by Frankie Piliere
Yes, we do makes posts on Easter Sunday. This post is worth it. Coming off a first-hand scouting report from the Cape Cod League, I wanted to follow up with a study I conducted using the Cape pitchers as my guinea pigs. Is it even sabermetrics or just for statheads? I don’t think so. This involves just some pure analysis and uses a number we don’t always use, perhaps because it isn’t always available. Pitches thrown. We know 100 pitches seems to be the magic number for pitchers these days and we know the difference between a good and a bad strike to ball ratio. They tell us of a pitcher’s a durability or how much a pitcher is laboring. However, what they’ve never told us before, without actually charting every game is where those pitches are being distributed, and more importantly how efficiently the pitcher is using them.
There is a stereotype given to strikeout pitchers that says they use more pitches (which they do) and somehow aren’t as efficient as pitchers who pitch to contact. But, have you ever wondered which pitchers achieve their positive outcome (an out in play or a strikeout) in the most efficient manner? Example: A pitcher who struck out every batter he faced on three pitches and induced outs in play on the first pitch would be considered extremely efficient. Obviously that is an extreme example, but you see my point.
I’ve come up with a formula that I believe can show us who gets their outs in the most efficient manner. Usually, pitchers who strikeout more than average are eliminated from the equation of being considered efficient but I think this idea and this formula elminates that and shows they are just as efficient, but instead just use three strikes instead of one.
I’ve discussed this with Kiley and some other good statistical minds to make sure I wasn’t headed down a blind path; because I will be the first to admit that I’m not as statistically inclined as I am pure scouting inclined. But, to better support what I gather from scouting reports and grassroots evaluations, I like to back it up with statistical studies. To make a long story short, though, I at least have a strong feel that I’m onto something with this concept, despite the fact that it has so far proved difficult to put into words.
Perhaps I’ll lay out the formula before going:
PTO% = (SO*3)+SH+SF+(AB-SO-H)/Total Pitches
Basically, the formula takes the number of pitches that go directly into getting strikeouts (three strikes) and inducing outs in play (one pitch each) and divides that by the total number of pitches. The end result is the percentage of pitches used directly towards outs. The three strikes needed to get a strikeout are figured in, which allows strikeout pitchers to get a fair shake in this test of efficiency. The remaining pitches not used towards out are “wasted pitches” or negative outcomes (walks or hits). Therefore, I pitcher who goes to a 3-2 count on every batter before getting an outcome would not be looked at as being very efficient under this line of thinking. Make sense?
I went through hours of modifications to this formula looking for the most telling one. I went through one in particular that considered all outcomes both positive and negative, which did include walks and hits. Under that theory, a pitcher who gave up his hits on the first pitch and allowed his walks on four pitches would have a high percentage and hence, would have a low number of pitches that didn’t go directly into the result. I hesitate to call those wasted pitches but its the temporary name I’ve used. I’d like some feedback on that particular concept, however, because I do think that variation of the formula has some legs as well.
As I said, I test my PTO% (Pitches Towards Outs %) on the Cape Cod League pitchers. Here is a look at some of my results. And, I also want to send out big props to Derek Falvey of CapeProspects.com, a site I hope one day returns to the web. He was able to provide me with the pitch totals that allowed to even conduct this test. Here are the top performers under this formula with the accompanying percentage.
Relievers
Bryan Shaw - 41.4%
Luke Burnett - 40.0%
Evan Crawford - 37.4%
Rob Wooten - 37.2%
Chris Hicks - 36.6%
Nick Cassavechia - 36.4%
Jordan Flasher - 36.0%
Garrett Sherrill - 35.9%
Starters
Alex White - 36.3%
Aaron Crow - 36.1%
Kyle Gibson - 36.0%
Andy Oliver - 34.8%
Jeremy Bleich - 34.7%
Bryce Stowell - 34.6%
Tom Milone - 34.6%
Christian Friedrich - 34.0%
———————————-
I found these results very interesting. I can only hope that I explained the theory well enough. I’d really like a lot of feedback from our readers on this to see what you think. It really seems to me like a great measure of efficiency, considering it shows how high of a percentage of a pitcher’s pitches are going into positive outcomes. The higher the percentage, the lower number of wasted pitches there are in deep counts and on pitches that go into surrendering walks and hits.
