 4.667 | 114 | 532 |
Andy Pettitte | 4.667 | 171 | 798 |
Brandon Webb | 4.669 | 172 | 803 |
Jon Garland | 4.670 | 115 | 537 |
Francisco Rodriguez | 4.670 | 91 | 425 |
Ryan Dempster | 4.674 | 89 | 416 |
Carlos Silva | 4.676 | 71 | 332 |
Brian Lawrence | 4.697 | 109 | 512 |
John Patterson | 4.697 | 185 | 869 |
Jake Westbrook | 4.697 | 119 | 559 |
Shawn Estes | 4.698 | 63 | 296 |
Matt Clement | 4.699 | 146 | 686 |
Jeff Fassero | 4.700 | 60 | 282 |
Scott Linebrink | 4.700 | 70 | 329 |
Aaron Harang | 4.706 | 163 | 767 |
Sidney Ponson | 4.706 | 68 | 320 |
Jason Isringhausen | 4.706 | 51 | 240 |
Randy Johnson | 4.706 | 211 | 993 |
Jason Johnson | 4.710 | 93 | 438 |
Kip Wells | 4.712 | 132 | 622 |
John Smoltz | 4.716 | 169 | 797 |
Matt Thornton | 4.719 | 57 | 269 |
C.C. Sabathia | 4.720 | 161 | 760 |
Doug Brocail | 4.721 | 61 | 288 |
Jorge Julio | 4.724 | 58 | 274 |
Jeff Francis | 4.727 | 128 | 605 |
Sean Douglass | 4.727 | 55 | 260 |
Jason Marquis | 4.730 | 100 | 473 |
Danny Haren | 4.730 | 163 | 771 |
Julio Mateo | 4.731 | 52 | 246 |
Victor Zambrano | 4.732 | 112 | 530 |
Derek Lowe | 4.733 | 146 | 691 |
Joe Nathan | 4.734 | 94 | 445 |
Greg Maddux | 4.735 | 136 | 644 |
Mariano Rivera | 4.737 | 80 | 379 |
Dave Williams | 4.739 | 88 | 417 |
John Lackey | 4.739 | 199 | 943 |
Jake Peavy | 4.741 | 216 | 1024 |
Jerome Williams | 4.743 | 70 | 332 |
Javier Vazquez | 4.745 | 192 | 911 |
Billy Wagner | 4.747 | 87 | 413 |
Mike Wood | 4.750 | 60 | 285 |
Runelvys Hernandez | 4.750 | 88 | 418 |
Doug Davis | 4.755 | 208 | 989 |
Claudio Vargas | 4.758 | 95 | 452 |
Dontrelle Willis | 4.759 | 170 | 809 |
Pedro Martinez | 4.760 | 208 | 990 |
Mike Wuertz | 4.764 | 89 | 424 |
Cory Lidle | 4.769 | 121 | 577 |
Johan Santana | 4.769 | 238 | 1135 |
Tim Wakefield | 4.775 | 151 | 721 |
Neal Cotts | 4.776 | 58 | 277 |
Jeff Weaver | 4.777 | 157 | 750 |
Roy Halladay | 4.778 | 108 | 516 |
Matt Wise | 4.778 | 63 | 301 |
Brian Fuentes | 4.780 | 91 | 435 |
Bartolo Colon | 4.785 | 158 | 756 |
Juan Rincon | 4.786 | 84 | 402 |
Josh Towers | 4.786 | 112 | 536 |
Vicente Padilla | 4.786 | 103 | 493 |
Mark Prior | 4.787 | 188 | 900 |
D.J. Houlton | 4.789 | 90 | 431 |
Hideo Nomo | 4.797 | 59 | 283 |
Kevin Brown | 4.800 | 50 | 240 |
Ramon Ortiz | 4.802 | 96 | 461 |
Mark Buehrle | 4.805 | 149 | 716 |
Mike DeJean | 4.808 | 52 | 250 |
Cliff Lee | 4.811 | 143 | 688 |
Brendan Donnelly | 4.811 | 53 | 255 |
David Wells | 4.813 | 107 | 515 |
J.P. Howell | 4.815 | 54 | 260 |
Jarrod Washburn | 4.819 | 94 | 453 |
Wandy Rodriguez | 4.825 | 80 | 386 |
Tom Gordon | 4.826 | 69 | 333 |
Brandon Claussen | 4.826 | 121 | 584 |
Rich Harden | 4.826 | 121 | 584 |
Joe Mays | 4.831 | 59 | 285 |
