11 May 2010

Mariano Rivera Watch

Having watched my fair share of games on MLB.com this season, it struck me that Mariano Rivera's velocity seemed to be dipping. And yet, in his 10 appearances this year this doesn't seem to have affected him (unlike the stiffness that has kept him on the bench for over a week). I decided to look at this more closely, using Brooks Baseball PitchFX data as well as FanGraphs' resources. As you can see in the chart below, Mo's velocity is dipping, although not as much as I thought.

At one point this season, Rivera had lost 0.6 mph on his fastball between this year and last, but in his last two starts he really let it fly, bringing his season average within 0.2 mph of the previous year's. The graph (below) represents the season-to-date average speed on Mo's famous cut-fastball, comparing his numbers from each of his first 10 starts in 2009 and 2010 (with a Lowess curve overlay; for some reason BB's data on Mariano's 9th outing in 2009 was unavailable, so I cheated a bit and pretended that appearance didn't happen).

Even when Rivera's cut fastball was coming up short through 8 starts, it certainly wasn't affecting overall outcomes—Mariano has yet to allow a run this year. Compare this to April 2009 when he had already blown two saves. So why is Mariano still able to compete without the zip on his cutter? Two reasons: break, and luck. While the cut fastball's velocity is dipping, Mo has added to the already maddening break of his cutter. As you can see in the graph (below), Cut Fastball v.2010 is breaking a full inch more than the 2009 version.

Methodology: I wasn't satisfied with the idea of  adding h-break and v-break together to obtain a "total break" value, since pitches don't move that way. Instead, I calculated the Pythagorean break, assuming that the h-break and v-break were the lengths of the perpendicular sides of a right triangle, and calculating the Pythagorean break as the hypotenuse (P-break = sqrt( h-break^2 + v-break^2 ). Yes, I know that pitches don't move that way either, but I think it's a bit closer to reality.

That said, I'm not so sure we can give all the credit to Mo in this case. First off, the strike rate on his cut fastball is down, indicating that the pitch is fundamentally less effective. Second, Rivera's opponents are only hitting 0.155 against him when they get the ball in play. That's a full 0.120 below his career average, and 0.230 below where he was this time last year.

Now, it could be that the added break is compensating for the decreased velocity by creating more ground ball outs. However, an opponent BABIP of 0.155 is simply unsustainable, even for the Sandman, whose career oBABIP is already strikingly low. Looking at the data below, it's pretty clear how fortunate Mo has been this year.

Cut Fastballs (April)
All Pitches (April)
Year Gm Pit Cut% Mean Velo. Mean Break Str%
BABIP SV SV% WPA WPA/LI
2009 10 126 89.4% 90.9 mph 6.7 in 72.2%
0.385 5 83.3% 0.36 0.22
2010 10 131 91.2% 90.7 mph 7.8 in 67.7%
0.155 7 100.0% 0.58 0.40
0 5 1.8% -0.2 mph 1.1 in -4.5%
-0.230 2 16.7% 0.22 0.18

It's interesting that Mo has been able to keep up with his opponents even considering the fact that he's been tossed into more high-leverage situations this year. Of course, we know that pressure isn't a problem for this particular star. By the way, if you think the secret lies in the employment of his four-seamer, think again: he's using it even less this year than last.

Will Mo get his velocity back, and will the 40-year-old future citizen of Cooperstown continue to shut down games as the season progresses? I'm not sure, but I'll keep tabs on his performance and peripherals throughout the season.

(Photo Credit: Me)

1 comment:

Sturgeon General said...

Well this drop in velocity is already similar to the drop in velocity had 6-7 years ago when he went from throwing mid 90s to low 90s. Mo survived and continued to thrive, no reason he cannot continue to now as long as he can continue the deception between fastball and cutter.