R.A. Dickey’s Velocity

Esteemed colleagues Dave Cameron and Carson Cistulli are correct — it’s about location and movement when it comes to R.A. Dickey’s excellent work so far this year. Certainly, a knuckleball pitcher isn’t blowing it by the batters he faces.

That doesn’t mean that velocity doesn’t have a lot to do with why Dickey’s been good this year — and why he might be able to keep it up.

There have been many knuckleball pitchers before, but none has ever finished higher than third in the Cy Young voting before. Dickey could break that glass ceiling this year — and it might be because of something he does differently from all the knuckleballers that came before him. R.A. Dickey throws his knuckleball fast. He throws some his knuckleballs (79+ mph) faster than Tim Wakefield threw his fastball (74.1 mph career average).

It’s a little strange that the hardest-throwing knuckleball pitcher of this era has found the most success, though. As our favorite physics professor Alan Nathan found, more velocity means less movement for the knuckleball:

Whereas Wakefield throws at a very consistent 66-67 mph, Dickey throw at two speeds: one in the 73-75 mph range, the other in the 75-80 mph range. The plot shows that the movement on the knuckleball is as random in magnitude as it is in direction. Moreover, the maximum movement appears to decrease with increasing speed.

Whet Moser at ChicagoMag.com re-published the following plot in which Nathan showed the phenomenon graphically.

What might be most amazing about R.A. Dickey is that he has as many as three distinct knuckleballs, though. Look at this plot of his knuckleball velocities from 2010, and you start to see some clusters.

There’s at least two distinct groupings — one around 73-74 mph, and one around 77-78 mph — but there might even be a third. Look at how many knuckleballs he throws above 80 mph. That’s pretty different from 77 mph. And, as Dickey has evolved, those velocity clusters have changed. Check out 2011:

Now the three clusters are more defined. Slow-mo at 74 mph, regular speed at 77 mph, and a fast knuckler around 79 mph. Of course, he still threw the odd super-slow-mo knucklers, but you can see where he settled in most days. You might notice something about the general graph in 2011, though. Try looking at this year’s graph, and the difference should come into focus.

It really looks like R.A Dickey is throwing his knuckleball faster this year. Or, to say it more correctly, it looks like R.A. Dickey is favoring his fast knuckler more this year. His PITCHf/x page tells us his average knuckler has gone up to 77 mph this year from 76 mph, but the graphs tell the story of how he got there better. He still throws his two or three knucklers, but he’s throwing the fast one more this year. And compare how often he throws a 74 mph knuckleball this year to how often he did so in 2010. There’s an evolution here.

Some of it might have to do with the situations involved. From the beginning of his stretch of dominance, there’s been evidence that he uses different knuckleballs in different counts. And even in his last game, you could see that he was using the harder, straighter knuckleball when he needed strikes. But these velocity graphs don’t know counts. They show that regardless of count, Dickey is focusing on the faster knuckleball.

Somehow, Dickey is using less movement and more velocity to make his knuckleball more effective. He’s also throwing his fastball less than ever (13.6% this year, 22.4% in 2011 and 16.2% in 2010). It’s a strange mix of less fastball gas and more gas on the knuckler, but it’s working. His swinging strike rate this year (12.7%) is far and away better than his career number (8.4%), and obviously the rest of his numbers are looking pretty positive.

It’s not your traditional approach — nothing about this knuckleball pitcher is — but it’s clear that gas somehow powers this machine. Next time you see Dickey pitch, watch the radar gun. The numbers won’t look impressive, but they might tell you an interesting story.

Per reader request and with Chad Young’s help, I’ve combined all three graphs into one and used a percent of the total instead of a pure count. I think you’ll clearly see the biggest difference is in the 72-74 mph range: