Pitch Tracking: Ricky Romero vs. Cleveland

Opening Day brings out all kinds of strange behaviors in baseball fans. For non-playoff teams, like the Blue Jays, fans have been waiting six full months for meaningful baseball to return to our television sets. That hiatus finally ended yesterday, and I chose to celebrate the occasion by tracking pitches for the Jays’ starter, Ricky Romero.

Pitch tracking is an aspect of advanced scouting that every team uses in preparation for starting pitchers in an upcoming series. On Thursday’s broadcast, Buck Martinez mentioned how former Blue Jays catcher Kevin Cash was an offseason addition to the advanced scouting department and that he had been helping prepare the team for the Cleveland series.

The PitchFX system has taken away from pitch tracking a little bit, but there’s something to be said for having human eyes watching a pitcher as opposed to relying entirely upon a computer algorithm. Computers can’t think for themselves, which can cause problems on occasion, including an issue I encountered during today’s game, which I’ll touch on later.

The first step in my preparation was to use the PitchFX data from FanGraphs to get a better idea of the specific velocity and movement of Romero’s pitches. As a Blue Jays fan I know what kind of an arsenal he has, but I wanted to have some actual numbers on my page that I could reference throughout the start for pitch classification. The data (from the 2011 season) was as follows:

47.5% | Four-seam fastball | 92.0 mph
23.4% | Two-seam fastball | 91.5 mph
18.7% | Changeup | 84.9 mph
9.1% | Curveball | 77.1 mph
1.2% | Slider | 84.5 mph

As you can immediately see, he’s heavily reliant upon his two fastballs, totaling 70.9% of pitches thrown. Romero’s changeup and curveball -– both excellent offerings -– were also thrown with significant frequency. The slider, however, was seldom used. To further that point, Romero threw 0 sliders in Thursday’s start against Cleveland. Now that we have some background information, onto the game.

The first inning was arguably Romero’s best of the day, as he efficiently set the Indians down in order while throwing only 11 pitches (seven strikes). He worked almost exclusively with the fastball — 10 of the 11 pitches were heaters ranging between 90 and 92 mph. The 11th pitch, and last of the inning, was a 79 mph curveball that Indians right fielder Shin-Soo Choo pounded into the dirt in front of Romero for an easy groundout. (Disclaimer: I found it extremely difficult to differentiate between a left handed pitcher’s two-seam and four-seam fastballs (particularly with such a small velocity difference) from a center field camera, so outside of pitches I was able to be definitive with, I’ll refer to them simply as fastballs)

The second inning was a completely different story. Romero allowed two walks, two doubles and gave up a huge three-run home run to Jack Hannahan while throwing a staggering 43 pitches (23 strikes). That insane pitch count took an immediate toll on Romero’s arm, as after sitting 91-92 with the fastball in the first (and peaking as high as 94 mph on a four-seamer to Jason Kipnis in the second), the velocity fell to 88-90 mph in the latter half of the inning. The inning did see Romero throw his first changeups of the afternoon, as he tossed seven in the frame. Romero was effective with the pitch, as five of the seven changeups went for strikes and the other two fell safely down below the strike zone. His one miss was a hanging changeup to Shelley Duncan which went for a double to left field, and the curveball had a similar fate. Romero threw the breaking ball seven times in the inning, with the one mistake being the curveball that split the plate, thigh-high to Jack Hannahan. The other six went for three balls and three strikes, with the balls being inside or in the dirt and the three strikes resulting in a swinging strike, groundout, and strikeout.

The third inning was where PitchFX began to show some of its inadequacies. Romero’s fastball velocity continued to drop, plummeting as low as 86 mph on a 2-0 pitch to Carlos Santana, and touching 90 mph on only two of the 13 fastballs thrown in the inning. The 86 mph pitch was wrongly classified as a changeup, as its velocity was far more similar to Romero’s 84.9 mph average changeup velocity than to his two-seam fastball average velocity of 91.5 mph. The issue continued to occur with more frequency in the fourth and fifth innings. PitchFX does not account for excruciatingly long innings wearing an arm out, which once again is why human eyes will always be superior to computer algorithms in terms of scouting. When fastball velocity drops 4-5 mph in a start, you must adjust expected off-speed velocity accordingly. The curveball showed up only once this inning -– the ball in dirt tagout play that ended the third –- while also showing a drop in velocity at 75 mph. Two changeups were thrown, both bounced in the dirt for balls.

Romero faced only three batters in the fourth, throwing 13 pitches (nine strikes) in the bounce-back inning. He once again worked almost exclusively off the fastball, with 11 thrown between 85 and 89 mph. I counted five misclassified fastballs this inning, as Romero’s velocity drop continued to terrorize the PitchFX software. The two off-speed pitches he threw were very impressive, both coming in the at-bat against Casey Kotchman. The first was an 0-2 changeup in the dirt that the batter wisely held back on. The curveball was the very next offering, clipping the inner part of the plate on its way to Arencibia’s glove. The pitch was wrongfully declared a ball instead of strike three by the home plate umpire, but Romero managed to induce a ground ball to short only two pitches later to get the out.

In his fifth and final inning, Romero once again retired Cleveland in order. Learning from his earlier mistake, Romero fed Hannahan nothing but fastballs, with the plate appearance ending on a spectacular diving catch in center field by Colby Rasmus. Michael Brantley saw the only off-speed pitch of the inning, taking the 76 mph curveball in the dirt for a ball. He grounded out to second base on the next pitch. Asdrubal Cabrera was Romero’s last batter of the game, and he was punched out on four fastballs. Of the 13 pitches (9 strikes) thrown, 12 were fastballs, ranging between 85 and 90 mph.

I was really hoping to see Romero be at his best -– and throw more than five innings -– but today’s game was perhaps even more interesting for two reasons. Prior to this exercise, I had never really put too much thought into the classification scheme of the PitchFX system. I knew how it worked (the basics, at least), but I never considered how one agonizing inning that causes a temporary velocity drop could throw the system off so much. For me, this only further emphasizes the need to inspect data for outliers when using statistical sites like FanGraphs to make an argument. It was also interesting to closely examine a start in which the pitcher didn’t have his best stuff, and see how he adapted to the situation. Romero failed to establish first pitch strikes early on in the game, and it cost him in the long run. With far too much frequency he fell behind 1-0, immediately giving the hitter the advantage. Furthermore, the constant two and three-ball counts forced Romero to try and locate his off-speed pitches within the strike zone to avoid the walk. Romero is a much better pitcher when he’s able to establish the fastball early in the count and then work the changeup and curveball into 0-2 and 1-2 situations, where he can target swinging strikes or weak contact outside the strike zone. In Thursday’s game, after seeing his off-speed pitches get lit up in the second inning, Romero threw 85.7% fastballs (36/42) over the final three innings, recovering both his command and confidence.

For anyone interested in the general data from Romero’s start:

Total pitches: 96 (55 strikes, 41 balls)
Fastballs: 75 (78.1%) –- 45 strikes, 30 balls -– 85-94 mph, averaged 89.2 mph
Curveballs: 11 (11.5%) –- 5 strikes, 6 balls (5 in dirt) –- 75-80 mph, averaged 77.1 mph
Changeups: 10 (10.4%) –- 5 strikes, 5 balls (4 in dirt) –- 80-85 mph, averaged 82.0 mph
Sliders: 0 (0.0%) –- N/A
First pitch: 11 strikes, 10 balls –- (20 fastballs, 1 changeup)

Like what you read and want to stay informed on all updates here at Jays Journal? Follow us on Twitter (@JaysJournal), “Like” our Facebook page, or grab our RSS feed!