Velocity, strikeouts, and Colby Lewis’ rough start

Colby Lewis
Colby Lewis is off to a rough start in 2011. Should his fantasy owners be worried? (Icon/SMI)

Yesterday, reader Andrew posted a comment to my personal site asking about Colby Lewis.

I drafted Lewis on three of my teams this year (Tout Wars, Cardrunners, and Yahoo! Friends & Family), in addition to two teams last year. I also made a bet with Eric Kesselman (co-commissioner of the Cardrunners expert league) that Lewis would be worth at least $18.5 in a 12-team, AL-only format this season. Needless to say, I was pretty high on Lewis coming into the year.

Through two starts this season, however, his fastball velocity is down. Andrew pointed out a Mike Podhorzer article at FanGraphs that listed the pitchers who have suffered a decrease in velocity to start 2011. Lewis’ average velocity is 87.9 mph this season, down 2.2 mph from 2010. That’s not something to be taken lightly.

The radar gun and sample size

Some concerns were raised in the comments of Mike’s post about the accuracy of the radar guns this year, and although the stadiums Lewis has pitched in weren’t mentioned, that’s always a possibility. It also should be noted that the velocity readings Mike noted are from Baseball Info Solutions. The PITCHf/x guns are a little more favorable for Lewis, who loses only 1.7 mph based on them.

Furthermore, Lewis hasn’t used his fastball as much this year, and therefore has only thrown a total of about 70 (according to PITCHf/x) or about 90 (according to BIS). When you consider that he threw over 1,800 last year, we’re not looking at very many.

Velocity’s impact on strikeouts

Before going any further, we must first figure out how important velocity is for striking out batters. Dave Cameron ran a study at FanGraphs a couple of years back that looked at the correlation between a pitcher’s fastball velocity and his strikeouts per nine innings. I’ve basically done the same thing here and reached similar conclusions. Looking at all pitchers since 2002 with at least 125 innings pitched, I found a 0.21 r-squared, which means that fastball velocity explains 21 percent of the variation in a pitcher’s K/9.

This is highly significant, but not damning for a pitcher. This means that nearly 80 percent of a pitcher’s strikeout rate comes from something other than his velocity. (Caveat: the pitcher may not necessarily control all 80 percent.)

Application to Lewis

That’s very good news, but we get even better news when we look at the regression equation derived from the above study. If we apply it to Lewis’ 2010 stats, based solely on fastball velocity, we’d expect Lewis to post a K/9 of 6.5. That’s much lower than his actual figure of 8.8*. Based on Lewis’s decreased 2011 velocity, his expected K/9 would be 6.0.

This can be seen as a positive thing for two reasons:

  1. Lewis’ “expected” K/9 was much lower than his actual K/9 last year despite him likely deserving his actual K/9. This likely means that Lewis is able to do things other than merely having a fast fastball to strike batters out. It’s possible that a pitcher like this can lose a tick off his fastball and still be successful.
  2. Lewis’ “expected” K/9 dropped only half a point anyway. That’s hardly going to destroy his numbers.

*You could argue that it’s one year and Lewis didn’t deserve the 8.8 K/9 last year, but I’d argue against you based on his stuff and his terrific MLEs in Japan the two seasons before. THT’s Oliver system projected an 8.4 K/9 for Lewis this season, and ZiPS is even more bullish at 9.0.

Historical velocity dips

I decided to run one more study to examine other pitchers who lost some velocity in April and see how often their velocity returned over the rest of the season. Using PITCHf/x, I looked at all pitchers since 2007 (the year it was introduced) who had lost at least 1.5 mph off their four-seam fastball between the previous season and April. I then looked at their velocity for May through October and compared the three.

To qualify, a pitcher needed to throw at least 1,000 fastballs in the previous season, 100 in April, and 600 in May through October, which gives us a sample of 23 pitchers.

+---------+-----------+---------+
| Y-1 MPH | APRIL MPH | M-O MPH |
+---------+-----------+---------+
|    91.2 |      89.3 |    90.2 |
+---------+-----------+---------+

While our sample size is less than ideal, this tells us that, on average, pitchers who lost at least 1.5 mph on their fastballs in April regained 52 percent of their velocity over the rest of the season.

For Lewis, that would mean that his fastball velocity would rise to 89.2, an overall drop of just 0.8 mph from 2010. If we plug this into our “expected K/9” equation, Lewis’s expected K/9 would be 6.3 with an 89.2 MPH fastball. That equates to just a three percent drop in strikeouts for Lewis based on his velocity loss.

If we look at the results a slightly different way, we see that 87 percent of the pitchers in our study improve their April fastball velocities in the following months. Just four percent end up reaching their previous velocity level, but 22 percent get within 0.5 mph.