Kevin Millwood | 4.836 | 146 | 706 |
Brad Penny | 4.836 | 122 | 590 |
Bruce Chen | 4.842 | 133 | 644 |
Cliff Politte | 4.842 | 57 | 276 |
Casey Fossum | 4.844 | 128 | 620 |
Esteban Loaiza | 4.844 | 173 | 838 |
Eric Milton | 4.846 | 123 | 596 |
Josh Fogg | 4.847 | 85 | 412 |
Chad Cordero | 4.852 | 61 | 296 |
Josh Beckett | 4.855 | 166 | 806 |
Tom Glavine | 4.857 | 105 | 510 |
Felix Hernandez | 4.857 | 77 | 374 |
Oliver Perez | 4.866 | 97 | 472 |
Zack Greinke | 4.868 | 114 | 555 |
Robinson Tejeda | 4.875 | 72 | 351 |
Chris Capuano | 4.875 | 176 | 858 |
Jeff Suppan | 4.877 | 114 | 556 |
Mike Maroth | 4.878 | 115 | 561 |
Ezequiel Astacio | 4.879 | 66 | 322 |
Zach Duke | 4.879 | 58 | 283 |
Brett Myers | 4.880 | 208 | 1015 |
Jorge Sosa | 4.882 | 85 | 415 |
Joaquin Benoit | 4.885 | 78 | 381 |
Scott Kazmir | 4.885 | 174 | 850 |
Barry Zito | 4.895 | 171 | 837 |
Kyle Farnsworth | 4.897 | 87 | 426 |
Rodrigo Lopez | 4.898 | 118 | 578 |
Bronson Arroyo | 4.900 | 100 | 490 |
Gary Majewski | 4.900 | 50 | 245 |
Ryan Franklin | 4.903 | 93 | 456 |
Joe Blanton | 4.905 | 116 | 569 |
Luke Hudson | 4.906 | 53 | 260 |
Kazuhisa Ishii | 4.906 | 53 | 260 |
Brian Moehler | 4.916 | 95 | 467 |
Noah Lowry | 4.919 | 172 | 846 |
Kenny Rogers | 4.920 | 87 | 428 |
Tim Hudson | 4.922 | 115 | 566 |
Brad Radke | 4.923 | 117 | 576 |
Ryan Vogelsong | 4.923 | 52 | 256 |
Ambiorix Burgos | 4.938 | 65 | 321 |
Adam Eaton | 4.940 | 100 | 494 |
Mark Redman | 4.941 | 101 | 499 |
Gustavo Chacin | 4.942 | 121 | 598 |
Ryan Madson | 4.949 | 79 | 391 |
Yhency Brazoban | 4.951 | 61 | 302 |
Kyle Davies | 4.952 | 62 | 307 |
Jason Schmidt | 4.952 | 166 | 822 |
Justin Duchscherer | 4.953 | 85 | 421 |
Glendon Rusch | 4.955 | 111 | 550 |
B.J. Ryan | 4.960 | 100 | 496 |
Danys Baez | 4.961 | 51 | 253 |
Pedro Astacio | 4.962 | 78 | 387 |
Kyle Lohse | 4.965 | 86 | 427 |
Roger Clemens | 4.968 | 185 | 919 |
Jason Frasor | 4.968 | 62 | 308 |
Tomo Ohka | 4.969 | 98 | 487 |
Ervin Santana | 4.970 | 99 | 492 |
Joel Pineiro | 4.972 | 107 | 532 |
Dave Bush | 4.973 | 75 | 373 |
Jose Lima | 4.975 | 80 | 398 |
Russ Springer | 4.981 | 54 | 269 |
Victor Santos | 4.989 | 89 | 444 |
Scot Shields | 4.990 | 98 | 489 |
Daniel Cabrera | 4.994 | 158 | 789 |
Carlos Zambrano | 4.995 | 202 | 1009 |
Tony Armas Jr. | 5.000 | 59 | 295 |
John Thomson | 5.000 | 62 | 310 |
Seth McClung | 5.011 | 92 | 461 |
Jay Witasick | 5.014 | 73 | 366 |
Roberto Hernandez | 5.016 | 61 | 306 |
Aaron Heilman | 5.019 | 106 | 532 |
Chris Young | 5.022 | 137 | 688 |
Ron Villone | 5.029 | 70 | 352 |
Chan Ho Park | 5.035 | 113 | 569 |
Orlando Hernandez | 5.055 | 91 | 460 |
Aaron Fultz | 5.056 | 54 | 273 |
Jim Brower | 5.057 | 53 | 268 |
