The Video History of a Prospect Legend

Let’s take a trip back to 2006. All anyone could talk about was a hot new video-sharing website called YouTube. For the first time, baseball fans could access at will clips of players they couldn’t watch every day. Fans and scouts who had traveled to minor-league games and showcases could take short clips—often only a few seconds long—and post them on the Internet for a wide audience. People who had once settled for what scarce minor-league box scores they could scrap together (or, perhaps at best, receive as an e-mail from some guy with a Prospect Report) could now spend hours poring over grainy video taken from unstabilized handicams. It was truly a breakthrough.

Legends, tradition has it, are either oral or written. Both forms lend themselves to hyperbole, embellishment, and exaggeration. By way of contrast, visual proof, whether still or moving, is something of an evidentiary gold standard. That is to say, video evidence is presumptively true and impeachable only with very careful proof. So it was very difficult for those of us who were just for the first time seizing upon the Internet video revolution to watch prospect clips at odd hours of the evening to accept what we began seeing in 2006. But there it was—proof of something that appeared to the untrained eye to be legendary.

Early in 2006, clips began appearing of a short, sinewy pitcher with a funky rotation and a push-off that left him momentarily and all but impossibly off the ground. He pitched for the University of Washington and clips of his odd delivery soon began to be passed around. Some aspiring home Dede Allens slowed down the video to show just how unorthodox the wiry kid was.

After being selected 10th overall by the San Francisco Giants in the 2006 draft, Tim Lincecum’s YouTube profile increased drastically. By the offseason, after compiling a 48/12 K/BB ratio in 27 ⅔ innings in his pro debut, Lincecum was an Internet video sensation. By the time he was called up in May of 2007, Giants fans weren’t the only ones eagerly anticipating the day.

Throw Hard Or Go Home

Part of what made Lincecum’s delivery so sensational was the velocity it allowed him to generate from his relatively slight frame. Upon his arrival in the big leagues, Lincecum’s velocity (no doubt aided by some hot guns and eager broadcasters) was clocked at 99 mph. He gave the impression that he threw upper 90s heat all the time, snapping off blazing heat with every pitch. It was a great way to set expectations high, especially because it apparently confirmed what everyone had already been told.

But hard-throwing righties are not especially remarkable commodities, and ones who are short of stature are typically considered oddities. What made Lincecum worth getting excited about was his complete dominance at every level of baseball. That, of course, did not come just from the gun reading on his fastball. For one thing, his delivery allows him to deliver the ball in an exaggerated, over-the-top fashion that hides the ball until the very last second—an effect reinforced by the push-off leg kick that causes him to leave terra firma with each toss. For another thing, Lincecum complemented the fastball with several off-speed offerings, most notably a yakker of a curve ball.

But since his debut, evidence suggests the velocity on Lincecum’s fastball has been declining gradually. Some commentators have begun to take notice. What could possibly explain this phenomenon (“Dance Hall Days”?), and why hasn’t it affected Lincecum’s performance?

In New-Wave Magazines

Using available Pitch F/X data from 2007-09, I gathered data on all of Lincecum’s pitches thrown at home. (Big thanks to Jeff Zimmerman for data assistance.) Two important explanations apply here. First, the data from 2007, and to a lesser degree 2008, is not complete. In other words, not every single pitch was accurately recorded by the Pitch F/X cameras. However, the sampling of pitches is more or less random and therefore ought not to affect the analysis. Second, I have chosen only to use data from home because of concerns about park effects, which do not bias batted ball data exclusively.

That being said, there were a total of 583 such pitches from 2007, 1,751 from 2008, and 1,855 from 2009. Of those, I classified 374 fastballs in 2007, 1,152 fastballs in 2008, and 1,013 fastballs in 2009 (I am using my own pitch classifications due to issues with Gameday’s classification algorithms, which continue to be improved). What is patent from that data is that Lincecum’s average fastball velocity did decline from 2008 to 2009. In 2007, Lincecum’s average fastball velocity at home was 92.9 mph. In 2008, it was 93.3 mph. Last year, it dropped to 92.1 mph. However, we’re dealing with a large number of pitches and these readings are certainly capable of fluctuation.

What if, instead of average speed, we looked at percentage of fastballs that exceeded a certain velocity. That is exactly what this chart does:

The effect is most dramatic if you focus on the percentage of fastballs that were faster than 94 and 95 mph. Whereas in 2007 and 2008, approximately 30 percent of Lincecum’s fastballs were faster than 94 mph, last year only about 10 percent were. Additionally, the 96-plus mph fastball effectively disappeared from his arsenal last year, as he threw just one in the sample I am using. This, it would seem, would not bode particularly well for the two-time Cy Young Award winner’s future.

It’s Not How Many It’s How

Of course, no mistake of analysis is worse than looking at only one category in isolation. To understand the effectiveness of Lincecum’s fastball, we must also look at its other characteristics, particularly since we know he throws predominantly two-seam, rather than four-seam, fastballs (something he has been doing since his debut). Although Lincecum may occasionally throw four-seamers as well, I have grouped all fastballs together for the purpose of the analysis, for reasons that should become clear below.

Looking merely at horizontal movement (in Pitch F/X jargon, that is the distance in inches the final location of the ball differs from a hypothetical straight-line path), his fastball last year had more movement than it did in the past. The average (really average of the absolute values) horizontal movement on his fastballs was 3.6 inches. In 2007, the same mark was 2.2 inches, and in 2008 it was 1.9 inches. But would it really be wise to trade a tick or two off the fastball in exchange for a measly inch and a half of break?

Well, it’s actually even more interesting than that. Because not only has the average break on Lincecum’s fastball increased over the past two years, the amount of variance from pitch to pitch in the break has been increasing steadily from 2007 through 2009. To calculate this fluctuation from pitch to pitch in the tail on Lincecum’s fastballs, I calculated the standard deviation of the horizontal movements for all the home field fastballs. Here are the results:

2007: 1.8
2008: 2.2
2009: 2.6

The steady increase in the standard deviation suggests Lincecum’s fastball is moving more and is also less predictable for batters. It is likely a safe assumption that unexpected movement is harder for a batter to pick up than expected movement, and the wider range of horizontal movement on Lincecum’s fastball would likely be unexpected even to a batter who had faced him in years past.

There is undoubtedly an optimal point in the tradeoff between velocity and movement, and one variable in calculating that optimal point may well be how long it is possible to maintain. Though Lincecum’s unique delivery effectively put him beyond the realm of even the top mechanics experts, it may well be the case that the lower-velo, higher-movement Lincecum fastball is a recipe for longevity.

Question of the Day

Is there any reason to believe that Lincecum’s declining velocity is a cause for concern? Have you noticed the increased movement on Lincecum’s fastball?