Notice anything weird about this photo?

Samuel Deduno is off and running in 2013. He leads Twins starters in ERA and FIP, as well as opponent batting average, WHIP, and ground-ball percentage. But he’s well known for his unpredictable pitches, and without a full major-league season at age 30, they've held him back. How wild are they?

R.A. Dickey is the league’s best definition of “effectively wild". Here’s a graph of his pitches on Friday, primarily knuckleballs. At FanGraphs, Drew Sheppard noted that “the ideal knuckleball makes just 1 to 1.5 rotations (~150 RPM) between release and home plate, causing the drag on the ball to shift significantly mid-flight as the leading seam of the ball rotates slightly, resulting in unpredictable and uniquely sudden movement.” Not all of Dickey’s knuckleballs spin quite that slowly, but many do, and he’s averaging a respectable 935 RPM on the season (vs. 1600 RPM on his fastball, a good benchmark for the league). Spin angle helps to keep the hitters guessing. As the fingertips push “through” the ball, they roll either under the leather, creating a slight amount of backspin, or over, creating a small amount of topspin. On this graph, anything between 90 and 270 degrees represents backspin. The pitches outside that range have topspin.

Here is a graph of Deduno’s 2013 season, prior to this weekend. I added a few notes on the application of spin angle. His non-curve pitches hardly spin consistently (and he’s credited with a “cutter” for that reason), but except for the occasional “errant change”, they have backspin, as a fastball and change-up normally do.

Now here are two graphs from Saturday’s game in Seattle, a non-televised blackout in both Minnesota and Washington.

Please return your seat backs to their full upright and locked position.*

Ignore pitch classification. On Saturday, Deduno threw a 90+ MPH pitch that frequently spun over the top, adding to gravity’s effect. His non-curve pitches averaged less than 550 RPM, and his change-up spun half as fast as Dickey’s knuckleball, moving at roughly the same velocity. Many of Deduno’s pitches approached that ideal "barely spinning" mark of 150 RPM. That’s essentially impossible with any traditional fastball/change-up grip. In July, his fastballs have averaged ~700 RPM. For a righty, a spin angle between 270 and 360 degrees means a ball breaking down, but in the opposite direction from their curve/slider. The numbers are incredible. Deduno has entered new territory.

The PITCHf/x machine isn’t broken; this isn’t the first time he’s done this. He brought out the same pitch in a hand-full of games last year. This time around, he met or set season bests in earned runs, strikeouts, WHIP, and ground-ball percentage. He looked** good, if not great, and he kept Doumit’s glove moving.

One year ago today, Parker Hageman introduced us to Deduno’s fastball. He noted that “when asked about the possibility of being ‘effectively wild’... Twins pitching coach Rick Anderson joked, ‘Ask Mauer, he says it’s like catching R.A. Dickey.’” What’s Deduno throwing? It’s tough – if not impossible – to say. But it might be time to call it what it is.

Notice anything weird about that photo?

* Fight Club reference. Hopefully I'm not losing anyone.

** Thank god for MLB.tv.

Great information and insight.

Notice anything weird about that photo?

It does look surprisingly like the grip for a knuckleball.

Wow. Great article.

I'd happily call it a knuckleball, if that is what it is.

Here is a dumb question...how in the hell does pitchFX figure out how many RPMs a ball makes? I get the break calculations...

Very interesting story, but I need some clarification. I've read the story a couple of times, but I'm still not sure I know what you're saying. I think what you're hinting at is that Deduno is throwing, if not a knuckleball, the equivalent of a knuckleball. I'd love it if you could be a little more clear on that. We could start with....

1) Which pitch is "knuckling"? Is it his fastball? Or are you saying it is one of the other pitches he throws?

2) Are you saying that his fastball has much less rotation than other fastballs, and that accounts for the movement on it?

3) Should we be evaluating Deduno as a knuckleball pitcher (even though he doesn't really throw a knuckleball) in terms of his future development, as opposed to comparing him to more traditional pitchers? If so, why?

4) What does a regular pitcher - say a Pelfrey or Correia or whoever - have as a graph for their rotation vs velocity?

5) So in his last start, he generated more spin angle? That's why he was more effective? But he doesn't do that all the time?

I guess I'd ask for the following, either as an introduction or a conclusion - can you plainly state, in 100 words or less, the main point of the story, and that which you think the evidence supports?

Edit: Hopefully even more succinct:

Deduno has three pitches:

Wild Fastball - Incredibly-slow (Back)spin

Wild Change-up - Incredibly-Slow (Back)spin

Curve Ball - Normal Rate of (Top)spin

The fastball and change-up have extremely inconsistent spin angles, wherein a pitch can resemble anything from a four-seam to a cutter. On Saturday*, he showed that he can - likely deliberately - also throw these two pitches with topspin, in directions that are most often achieved with a knuckleball.

*And in several games last year.

_ _ _

Did this and the graph as quickly as possible, hopefully it answers the above...

Based on his incredibly low spin rate on all pitches except the curve, it must be assumed that he has always (to some extent) 'pushed' the ball, as opposed to creating a 'normal amount' of backspin by rolling the ball off the fingers, the traditional baseball throw/pitch. This is odd, accounts for the erratic movement, and suggests a knuckleball-type grip.

A curve/slider with topspin falls in the 0 to 90 degree range of spin angle for RHPs, 270 to 360 for LHP. This is easy to illustrate (once I had read a few papers on magnus force ); imagine you are the catcher with this graph in front of your mask. The ball is coming at you spinning in the given direction.

http://i.imgur.com/0czG6Pe.png

The graph is simply reversed (though the grid stays the same) for a LHP. Until Deduno (or a knuckleballer), I hadn't encountered a pitcher that consistently enters the 'dead zone' as illustrated above. The pitcher would need to somehow create topspin, and force the ball to break in the opposite direction from their curve/slider. The occasional odd 'sinkerballs' can track that way, the errant ball (typically a weird change-up grip) could slip out of the hand, the ball could be spit on, or they could be 'pushing' the ball with knuckle (figertip) grip. Based on the 70 or so Deduno photos I've looked at, the last seems the most probable. He's thrown this way in a handful of games in the last two seasons, but consistently on Saturday.

Here's Correia. Again, a fastbal/change-up can 'slip' out of the hand and track a small amount of topsin. But the vast majority of his hard stuff falls between 90 and 270, backspin.

http://imgur.com/YbbFUOP

Kevin Correia, Minnesota Twins - PITCHf/x Pitcher Profile - TexasLeaguers.com

I can't accurately speak to Deduno's future development / categorization. I have to guess that at the very least Cuellar, Anderson, Mauer, Doumit, and Hermann know about the weird spin.

The implication is that Deduno throws 3 pitches (as labeled per graph 3: FF, FC and CH) that have less rotation than R.A Dickey's Knuckleball.

For these three pitches here are there numbers:

CH: 14 pitches, 12 rotated slower than 935 rpm (9 slower than 400 rpm)

FC: 8 pitches, all 8 rotated slower than 935 rpm (6 rotated slower than 400 rpm)

FF: 65 pitches, 61 rotated slower than 935 rpm (10 rotated slower than 400 rpm)

What I take this to mean is that Deduno has three pitches the at least in terms of rotation speed, are comparable to a knuckleball.

Given:

1."The ideal knuckleball makes just 1 to 1.5 rotations (~150 RPM) between release and home plate, causing the drag on the ball to shift significantly mid-flight as the leading seam of the ball rotates slightly, resulting in unpredictable and uniquely sudden movement.”

2. R.A. Dickey is one of the premier knuckleballers in MLB

3. Deduno's fastballs and change-ups on average have less rotation than Dickey's knuckleball.

One would expect to make the following conclusion:

1. Deduno's fastballs and change-ups will show "unpredictable and uniquely sudden movement." Behavior similar to that of a knuckleball.

Thus implying the answer to admin's question 2 is yes.

Thanks guys, this helps me out a lot. (I was the admin.) But what puzzles me is that is seems it is only happening every once in a while? He's not doing this every game? So a couple times per year he looks like a knuckleballer, and the rest of the time he's an average (both in quality and approach) pitcher?

And does the data suggest he is trying to do it but just not quite as good on those days, or that he saves it for some special pitches, or that some days he can do it and some days he just can't at all?

I am going to figure out how to download the data from Deduno's appearances this year, and try analyze it. But just from looking at the images available online I would say he is throwing slowly rotating fastballs consistently.

But that is just from quickly looking through them.

I hope to have more for you by Tuesday evening.

This will be a fun learning process.

It's only the topspin on his fastball that was 'new' and came out consistently on that Seattle start, all of his hard/normal pitches have slow spin.

2013 Averages:

Four-Seam: 763 RPM

(Technically a..) Cutter: 621 RPM

Change-up: 520 RPM

Samuel Deduno, Minnesota Twins - PITCHf/x Pitcher Profile - TexasLeaguers.com

Sorry I missed the further comments til now, I very much appreciate everyone's interest!

Hey E. Andrew,

Just wanted to say this is some mighty fine research and presentation done for this piece. Very entertaining too, which isn't always easy when you are presenting topics such as this.

In regards to the picture, I'm not sure you are attempting to draw this conclusion or that you are just hinting that the grip is giving him a knuckleball-effect, however, the above image of Deduno's grip is for his spike curve and not his fastball.

I do believe are on to something about his spin rate and why his fastball has been so difficult to hit. Keep up the good work.