Curt Schilling | 5.057 | 87 | 440 |
Jamie Moyer | 5.059 | 102 | 516 |
Woody Williams | 5.066 | 106 | 537 |
Mike Gonzalez | 5.069 | 58 | 294 |
Gary Glover | 5.069 | 58 | 294 |
Akinori Otsuka | 5.083 | 60 | 305 |
Al Reyes | 5.088 | 68 | 346 |
Livan Hernandez | 5.088 | 147 | 748 |
Francisco Cordero | 5.101 | 79 | 403 |
Gil Meche | 5.108 | 83 | 424 |
Aaron Sele | 5.113 | 53 | 271 |
Hector Carrasco | 5.120 | 75 | 384 |
Ugueth Urbina | 5.124 | 97 | 497 |
Kevin Gregg | 5.135 | 52 | 267 |
Mike Mussina | 5.141 | 142 | 730 |
Erik Bedard | 5.144 | 125 | 643 |
Damaso Marte | 5.148 | 54 | 278 |
Shawn Chacon | 5.152 | 79 | 407 |
Jason Vargas | 5.169 | 59 | 305 |
Jae Seo | 5.169 | 59 | 305 |
Kirk Saarloos | 5.170 | 53 | 274 |
Mark Hendrickson | 5.180 | 89 | 461 |
Justin Speier | 5.196 | 56 | 291 |
Scott Eyre | 5.215 | 65 | 339 |
Brad Hennessey | 5.219 | 64 | 334 |
Mike MacDougal | 5.264 | 72 | 379 |
Andy Sisco | 5.303 | 76 | 403 |
Guillermo Mota | 5.317 | 60 | 319 |
Wade Miller | 5.406 | 64 | 346 |
Al Leiter | 5.443 | 97 | 528 |
The mean for this set of data is 4.815 Pitches per K and the standard deviation is 0.206. That should make the 95% confidence interval about 4.4 to 5.2. There are 236 pitchers in the set.
If Rich's conjecture is true, then we should expect his top pitchers to be near the top of the list. This is not true. The highest ranked pitcher in his top 6 is Carpenter at 21, 4.535 pitches per K. Santana ranks 102, Peavy 90, Pedro 99, Prior 113, and Randy Johnson 70. These five are all between 4.70 and 4.79.
The bottom six do on average throw more pitches per strikeout. Four of them, Lima, Rogers, Lohse and Arroyo were between 4.9 and 4.98 pitches per strikeout. In other words per 100 strikeouts, the bad goup throws about 20 more pitches than the good group! On top of that, Horacio Ramirez, the lowest ranked pitcher in Rich's list, ranks 30th in Pitches per K, right near Chris Carpenter. Marquis ranks 80, better than most of the top six.
This was the point I was trying to make last night. It takes a few pitches to set a batter up for the strikeout. Efficiency comes from other things; not walking batters, or getting players to chase bad pitches early. It's not just from striking out batters on fewer pitches. Strikeouts per 100 pitches does help separate good from bad pitchers, but it's not just because of fewer pitches per strikeout.
Posted by David Pinto at
08:48 PM
|
Pitchers
|
TrackBack (0)
Gosh, I don't know where to even start. First of all, I apologize if I misled you or anyone else. I never meant to suggest that pitchers with high K/100P rates strike out batters using fewer pitchers than pitchers with low K/100P rates.
