keyboard_arrow_uptop

Last year, Kevin Goldstein discussed the magic mark where pitching prospects needed to be to sustain any success in the major leagues. With the advent of PITCHf/x, we’re now able to virtually look over the shoulder of scouts and get those radar-gun readings that teams have always had access to, plus location, movement, and more. While scouts had the information, they didn’t do a whole lot with it. Fast was good, slow was bad, and in the end, that’s really all they cared about. Comparisons were superficial.

That 90.5/50 scale mark was more anecdotal than anything; there’s obviously pitchers that have been successful with fastballs that weren’t menacing. It’s just that they’re clearly the exception. Goldstein pointed out that most of the pitchers last season who were living below 90 with any level of success had some gimmick, be it sidearming or showing an angle that most can’t do. While it’s hard to throw in the 90s, it’s harder to grow up to be Chris Young.

A recent piece at FanGraphs was pointed out to me, discussing which pitchers had lost the most off their fastballs. Putting that idea together with what Kevin had put in my head, I wondered if that velocity floor was an absolute and whether pitchers who started above it, but dropped below it would have problems.

I got with Eric Seidman, one of our sharp knives, and he did a little data dissection. I asked him to take a look and see if there was any way to tell whether a pitcher’s game-to-game or even in-game loss of velocity resulted in a noticeable pattern. I tried to keep it simple and just looked for pitchers who had an average velocity over 91 mph who then went to 89 mph or below their next time out. It was easier to go from game to game, so Eric looked at results in terms of the game’s score. The entire look at this isn’t very deep, but remember, I wasn’t trying to do anything more than see if there was any kind of pattern. Here’s the answer in handy spreadsheet form.

The technical answer here would be that there’s no answer at all. We’d have to look deeper to figure out anything more definitively. On the more superficial level, it’s clear that there’s no real pattern here. Some pitchers get better while some get worse. There’s no reason that seems to jump out. A more in-depth look at the pitchers themselves offers no further clarity. There’s no advantage to a certain type, a certain repertoire or secondary pitch, or anything noticeable. There’s aces and future Hall of Famers as well as journeymen and LOOGYs.

So if the loss of velocity on a game-to-game level doesn’t show any pattern of loss of effectiveness in step with the loss of velocity, is there any takeaway? I think as Goldstein’s original work showed, there’s a floor where pitchers-at least right-handers-need to be to be consistently effective. Dropping below that on occasion or as the result of fatigue can have effects if there are not additional elements of the pitching tool kit to make up for it. A pitcher has velocity, location, command, and deception as he stands out on the mound alone. There have been extreme elements of each, but no one is a real Sidd Finch and can rely solely on velocity.

Over a broader sample, and certainly for a player trying to establish himself at the major-league level-or even just announce his presence with authority-a fastball of average or better velocity is a must. Beyond that, it’s up to that thrower to become a pitcher. We’ve seen over and over that while plus-plus velocity will get you drafted, it is no guarantee of even a small measure of success. For a better predictor of success, we’ll have to do more work than this attempt at finding a predictor of impending failure.

It’s a bit bothersome to spend time and effort-especially that of not just me, but that of Seidman, who is working on about a million things in front of and behind the scenes here-but I think it’s important sometimes that we get more than just the formulae and the proofs out there. Baseball needs to constantly be asking questions. Goldstein’s original article is the result of a couple casual conversations and some intelligent curiousity. The bias to velocity is very real in baseball. Finding out that at least at the most basic level-success or failure-it’s not that predictive is interesting.

If nothing else, I’m hoping that someone will look at this and see where this could go. I’m not a doctor, and I’m certainly not a statistician. What I am is curious about the game of baseball, one that rewards our passion with the constant surprise of discovery.

You need to be logged in to comment. Login or Subscribe
Richie
2/18
Ummm. So this is 'Under the Knife' how? I think maybe perhaps you were looking to see if a loss of velocity suggested injury. But found out it suggested nothing clearly. And so then expanded your article, point-wise. Feel free to write about additional things, Will. But don't label them 'UTK' if there's no actual knife in there anywhere. Unless they won't let you or demand 'X' amount of 'Under the Knife's per season. Then I'll approve of your creative labeling.
BeplerP
2/18
What's your point? That the BP guy who has studied more closely than any other the physical task of pitching a baseball and the reasons for the crazy variation of outcomes which affect pitchers' results should cabin his scope of inquiry because it trespasses outside some little box you have set up for him? Please, re-examine your premise for this Comment. You may find, as I do, that it's ugly and pointless.
Richie
2/19
It's not an 'Under the Knife' article, is all. Has nothing to do with injuries. I'm happy to read Will write about something else. It just seems silly to label it something it's not. Like if Goldstein wrote something about Cliff Lee while titling it 'Mariners Top 11'.
DrDave
2/19
One of the things academia struggles with is that there are no incentives to publish inconclusive or negative findings, and strong incentives to hide them. Kudos for saying "we looked at this, and found nothing we can use". That opens the door for others to either look in a different way (and find something), or look more carefully and confirm that there's nothing there to see.
bctowns
2/20
"The bias to velocity is very real in baseball." I guess I'm confused by this column, Will, and this sentence seems to encapsulate it. Do you mean that Kevin's conclusions about the efficacy of velocity in succeeding at the major league level were wrong, and pitchers with below average velocity are unfairly discriminated against despite success in the minors? Or that major league pitchers display a wide randomness of outcomes even when pitching with atypically below average velocity because they've learned the fine art of pitching? I understand you and Eric are very busy with lots of important things (sarcasm), and this question will probably need to be addressed further, but perhaps you could take the time to lay out your position more clearly so those others can proceed from a position of clarity? As it stands, if this column were a comment, I'd rate it below the viewing threshold.
bctowns
2/20
To clarify, I understand that you've laid out your conclusions, but don't you have a faulty premise for this quick and dirty study, with small sample sizes for individual pitchers. It's one game in a seasonal collection of fastball data. Of course there are going to be random occurrences looking at data like this. You also have a sizeable selection bias looking at major league pitchers. I know the pitchfx data isn't there for minor leaguers, but until it is, it seems like an actually meaningful study can't be conducted, unless you are going to look at pitchers' results after their velocity decreases from above average on the scouting scale to below average over a longer period of time/sample size. But even then you'll have selection bias problems.
Severe
2/21
OK, so no one (including Will) thinks this data is conclusive, but what's so wrong about an article just aiming to open up a line of inquiry? Jeez. Looking at the data myself, my uninformed gestalt is of a pretty big pile of mediocrity -- very few examples from really good pitchers who didn't also have injuries last year, but plenty o' Kevin Millwood. I wonder how much some of the variations have to do with factors like weather, wind, or even the opposing teams? Is a relatively small drop in mph turn out to be a leading indicator for what is eventually revealed to be an (ahem) injury? But is the more important thing the 90 mph threshhold, or just the absolute drop (or variation) for the pitcher from one start to the next?