Wednesday, May 26, 2010

Maybe Bert is Right (Part 1)

Pitch Counts May Be BS

(This is the first of a 3-part series that I’ll be running on the TwinsCentric blog and at TwinsGeek.com. Part 2 will be published on Memorial Day and Part 3 on June 3rd.)

Let’s be honest: for the first one hundred years or so of major league baseball, the players were chattel. That’s the biggest reason that starting pitchers were allowed to throw until their arms fell off. Management didn’t really give a damn if they fell off or not.

That’s also why things have changed. With the introduction of guaranteed contracts, a fragile arm can sink an entire front office. (Just ask Omar Minaya next fall.) So teams, coaches, agents and certainly players are looking for a way to protect those investments. Pitch counts seemed like a good place to start. And 100 is such a nice round number.

But don’t make the mistake of thinking there is any science behind it. Or at least that it isn’t paper thin. The current wisdom that 100 pitches is some kind of limit is an overly simplified interpretation of very specific studies that weren’t afraid to point out their own limitations.

The initial and most significant research on pitching abuse was a pair of essays in the 2001 edition of Baseball Prospectus. The first is called “Re-Thinking Pitcher Abuse” by Rany Jazayerli, which gives a brief history of his attempts at studying pitching abuse and summarizes a new measurement called PAP3. The second is “Analyzing PAP,” written by Jazayerli and Keith Woolner, which details the study that led them to develop the new measurement.

Both essays are very candid about what they found and what their limitations are, mostly without an ax to grind. Unfortunately, the measurements they created have been misinterpreted and oversimplified to become some deranged gospel of truth that doesn’t exist. So let’s take a look at what we REALLY know about pitch counts from those essays.

Re-Thinking Pitcher Abuse Essay
Jayazerli introduces both essays by explaining his original theory: that all pitches are not created equal. In particular, pitches thrown later in a game, once the arm is tired, are more damaging than those thrown earlier in a game.

When Jayazerli had proposed that idea earlier, he also devised a statistic to try and measure it called Pitching Abuse Points, or PAP. The original idea was that the first 10 pitches over 100 would be worth one point each. The next 10, two points each. The next 10, three points each, and so on. The more points, the worse the risk to the pitcher.

(In the later essay, Jayazerli says he chose 100 as a starting point because of research dating back to Craig Wright’s The Diamond Appraised, which suggested the 100-pitch limit for developing pitchers. I’m afraid I haven’t procured a copy of that book to see exactly where it came from.)

Jazayerli had thrown out this statistic as a starting point, but BaseballProspectus.com was exploding in popularity at that time, and he noted that a strange thing happened:

“And for two years, I have tried to use PAP as a framework in which to center the ongoing discussion of pitcher usage. In the process, though, PAP became more than a framework for measurement; it became the standard for measurement. Which it was never intended to do.”

Jazayerli then points out that he had never found any evidence that this PAP score is tied to injuries. He explains that it is a very difficult thing to measure because of all the confounding factors. So he enlisted Keith Woolner’s help and they conducted another study (detailed in the second essay) which resulted in a new measure called PAP3.

PAP3 was similar to PAP except that the points increase exponentially once you get over 100 pitches. Basically, you cube the number of pitches over 100, so 105 pitches would be 5^3 or 125 points. But 110 pitches would be 10^3 or 1000 points. And 120 pitches would be 20^3 or 8000 points.

You can see, that creates some very scary looking numbers in a hurry. However, the standard for what was truly damaging was also raised considerably. So they also included a table which listed the pitch counts along with their risk. Anything below 105 pitches was “virtually none.” Anything under 122 pitches was “moderate” and anything over 133 pitches was “severe.”

So let’s review what this essay just said. First, it explains that there was never any evidence that a previous metric (PAP) was ever valid. It pushes any significant risk in pitch counts up to 120+ pitches. And finally it explains a new metric (PAP3) for evaluating pitcher risk.

Of course, the basis for PAP3 and those conclusions are in the second essay, and we’ll start evaluating that in Part 2 on Monday.

------------------------

I really, really, really cannot believe that I haven’t covered this next item yet. I’ve just been distracted by some life stuff. Many of you may have heard that one of the TwinsCentrick authors, Parker Hageman, has been designing some t-shirts for Twins fans. His initial one is a “Thome is my Homey” t-shirt and the first batch already sold out, but they’ve ordered a second batch. You know you’re going to want to rock this shirt at your next Twins game, so get it now, because I don’t think there will be a third batch.

5 comments:

Jack Ungerleider said...

I thought all the questions about use and abuse with pitchers grew out of the early '80s Oakland team and the belief that Billy Martin basically destroyed a group of promising young pitchers by over using them.

Also hasn't Nolan Ryan instituted a program in the Rangers system that will stretch the minor league pitchers out to 120 pitches on a regular basis? At least that's what I remember from an interview Bert did with him when the Twins went to Arlington the first time after Ryan was named to his front office position.

TT said...

The 100 pitch count is nothing new and it predates any "studies" that bloggers have done. It IS arbitrary, but it also reflects the point at which pitchers generally start to tire. So instead of waiting for a pitcher to get in trouble, managers have used it as guide to when they should turn to the bullpen.

Once you use it as a standard for how long you will stick with the starter, starters get acclimated to throwing 100 pitches. It becomes a bit of a self-fulfilling prophecy.

Whether it prevents injury or not is questionable. In general, injuries in most endeavors increase as people get tired. And, as anyone who has lifted weights knows, if you increase the repetitions a previously easy task becomes much harder.

So I think there is quite a bit of "scientific" evidence out there that a pitch count will reduce the chance of injury. Unless by "scientific" you mean statistics.

Nihilist in Golf Pants said...

Do you have any data on how many times a guy like Bert or Jim Palmer threw 120+ pitches, 130+, or 140+?

I have to believe that a slider is more taxing on the arm than a fastball, so that should be taken into account as well.

I believe that every pitcher is different. Johan Santana seems more fragile than Roy Halliday. I would think that a manager would treat them differently with regard to pitch count.

Cris E said...

A study I'd like to see is the number of really bad hitters there are in the league over time, and how that related to pitch counts. My hypothesis is that a lot of the great pitching (or vast quantity of innings) of the 60s or 70s or 80s could be attributed to the existance and significant playing time of non-hitters like Steve Jeltz, Mark Belanger, Bobby Wine, Hal Lanier, Dal Maxvill, etc. There are not a lot of guys like this amassing 500+ ABs any more, as teams are not willing to settle for defense-only performance. (These guys still exist, of course. I'm just contending that the distribution of talent has tightened up on the low end and the worst hitters on the bench aren't as bad as they were.) Olde Timee Pitcherse like to talk about coasting or saving a little something for the pinch, but in the modern AL there aren't many cheap ABs to coast through. Almost every batter has to be focussed on. Palmer may have thrown 130 or whatever, but if you allow that about 30 were to pitchers and guys like Horace Clarke then it still comes down to a similar number of tough pitches, or at least tough batters.

Do a study where you grade hitters and situations, and then weigh the pitch counts accordingly. I wonder how the bottom third of a modern lineup card compares to twenty or thirty years ago, and how it affects pitchers' endurance.

Anonymous said...

You know what will be interesting is a study that might be able to determine how many pitches a very old time pitcher. Some pitchers threw complete games on only one day rest on a regular basis.

Part of that might be that there weren't quality hitters like one person wrote. But still a bunch of pitchers seem to have long careers who started more games and threw more innings than today's starters.

Walter Hanson
Minneapolis, MN