This was a long winded post, but hey it’s a holiday; I felt pretty good about myself.
Help Saber-Scouting Grow: Submit or Comment on this story at BallHype and BTF.
19 responses so far ↓
1 Dave Rouleau // Mar 23, 2008 at 3:08 pm
Frankie,
I think that’s a great way to a pitcher’s efficiency toward getting the outs, especially after checking the stats on the Cape Cod League site.
An interesting comparison:
Andy Oliver: 44.2 IP, 22 H, 24 BB, 54 K
Jeremy Bleich: 44.1 IP, 39 H, 12 BB, 47 K
If you look at #2 in your ranking Aaron Crow:
40.1 IP, 19 H, 9 BB, 36 K
In these cases, it could demonstrate the need to put a value on the outs, because you can see the big difference in approach with these three pitchers. They fanned about the same number of batters, but one must still walked 24 batters in 44 innings, compared to the others who seems to throw a lot more strikes, but relies on his defense.
Just my thoughts…
Dave Rouleau
Baseball Digest Daily
2 studes // Mar 23, 2008 at 3:09 pm
Nice Easter article and something to think about. Thanks. But I’m not sure that striking out batters necessarily leads to higher pitch counts. See David Gassko’s article for an explanation:
http://www.hardballtimes.com/main/article/the-kazmir-conundrum/
3 fpiliere44 // Mar 23, 2008 at 3:24 pm
You could be very right in that striking out batters not leading to higher pitch counts. I just wanted to be sure I pointed out that stereotype. But, mainly what I think this does is tell us who is getting their outs efficiently, regardless of how they’re getting them…without giving up many hits or walks in the process.
4 fpiliere44 // Mar 23, 2008 at 3:27 pm
In regards to the first post….yeah I see what you mean with Bleich and Oliver. Despite the difference in how they got there, similarly efficient.
My gut feeling is that this is a concept that even I as the one who conducted the test…..that will take some time before fully understanding what it can tell us.
5 Ethan S. // Mar 23, 2008 at 4:09 pm
It’s an interesting measurement, and an approach that’s not usually the kind of thing sabremetricians approach. It’s the kind of stat that’s more interesting to scouts and managers worrying about preserving young arms than G.M.’s looking for who to draft, as it’s a method of looking behind the relatively simply Innings Pitched measurement. I’d be interested to see if there’s any correlation to actual success in this regard, but I imagine it would be pretty weak.
I do think it’s important to address overall efficiency in a way that both addresses successful outs versus walks and hits, but I’m not sure if it’s possible to get a stat as clean as this one. I also think it would be interesting to compare efficiency in quality appearances versus poor ones, as there’s no reason to think that the efficiency to vary in either case, at least by how the stat is defined. If there is a difference, however, it could prove an important point for managerial tactics in terms of when to replace a pitcher.
6 fpiliere44 // Mar 23, 2008 at 5:24 pm
Great points ethan. I think that’s a great idea. See, this is why I ask for feedback, lol. I love it.
And, I think you’re right on your first point…and it’s the kind of stat I prefer. A stat that kind that isn’t easy to arrive at but extremely interesting when you get the outcome. And, that’s the kind of stat I search for.
The results I got actually correlated quite neatly as some of the more successful pitchers in the league and also correlated well with the pitchers that impressed me as being dominant when I watched them first hand.
I was trying to find a way to better incorporate what % of pitches were used for negative outcome but perhaps the only way to do that is have a separate stat.
Great suggestions, and I think I may use them in some form for future posts.
FP
7 Mark // Mar 25, 2008 at 7:20 pm
It’s interesting, but the trouble with it is that it assumes that a pitch that isn’t one strike of three in a K or made contact on isn’t a pitch to an out. Often in an at-bat, a pitch may be essential to retiring a batter even if it doesn’t fit the formula.
8 fpiliere44 // Mar 25, 2008 at 7:43 pm
I do totally agree with you, Mark, but I’d chalk that up more to the name of the stat itself, which I struggled with. Any suggestions, btw? lol. I don’t think this stats say who is or isn’t good at getting people out, just who does it with the least of pitches used for setup/balls/walks/hits. Someone like Greg Maddux may do awful in this formula (perhaps I’ll check that out) but it doesn’t mean he’s terrible. That’s why this may not be your typical sabermetric stat per say…more of a form of analysis on how guys get their outs. Basically if you score high, this is what that pitcher is doing:
1. Not walking many and not many hits allowed.
2. Striking guys out quickly with not many deep counts
3. Not many deep counts in general.
The guys with the real high scores have all that going for him. The best someone can do is get all their in-play outs on the first pitch and get all their K’s on 3 pitches, never give up a hit or a walk. Man, I’d love to find that pitcher, lol.
