THE BOOK--Playing The Percentages In Baseball

It’s nice that Verducci’s been “tracking” (what does that even mean?) this, but has he done any proper analysis on the data?  If not, what is his point other than it SEEMS as if what he is saying is true.  That and 50 cents will get you a rid on the NYC subway (actually it’s like $2.00 now).

It may be true and it may not, although it seems plausible that there is some relationship between increasing workload and future chance of injury (the important thing, of course, is the extent of the relationship).  Or maybe it is just workload per se.  Who knows without some pretty complicated statistical analysis.

I think it is irresponsible of Verducci to make the claims he is making without employing someone (I assume he is incapable) to do the proper analysis.

That being said, of his list of “overworked” pitchers in 06, not a whole lot of them were injured in 07 from what I can tell from a brief glance…

I also want to point out that one important reasons pitcher who pitch alot get injured is because they have more opportunity to get injured!

If Schilling faces 1000 batters, and Mariano faces 250, then, Schilling has 4 times more chance to get injured!  Of course, a reliever throws harder, so maybe the 250 batters Mariano faces “feels” like 400 or something.

But, I never see the injury rate in relation to number of pitches thrown or batters faced.

Tango, you have to be careful with that logic.  We are not talking about catastrophic injuries like getting hit in the face with a ball.  We are talking mostly about repetetive stress type injuries, so your argument may be tautological.  IOW, that is exactly what Verducci and others are talking about in terms of pitchers who throw a lot (or have their workload increased too much too quickly) - that the more they throw, the more likely chance of injury.  But the idea of that is different for a pitcher than for a position player.

It wouldn’t be tautological.  What I’m saying is that every pitcher has a, say, .0001 chance of getting injured for every pitch they throw.  Within 3500 pitches thrown, that means they had a 30% chance of getting injured.

What others would be saying is that for the already overworked pitchers, those pitchers have a .0002 chance of getting injured for every pitch thrown.  In that case, within 3500 pitches, that means they had a 50% chance of getting injured.

But, say you do have an overworked pitcher, but he was given a very short leash.  He STILL has a .0002 chance of getting injured for every pitch thrown, but he’s only going to pitch 1750 pitches for the year (i.e., pitch half as much).  Within 1750 pitches, he has a 30% chance of getting injured.

See where I’m going here?  A guy can have a super high chance of getting injured, say .0010 per pitch (10x more than normal), and if he only throws 350 pitches (10x less than normal) for the season, he STILL has only a 30% chance of getting injured during the year.

***

Obviously, the .0001 is not some static number, as it’ll go up or down, after every pitch, and based on how much he’s been used, and how much he’s pushing himself, etc.

But, the opportunity to pitch is a requiremetn to getting injured.  I know I have a 0% chance of getting injured on a MLB mound.  But, if I actually pitch, I’ll have a 99% chance of getting injured.

Tango, I think I agree with MGL here.  There are some small probability injuries that are a function of PT, like colliding with a runner while covering 1B.  But the kind of pitcher injuries we’re really worried about don’t happen during a single pitch, they result from cumulative wear and tear.  There may be a single pitch on which the pitcher first notices the injury (often not), but if he hadn’t noticed it then it would have been the next pitch, or the one after that—it’s not a case of a series of independent low-probability (.0001) chances. (An exception would be something like slipping on a wet mound while making a pitch and getting hurt; but again, most pitcher injuries do not stem from a single pitch.)

I would think it’s more like a pitcher has X% chance of breaking down in any given season.  If an injury occurs, it’s likelihood of being detected is essentially 100%, whether the pitcher throws 60 or 240 IP.  The question then becomes, what’s the relationship between prior usage and X?

I think I will disagree totally here.  I can’t see how making 350 pitches or 3500 pitches could possibly give you the same likelihood of injury, anymore than driving 3500 miles gives you the same chance of an accident as 35,000 miles.

I think the right analogy is not an accident—which would be roughly proportional to mileage (controlling for city vs hwy driving)—but having your car’s muffler give out.  The muffler will give out at some point due to cumulative use, the only question is when.  It’s not like the probability of failure is identical for each trip, and independent.  Rather, the probability increases with each successive trip, and at some point probably grows exponentially.  You may dodge the bullet today, but the odds of failure tomorrow then increase.  A handful of pitchers may retire before they ever have an injury, but not many.  And even these would get hurt if they pitched long enough.