Prognosis for Lewis

Summing everything up, things don’t look terrible for Lewis. He has a few things going for him:

  1. The PITCHf/x gun sees only a 1.7 mph drop on his fastball.
  2. Lewis has thrown a relatively small sample of fastballs so far this year.
  3. Lewis uses a lot more than his fastball velocity to strike batters out.
  4. Since 2007, pitchers in this situation end up regaining 52 percent of their lost velocity.
  5. For Lewis, regaining 52 percent of his lost velocity would result in just a three percent total reduction in strikeouts the rest of the way.
  6. Even if Lewis doesn’t regain any of his velocity, his expected K/9 would still drop by only half a point.

Concluding thoughts

Maybe I’m trying to put a positive spin on things because I have such a vested interest in Lewis’ success this year, but I don’t think it’s time to hit the panic button yet. We’re just two starts into the season, and a lot can still happen. If a Lewis owner is panicking, it might be wise to acquire Lewis cheaply. If you’re the Lewis owner, I don’t think we have much of a choice but to wait and see what happens. I have a feeling we’ll be fine, though.


11 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
Dave
12 years ago

Who did Dave Cameron runt that original study on at Fangraphs a couple years back?  Im interested, it may reveal an underlying point of the same kind of “loyalism”  you give to Lewis here.  Not that I don’t think Lewis will be successful, actually quite the opposite. Just curious.

Derek Ambrosino
12 years ago

I hope you’re right. The small text counter argument is a bit worrisome though, because at least anecdotally, one might also be tempted to attribute a lower K-rate, when not fastball-reliant, to increased familiarity with Lewis and his stuff in his second year.

-Fellow DOL Society member
(Dereks Owning Lewis)

Derek Carty
12 years ago

Dave,
I’m not sure who it was.  In the article I linked, it doesn’t seem like he’s talking about anyone in particular, but there might have been companion pieces to it.  I don’t know.

Derek A,
That’s possible, and that kind of thing would make for an interesting study.  I wouldn’t too much weight on it without having done one, though, as there are plenty of pitchers who are successful year-after-year with the same stuff.

Scott
12 years ago

I hope you’re right, as this SOL (Scott Owner of Lewis) is feeling a little… SOL.

What, you didn’t think I was going to go there? smile

Mike Podhorzer
12 years ago

More good news on Lewis: he primarily uses his slider to generate punchouts. Last year according to the TexasLeaguers site, he had a 19.6% whiff rate on the pitch. His fastball whiff rate was just 5.4%, which is close to what it has been this year. So the fastball is certainly not what make Lewis successful, however a sustained decline in his velocity should make it easier for hitters to sit on the off-speed stuff.

Derek Carty
12 years ago

Perhaps one more positive thing for Lewis: http://www.beyondtheboxscore.com/2011/3/30/2080943/by-mike-fast-who-else-fastball-velocity-trends-over-the-course-of-the

Although, one of his games was in Texas when it was I think 70 or 80 degrees.  The other was in Baltimore when it was 50-ish.

Jonathan Halket
12 years ago

Nice article. I think the key question is Why do we worry about Lewis’s lost velocity? 2 potential answers (and perhaps more)

1) velocity is a component of strikeouts and Ks are key to success.
-if this, and if Lewis is good at getting Ks with other pitches as MP notes, then maybe we don’t have to worry to much.

2) decreased velocity might point to arm trouble.  If this, then we should be worrying. One wonders how many pitchers had this decrease in velocity and didn’t manage to accrue enough innings to count in some of Derek’s studies above…

Derek Carty
12 years ago

You raise a very good point, Jonathan, in there there could be some survivor bias here.  If I rerun the study with all of the parameters exactly the same, except that I only look at pitchers who threw less than 600 fastballs from May-October (as opposed to more than 600), I get just five pitchers:
– Rich Hill (2008)
– Oliver Perez (2009)
– Daniel Cabrera (2009)
– Dennis Sarfate (2009)
– Todd Wellemeyer (2010)
– Paul Maholm (2010)

Jonathan Halket
12 years ago

Interesting. There are definitely a few injury concerns on that list. But the overall number is small (6, not 5, no?). Still if that meant a 20% injury concern, that’d be more than peanuts.

In the end, I think, as you say, the warmer weather will tell… if in June we’ve not seen an increase in velocity, then perhaps we need to watch him like a hawk even if his ephemeral numbers are good…

Derek Ambrosino
12 years ago

I agree, Derek C.

But, just playing devil’s advocate, the question isn’t whether there are so many pitchers consistently successful with the same stuff year after year, but how many there are who are able to maintain nearly a strikeout per inning with a fastball that tops out at around 90mph?

We know Randy Johnson K’ed the world with the same stuff every year, but that isn’t the stuff that Lewis has. So, perhaps the more interesting study would involve pitchers who outproduce the K/velocity projections in their early years and whether that specific trend continues over time.

I’m remain confident that Lewis will be good for somewhere in the neighborhood of 175-185 Ks though.

howard
12 years ago

To be fair, who knows how the pending birth of a child may affect him.