If anything, this analysis may almost be a teller of pitching styles. Now, that’s an interesting thought.
9 fpiliere44 // Mar 25, 2008 at 8:37 pm
I couldn’t help myself and I don’t have enough to make a new post about it but I took a look at Maddux and some of his PTO% thru the years.
1992 - 26.8%
2001 - 26.1%
2007 - 21.4%
Thoughts? Makes sense I suppose. He’s had to use more and more pitches to get guys out and go deeper into counts to make it happen. There’s a lot of things we can draw from this.
We know this already but Maddux is very smart and has made immense adjustments to his attack. (He isn’t going into these deeper counts accidentally)
I just found it interesting.
10 Alex S. // Mar 27, 2008 at 3:07 pm
This was definitely a thought provoking article. The big question for me, especially from the perspective of player projection, is what does PTO% really capture? Or, put another way, what new information does it give us?
As others have pointed out, it doesn’t describe a pitcher’s likely success or effectiveness, nor is it intended to.
It might show a pitcher’s ability to go deep into games but pitches/inning might be a better and even simpler stat for that.
One hypothesis I would interested to see tested is that PTO% might provide a more objective measure of a pitcher’s “stuff”. From a purely theoretical point of view PTO% seems to particularly measure how hard it is to put bat on ball.
How well does PTO% correlate with your rating of pitcher’s pitches?
11 fpiliere44 // Mar 27, 2008 at 3:50 pm
Well, you may be onto something. Because that list of Cape pitchers, while not a completely clean correlation, was a pretty good representation of guys with some of the best “stuff” in the league.
One thing I’ve completely decided that this DOES tell us is HOW pitcher’s get their outs. So, maybe it’s a style measurement.
It tells us the difference between a guy that needs to go into deep counts everytime and throw everything but the kitchen sink to get their positive outcome and guys that get easy first pitch outs or quick strikeouts.
A guy with 3-2 counts on every hitter would probably do horrendously in this formula. I think overall it probably shows “ease of outs”. Do you think that captures it a little better?
I’d definitely like to keep getting feedback on this in order to be better understand how to apply it.
12 jyates129 // Mar 29, 2008 at 2:37 pm
I think a good way to test this out would be to test some of the more successful major leaguers and less successful major leaguers and see if the stat adjusts to fit what you expect. Major leaguers have a track record and generally a reputation; it would be interesting to see if your stat can prove a pitchers reputation. It would also be interesting to judge a pitcher’s PTO% in a problem inning. For example, in innings where guys are getting on base, how does a pitchers PTO% adjust to the situation. It could be telling as to who has the makeup/stuff to get through tough situations and who has to pick at the corners and hope for the best. To pick two people I would love to look at in those scenarios would be a ground-ball inducing sinkerballer like a Brandon Webb or Chien-Ming Wang vs. a strikeout pitcher who picks at corners like Al Leiter back in the day. Pitch efficiency is something I love in pitchers, starters especially because in todays game where starters are allowed 110-120 pitches per outing, the most effecient guys are going deeper into games and giving the team a better chance to win. I get frustrated with scouting because a 80 fastball is valued higher than a 70 sinker, but the latter might end up being more effective.
13 jyates129 // Mar 29, 2008 at 2:43 pm
Mike Mussina said it best when asked about Chien-Ming Wang, to paraphrase: He can throw one pitch to a guy and get him out, If I could do that I would do it every time.
14 fpiliere44 // Mar 29, 2008 at 3:08 pm
Thanks for reading. Just to clarify, a scout who does his due diligence is suppose to grade a fastball not just on velocity. If I see a guy sitting at 91 MPH and its diving, darting I’m not just going to give him his standard grade.
And, I really really think you are going to next the next article I post today. Let’s just say your comment is well timed, lol.
