---

Posted by Paul Singman at 6:13am (11) Comments

Thursday, June 21, 2012

The real deal? ... R.A. Dickey

R.A. Dickey:

Call me a cynic, but I’ve never been one to fully trust the capriciousness of a knuckleball. Don’t get me wrong, it’s a very effective pitch—as evidenced by the numerous careers it has spawned. But, as a Boston Red Sox fan, I’ve had my frame of reference skewed by witnessing the alternating brilliance and ineptness throughout Tim Wakefield’s career.

As such, its difficult for me to see R.A. Dickey in a different light—even though he’s having a third superb season in a row. To me, the vagaries of the knuckleball make him difficult to trust. Just ask Wakefield; one day the ball is dancing past bats, the next day it’s a batting practice meatball. Worse yet, the pattern changes inning to inning on occasion, and sometimes pitch to pitch. For as much as the curveball or slider relies on “feel”, the knuckleball is the ultimate “feel” pitch. Though some days you might have it and some you don’t, your only option is to battle through and trust your “feel.” The guy who can stop overthinking and “let ‘er rip” is oftentimes the best at his craft.

For the sake of argument, however, I’ll forget for a second that Dickey is a knuckleballer. He’s got an excellent groundball rate, he’s got no BABIP or HR/FB problems, and, perhaps counterintuitive for a knuckleballer, he has no problem throwing strikes.

Standing alone, those qualities are enough to place just about any MLB pitcher in the above average range. Last year, that sentence would have summed up R.A. Dickey quite nicely. His expected line reflected that, standing at a serviceable 4.006 ERA with a 5.763 K/9 and 2.216 BB/9.

Unfortunately, that R.A. Dickey couldn’t hack it in fantasy with such a poor strikeout rate. An inflated 3.23 ERA certainly got him on the majority of rosters, but that kind of production couldn’t be expected year-in and year-out without more whiffs. The new R.A. Dickey seems to have solved that problem, however, jacking up the K-rate while turning into one of the most valuable pitchers in the league.

The changes have been staggering: a 7.4 percent increase in O-Swing rate, accompanied by a 6.6 percent drop in O-Contact and 9.1 percent decline in Z-Contact.
It is in these indicators that we usually see the root causes of a breakout, often stemming from an altered approach.

Improvements in O-Swing are usually accompanied by declines in Zone percentage. This shouldn’t be too surprising: when batters show a tendency to chase, pitchers will throw out of the zone more. However, that steep drop in Zone percentage hasn’t happened – Dickey’s Zone percentage has only fallen 1.6 percent this season and only 0.5 percent from his career levels. The oft-used approach of throwing out of the zone and chasing does not seem to apply here.

Further, it is not as though Dickey has added a slider or curve that induces large O-Swing rates. Working with much the same repertoire as years past suggests that an overhaul in pitching philosophy has not occurred.

In fact, from a statistical standpoint, his approach is largely indistinguishable from 2011.

Additionally, because he is a knuckleballer, I have doubts that there’s much of a chance that Dickey could change his approach in ways that wouldn't get picked up in current statistical measurements. Unlike a three-pitch pitcher who can mix locations or add deception via pitch selection, a knuckleballer doesn’t have that same luxury.

By most all accounts, the knuckleball is too inaccurate for the pitcher to confidently place in the strike zone, so location selectivity is out (other than general in-out, up-down placement). Pitch selection could be an option, but evidence points to Dickey being less diverse in his pitch selection than in past years—he’s reduced the use of his #2 pitch, the fastball, to 15.8 percent this season (down from 23.3 percent in 2011). Had the use of his fastball risen compared to previous years, I would concede that this was a possibility, though even then I would have had my doubts.

As for the question of how a knuckleballer may use his additional offerings, Wakefield used to throw a fastball (and a curve now and again), though it was usually reserved for a count where there was a clear red light (3-0, no one on), or as a trick pitch to sneak in a strike, again in low leverage situations. Though Dickey does have 10 mph on Wakefield on the “heater”, I can’t imagine (even in the most aggressive of forecasts) that an improved approach with the fastball would be all that different to the point where it would bring his K-rate up by anything more than 0.5-1 K/9, let alone almost 4 K/9. (Simple math: Dickey throws 16 fastballs per game. Assuming perhaps half of these are strategically placed to take advantage of hitters’ timing or count. Half of these are thrown in two-strike counts. With four fastballs thrown in two-strike counts, a 75 percent contact rate yields one K/9. No numbers to back these up, just a back-of-the-envelope/”finger in the air” type of assumption).

Dickey’s knuckleball itself is another place to look. However, turning over that rock doesn’t yield much either. Its movement profile doesn’t seem to have changed from 2011—and movement deltas within an inch in either direction can often be attributed to statistical variance. With the same movement profile, its largely the same pitch.

So, we’re back where we started: statistically speaking, 2012 R.A. Dickey is virtually indistinguishable from his 2011 self. There is nothing in his profile that we can attribute to his sudden rise in O-Swing percentage and rapid drop in Z-Contact percentage, those being the two greatest determinants of his strikeout rate.

One thing that may yield clues, however, is how my ERA and K models interpreted Dickey’s performance in 2011. His 2011 K and BB results, if you can recall, were very similar to his expected values based on his plate discipline indicators (actual 2011 K/9: 5.78, expected 2011 K/9: 5.763; actual 2011 BB/9: 2.33, expected 2011 BB/9: 2.216).

In my research, I have found that pitchers who exceed their previous year’s expected K/9 rate have a statistically significant tendency to exceed their expected K/9 rate in succeeding years. In short, if you outperformed you K/9 last year, you're more likely to do it again this year.

This variable was introduced to try to account for some of the “X” factor in statistical modeling that is a pitcher's approach. In economic modeling, these kinds of variables are used to pick up on that which we cannot directly measure. In baseball, that “X” factor would attempt to account for a pitcher who saves a little extra velocity for two-strike counts, or a batter who chokes up to avoid striking out. In short, since we can’t record this variable directly, we try to pick up on glimpses of it from the residual inaccuracies in our model.

For Dickey, since he didn’t surpass his previous year’s rates, you wouldn’t expect him to surpass this year’s either. However, Dickey’s 2012 K/9 rate of 9.36 is far higher than his expected 8.35.

Why is this significant?

From his earlier seasons, we would have expected Dickey to have a low discrepancy between these two figures. Since Dickey is outperforming his K/9 rate by about a strikeout per nine, we now have some evidence to posit that there is probably something going on behind the numbers contributing to Dickey’s performance.

Then again, the contrarian would point out that this could all be statistical variance—a position that I would not argue against the possibility of. Either way, as we have used up so many of our tools and are still grasping at straws, you get to the point where you have to put forth some hypothesis to chip away at the question, especially when conclusive data you need to verify it doesn’t exist.

So, after all this, what is the answer?

It hurts my ego to say this, but on this one fellas, I’m stumped. There’s no clear cut answer and there isn’t much out there in the public realm that can provide a solution.

Since I would like to report something, I will say that I think this is part statistical anomaly, part unexplained uptick in performance.

Sorry readers. I hate giving answers like this. I'd rather be bold and beautiful, instead of giving a middle-of-the-road explanation that tends to be boring.

Either way, my expectation is that he keeps the K/9 rate up in the 7.5-8.5 range, posts a dandy ERA, and makes his owners very happy. You probably have a #2 starting pitcher on your hands whom I wouldn’t let him go just because his name is “R.A. Dickey.”

3.333 ERA, 1.187 WHIP, 176.567 K, 12.70 W, 198.41 IP—8.174 K/9, 1.85 BB/9

Skinny: 3.233 points above average at FantasyPlayerRater.com’s roto-points calculator