It's too bad that one sentence (taken out of context, I might add) led you to criticize my idea and findings. I still stand by the point I made in that sentence: "a pitcher who strikes out hitters on three pitches is more effective than those who take five or six to get the job done." The operative words are "more effective." Quite literally, a pitcher is "more effective" if he strikes out a batter on fewer pitches. But striking out individual batters on fewer pitches wasn't the point of the article.
Your list above compares apples to oranges. As I explained, my universe of pitchers consists of those who tossed 162 or more innings. It did not include relievers like Brad Lidge, Bobby Jenks, etc. nor did it include starters with less than 162 IP. Moreover, unlike K/100P, your list doesn't identify anything other than the average number of pitches per strikout.
It's too bad you didn't focus on my conclusion:
"We have known for some time that strikeouts are the out of choice. The more Ks, the better. We also know that the fewer pitches, the better. Combining high strikeout and low pitch totals is a recipe for success. The best way to measure such effectiveness is via K/100 pitches...I believe this stat just might be the best way to measure pitcher dominance, if not overall performance."
The bottom line is that if you believe in the power of strikeouts, you should believe in the power of K/100P as an indicator of pitching success. K/100P does a better job of identifying run prevention than K/9 or K/BF.
That is honestly one of the most interesting lists I've seen in a while.
There's an interesting trend I notice, but need to look up on before I make any conjecture.
I have to agree with David Pinto on this one. It is well known that most strikeout pitchers throw more pitches than guys who rely on hitters putting the ball in play. Strikeouts per 100 pitches rewards the guys with pitches like a sinker which will cause more outs early in counts. Most of the starters on this list are guys who don't have dominant stuff and therefore hittters put the ball into play more often. Guys like Peavy, Santana, Martinez, etc... will cause more hitters to swing and miss which means they will have to go deep into the count more often.
I posted this on Rich's site, but I thought why not annoy more people. If K/BB is so good, then does that mean Carlos Silva with 7.89K/BB is much better than Johan Santana at 5.28K/BB?
But on the other hand, why not just look at #P/Out? Or to get it more linear, by subtracting the ratio by 3? Do people actually use this?
For the record, pitches per strikeout are not the issue here. I'm measuring strikeouts per pitch, not the other way around.
A similar idea to Rich's is The Hardball Times' K/G (strikeouts per "game") which adjusts for efficiency:
"Strikeouts per games pitched. This stat is based on the number of strikeouts divided by total number of batters faced, times the average number of batters per game in that specific league (generally around 38 batters a game)."
Mr. Pinto –
What is the problem here? Do you dispute Rich’s finding that K/P has a higher correlation to the various measures of effectiveness (ERA/ R/G, ERC, FIP, DIPS)? If so, I didn’t read that anywhere.
If you don’t dispute his finding, acknowledge it and then forge your own path to prove the point – which has no relevance to Rich’s conclusion – that looking at the average number of pitches within each strikeout event does not produce a list of impressive pitchers, it just produces a non-impressive list.
I would like to inject that there is one thing I find to be of the utmost value in this list. Closers that are strikeout pitchers like Lidge producing results with lower pitch counts. Otherwise... no effect at all.
As a guy who pitched and caught (with limited success but I was told I played smart) sometimes an extra pitch or two to set up a batter is valuable in that at bat and perhaps even the next few following at bats.
While I enjoy stats there is a certain amount of gamesmanship that I do not believe can be quantified with numbers. Anyways, based on this post and the discussion it is obvious that it is Feburary.
I'm not sure I quite understand your point Dave.
What you're measuring seems to be quite different from what Lederer was measuring, and he pretty explicitly indicated as much in his original article. It shouldn't surprise us that the lists don't match. You seem to be going at things from a top down perspective(strikeouts to pitches), while he's going from bottom up one(pitches to strikeouts), while explains the difference pretty well.
Personally, I'm much more interested in the Lederer list, as it seems to indicate something pretty significant about pitching ability. Pitches per strikeout, I'm not sure what that's measuring. It might be significant as well, but I can't quite get my head around what that would be just yet...