15 jyates129 // Mar 29, 2008 at 4:08 pm
Just chalked these numbers with stats from Baseball-Reference for the American League East “Aces” in 07′:
Josh Beckett PTO% = 36.2%
Eric Bedard PTO% = 33%
Chien-Ming Wang PTO% = 29.9%
Scott Kamir PTO% = 29.8%
Roy Halladay PTPO% = 27.7%
To put this in perspective here are each persons GS W IP PITT SO BB
Josh Beckett = GS: 34 IP: 200.7 PITT: 2692 SO:194 BB:40
Eric Bedard = GS: 28 IP: 182 PITT: 2939 SO: 221 BB: 57
Chien-Ming Wang = GS: 30 IP: 199.3 PITT: 2859 SO: 104 BB: 59
Scott Kazmir = GS: 34 IP: 206.7 PITT: 3604 SO: 239 BB: 89
Roy Halladay = GS: 31 IP: 225.3 PITT: 3323 SO: 139 BB: 48
To average some of those things out:
Josh Beckett averaged 5.9 innings per start. He averaged 4.5 pitches per out with a 36.2% PTO.
Eric Bedard averaged 6.5 innings per start. He averaged 5.38 pitches per out with a 33% PTO.
Chien-Ming Wang averaged 6.64 innings per start. He averaged 4.78 pitches per out with a 29.9% PTO.
Scott Kazmir averaged 6.07 innings per start. He averaged 5.8 pitches per out with a 29.8% PTO.
Roy Halladay averaged 7.26 innings per start. He averaged 4.9 pitches per out with a 27.7% PTO.
What I found most puzzling about these numbers is that Josh Beckett, although he picked up the most wins of any of these pitchers, through the fewest innings per start of any “ace” in the NL east. If you look at the gamelogs for Josh Beckett from 07′ he was very carefully watched by John Farrell and rarely touched 110 pitches in a start. This speaks volumes about his durability as a pitcher. The Red Sox were clearly afraid he would get injured, and babied him along through the season. It also speaks volumes about the Red Sox bullpen that they were able to preserve his wins when he didn’t average 6 innings per start with is one of the major qualifications of a quality start.
Roy Halladay by contrast through a lot more pitches, he also made 4 more starts and averaged over 7 innings a start even though only 27.7% of his pitches were made towards outs. This means Halladay was a much more durable and dependable starter than Josh Beckett 07′, which is surprising given Beckett’s reputation as the most dominant starter of 07′ (I’ve heard it said on ESPN I don’t know how many times).
What this also shows is the relationship between a groundball pitcher and strike out pitcher. Chien-Ming Wang averaged half an inning more than Scott Kazmir per start and a full pitch less per out than Scott Kazmir. While Chien-Ming Wang might not look as dazzling as Scott Kazmir when he goes about his business, and their PTO% is almost exactly the same, he is able to go deeper into games and throw less pitches in the process. He also walked 32 fewer hitters than Kazmir. All this leads to a faster, more exciting ball game for the players and the fans.
What this means for PTO% is that it is not a good indicator of pitching style and durability amongst those with comparable talent. What else it says about PTO% I’ll leave up to debate.
16 PTO % « Yankee Psycho-Fan-t // Mar 29, 2008 at 4:15 pm
[…] 29, 2008 · No Comments The guys over a saber-scouting pitched a new stat called PTO% (Pitches Towards Outs %) which measures how many pitches per start a […]
17 fpiliere44 // Mar 29, 2008 at 4:23 pm
jyates, I have a piece going up soon that examines even more guys around the league. I’ve altered the formula slightly though. It’s simpler now I’d say and I also cannot consistently find SH and SF numbers.
I’ve altered it to PTO%= ((AB*3)-SO)+(SO*3)/Pitches
Hence the reason I might be getting slightly different numbers than you.
Anyway, I really think the application for this is simple the ease of getting outs. How easily is a pitcher getting his outs with the lowest % of pitches going elsewhere. Article coming soon.
18 jyates129 // Mar 29, 2008 at 4:24 pm
Can’t wait…I got really into that last one!
19 Jon Crittenden // Dec 10, 2009 at 12:46 am
great discussion. Anyone know what the fantasy baseball people are talking about when they use a stat called ‘efficienct’ for pitchers? You can see it being used at all the offbrand sites like barracudafantasysports and fantasysportscommish, etc. -Jon
Leave a Comment