Sunday, July 11, 2004

Pitch Counts

I've been having a dispute with Ed Cho about pitch counts.

Cho thinks they're given too much importance. I disagree. I think that for young pitchers, it's very important to keep a low pitch count. Pitching puts extreme stress on a pitcher's arm. I read an article once about how guys who pitch a ton of innings early in their career burn out. (Note, I'm using innings as a proxy for pitch count here.)

To test this, I looked at career leaders in wins. (Because the general consensus is that pitching these days is tougher than the 50's because the 7-9 hitters are so much better, I arbitrarily required a pitcher to start his career in 1960 or later to be considered.) I look at how many innings these guys threw. Because they are the career leaders, they are the most durable arms in baseball over the last 40 years. I assign an age based on that pitcher's age at the start of the season. I only look up to the age 26. The first number is the age the pitcher first threw 100 innings, the second number is 200, the third is 300 (if he pitched 300 before turning 27)

Pitcher
Carlton 22 23
Ryan 21 25 26
Sutton 21 21
Clemens 22 24
P. Niekro (a knuckleballer, he's obviously not too relevant)
Perry 25 25
Seaver 23 23
Maddux 20 21
Blyleven 19 20
Jenkins 22 23
Palmer 20 20 23
Glavine 22 24
Marichal 23 24 25
R. Johnson 25 26
Tanana 20 20
Tiant 23 26
Hunter 18 20

Note: Catfish Hunter's last season was when he was 33.

My point is, it looks pretty clear to me that pitchers who last a long time didn't throw a ton of innings when they were in their early twenties. This specifically came up with Cho when we were talking about the Cubbies. Kerry Wood threw 166 innings when he was 20. He lost the next year to injury, and surpassed 200 innings when he was 24 (which is about right, I think). We'll see if he lasts. Mark Prior threw 116 innings at 21, 211 at 22. To me, that's worrisome. Why take a chance?

Mulder 154 when he was 23, 200 when he was 24. Zito threw 92 when he was 21, 200 when he was 22, and he has arm troubles. Hudson threw 136 when he was 23, 200 when he was 24.

I don't know, but it looks like a trend to me. If I were a GM, I'd limit innings until a pitcher turned 24.

2 comments:

Anonymous said...

this is cho. i didn't want to login.

drake is smoking crack. not really, i'm just mad that the cubs are 7 games back of the cards, after just getting spanked in st louie.

my claim is that pitch counts are thoroughly arbitrary, and the magic "120" number is numerically non-significant. know why the number is set at 120? PAP^3. come with me and learn.

http://www.baseballprospectus.com/news/images/20020521_03_woolner.gif

this is from basball prospectus which is generally a viable source of info. the graph shows performance for 21 days after a pitcher throws X number of pitches. example: livan hernandez has a 3.90 ERA, then throws 134 pitches. they track livan's performance for 21 days after his 134 pitch outing, and measure the increase in RA (runs against). increasing is bad. the data goes from 1988 to 1998 (when this was published).

so looking at that graph, there is a clear slope change at 120 pitches. you know what that change is?
- after pitching 120 pitches, a pitcher's RA goes up 0.015% for the next 21 days.
- after pitching 130 pitches, a pitcher's RA goes up by 0.035%

an increase of 0.02%, that is 0.0002!! those people with any sense of numbers realizes that this data a trend does not make. this percent change is NOISE. if you change the scale on the y-axis, i can prove that there is an enormous change between 100 and 110 pitches. on a graph magnified to this level, sure, it looks obvious. but understand the scale. the data they pull does not give this amount of accuracy. if i have four apples, and there are three kids, do i give each kid 1.33333333 apples? no, i give each one apple, and i keep the fourth.

my horror is that ultimately, it's the media that is stoopid and they make the general public stupid. these are the same media people that talk about the "differential" instead of the "difference." 120 pitches came from this diagram and as you can see it ain't really significant. but numerically-challenged people have decided this is statistically viable. now, we have it in the heads of every pitching coach from college, to double A to the majors that 120 is "alot." now, they won't pitch these kids hardly past 100 without catching hell. as a (lazy) sprinter i would have loved this because i never would have had to run more than 100m for a workout.

anyway, my crusade is to have people understand that 120 pitches is a stupid number that came because the scale on the y-axis on a graph was incredibly magnified. as i've said to drake 15 times, 120 pitches for maddux would end his career tomorrow. 120 pitches for randy johnson produces 1.01211314 grams of sweat.

pitches are easy to count, but there are too many factors to consider. 15 pitches/inning for 6 innings is very different that 30 for 3 innings, but both are 90 pitches total!

no one is still reading, so i quit here.

Anonymous said...

cho again.

kerry wood blew out his elbow not because of innings pitched but essentially because of the 20k game against houston. when he looks back at tape of that game, he cringes. basically, he started throwing this slurve pitch that had wicked movement, but also did wicked things to his arm. it worked in that game, so he kept throwing it in the games after. he was 20 years old and had a pitch that major league hitters couldn't touch. that is what destroyed his elbow. it wasn't the number of total pitches, it was the number of times he threw that slurve.

blame the pitching coach for not catching that.

also, bull durham is a horrible movie, if for no other reason than robbins couldn't have made the rockford peaches and that rosie o'donnell had a better throwing motion.