I guess where I disagree is the idea of linking the injury to “a pitch.” Like a muffler, a shoulder or elbow breaks down over time.  When it gives way, it’s not one particular pitch that “caused” the injury (even though it may become apparent on a single identifiable pitch).

But can a muffler give out if you don’t drive the car.

Your analogy is more like the rust of the car.  Whether you drive it or not, your car will rust.  Driving the car more or not, won’t change that.

That’s what you and MGL seem to be implying (time is the determinant factor).  I’m saying that the pitcher’s arm is far more like tires or any other mechanical piece (usage is the determinant factor).

Now I’m confused.  I thought your point was that more IP created more opportunities for injury, SEPARATE from any increased likelihood of injury resulting from usage. If not, I’m not sure what your point is, or how it differs from the traditional idea that increased usage = higher injury risk.

*

I’m not sure how you would measure an injury/PT rate, as you suggest, since injuries directly affect PT.  How would you sort that out?

Sticking with the car-theme:

- driving a car from 0 to 100mph repeatedly, or driving a car for 10,000 miles, then going to 30,000 miles

- driving a car, period (more miles, more wear and tear)

- just having a car, period (rust builds up over time, regardless of use)

The first point is always brought up, the 3-days rest, the complete games, etc, as to the impact of injuries.  PAP, if you will.

The second point is never brought up.  The more you bat, the more you will get hit by a pitch… the more you pitch, the more you will get injured.

The last point, I don’t know if it applies.

I’m arguing for both of the first two points.

http://yankees.lhblogs.com/2008/01/21/pinch-hitting-will-carroll-of-bp/

Minor league innings are somehow not the same as major league innings.

This is an important point. Why are minor league innings any different than major league innings? There are only theories, but the best and most testable center around a selection bias. A pitcher good enough to go over 100 innings in the major leagues is almost by definition a quality pitcher. We know that major league hitters are harder to get out than minor league hitters, not to mention the stress of pitching in front of big crowds. The type of pitcher that can get over 100 innings in the majors is likely to be coasting through the minors on less than his best effort. He’s seldom taxed. He’s seldom forced to bear down or throw long innings. Granted, we don’t know this is the reason why and mathematically and physiologically, it shouldn’t be the case, but until someone can develop a working model for translation, we have to simply ignore those minor league innings. It should be noted that Verducci includes minor league innings in his formula.

This sounds like b.s., doesn’t it?  Do top minor league pitchers coast?

And in any case, Verducci’s model hasn’t even been proven, I don’t care how many years Verducci claims to be tracking things.  Why not present his results in a systematic fashion, rather than just making claims.

There’s a good study in there, and I have yet to see it.

I would also think that throwing 100 pitches in 4 IP is more taxing than 100 pitches in 8 IP. A top pitching prospect dominating the minors is presumably having more stress-free 1-2-3 innings. So it may not so much be coasting as much as staying out of exhausting jams.

Re-reading Guy/5, I’m surprised I didn’t get the independence idea.

The model therefore would be something like this:

Let’s say you are paced to throw 2000 pitches in a season (say 80 pitches a game, 25 games a season, one game every week). 

And let’s say that for the first 100 pitches you throw you have a .001 chance of injury (.999 chance of no-injury), and the second 100 pitches you have a .002 chance of injury (.998 chance of no-injury), and so on.  For your last 100 pitches, you have a .020 chance of injury.

Your chances of pitching non-injured therefore is .999 x .998 x… x.980 = 81%, or a 19% chance of injury.

If you keep extending the pattern to the 3000th pitch, your chance of injury is now 37.5% and on your 4000th pitch, your chance of injury 56.5%.

In order to see if there is this “cumulative” effect, we should be seeing more injuries late than early in the season. 

This would be like a 10yr old car being more likely to break down than a new car.

Otherwise, my original theory would hold, that it would simply be an independent point in a pitcher’s pitching arm life.

Data please.

Jeremy Greenhouse has a good article up at Baseball Analysts today on the Verducci Year After Effect.

http://baseballanalysts.com/archives/2010/02/verducci_effect.php