Dave; it seems that a pitcher who uses a lot of pitches but gets the K is rewarded under Rich's system, because the K outweighs the extra pitch or two he uses to set up the batter?
It's not the pitches per K that hurt, it's all the pitches thrown that result in batted balls and walks, right? I think Rich's idea makes perfect sense.
Gee. What a shock. A fresh reliever throwing heat for one inning, two max, is more likely to strike out a relatively tired batter than a starter. I'm gonna party like it's 1959.
In all seriousness, there is an assumption here that strikeouts are the best measure of a pitcher's effectiveness. While Bill James has pointed out that guys that strike out inordinately few batters tend to wash out (e.g. Mark Fidrych), as others have pointed out, groundball pitchers with good control aren't likely to get many strikeouts, either.
Bottom line: HOW a pitcher gets outs isn't as important THAT a pitcher gets outs.
I've now read Rich's piece four or five times. His thesis is that pitchers who strikeout lots of batters and are efficient in their use of pitches are very effective pitchers. I don't dispute this. What I dispute is that efficiency comes from striking out batters on fewer pitches. From my read of the article, this appears to be what Rich is implying. The point of the above chart is to demonstrate that the efficiency doesn't come from fewer pitches per strikeout, but from other places.
Note David's conclusion:
"Strikeouts per 100 pitches does help separate good from bad pitchers, but it's not just because of fewer pitches per strikeout."
It's worth clarifying that Rich's article had several hypotheses in it. (I think David was clear that his commentary was not focussed on the main hypotheses, one way or the other.)
1) One main hypothesis is that K/100 pitch is the best measure of a pitcher's strikeout proficiency.
2) In a closely related claim, K/100 pitch is "a recipe for success" or "the single greatest Defense Independent Pitching stat", since both high strikeouts and low pitch counts are good.
Then there is the subordinate hypothesis which Rich advanced as an reason of why hypotheses 1 and 2 might be believed:
3) strikeouts achieved on fewer pitches are better, because a)the pitcher getting low pitch strikeouts is better at missing bats, b)more likely to pitch deeper into game, c)more likely to record a greater number of outs.
[It's not clear if 3a is good because it means that such a pitcher gives up fewer balls in play, period, or because the balls in play are less likely to be hard hit.]
David has objected to the last (explanatory) hypothesis, and presented rather compelling evidence against it.
I think you guys are both arguing different things, and I think you're both right. I'm not sure Rich was implying that pitches/strikeout was important at all; he echoes that here in a comment above. And you're right David, a correlation between ERA and your table in your post here definitely disproves that idea I think. I think the bottom line of his article is that K/100 is better than K/9 or even K/BB, which is pretty striking to me. And if we trust his numbers, he's right.
David's findings DO shed important light on Rich's thesis. By showing that a high K/P ratio does not primarily reflect "efficient" strikeouts, it tends to confirm that a lower K/P (when comparing pitchers with similar K/BF rates) is telling us that a pitcher is making more pitches on other PAs. That is another indication that what Rich is really picking up is a higher walk rate and/or a higher hits on BIP rate. K/P is a combination of K/BF, BB/BF, and H/BIP, so it's not surprising that it correlates a little better with RA than K/BF alone. But as a hybrid stat that contains far less info than something like FIP or DIPS ERA, it's not clear what value it has.
What's the correlation on pitches per K to ERA or K/BFP rates?
To answer "Wilson"'s question: While K/BB is a descriptive and valued statistic, it does not tell the entire (defense independent) story, as does dERA or FIP. And also, I think Mr. Lederer was trying to analyze defense independent situations only, so pitches/out would not be defense independent. If you want to look up defense independent pitching (DIP) more, look up Voros McCracken, the "father" of DIPS, and mastermind of dERA.
Looking at both Rich's and David's analysis, I'm stuck by what appears to be a glaring problem. And that is that all the pitchers are being lumped together, rather than being separated by league. If you look at Rich's top 10 K/100P list, only two (Santana, Johnson) pitched in the AL last year. The difference between facing a DH or a pitcher is going to skew the statistics.