From Yu Darvish‘s first start in the major leagues — before, for those who followed his international career — it was clear the 26-year-old phenom had the pure stuff necessary to handle major league hitters, even to thrive against them. The arsenal: a mid-to-high 90s fastball and five other pitches ranging anywhere from cutter to slider to lollipop curve, all with sharp break and solid control.
Darvish had his flashes of brilliance through the first four months of the season, as all with his talent inevitably do, but the results were disappointing: a 4.05 ERA through July with a slightly better FIP. Two issues plagued Darvish: walks, and an inability to retire left-handed hitters. Even as he struggled with his control, righties managed just a .648 OPS off Darvish in his first four months. Lefties — buoyed by 23 extra base hits compared to just 12 for righties — mustered a .721 OPS. With managers playing platoons constantly, Darvish actually faced significantly more left-handers (321) than right-handers (217). This was Darvish’s major trial.
On August 6th, Yu Darvish pitched his first game with Geovany Soto behind the plate. From then on, there was one major difference in Darvish’s pitch mix: more cutters. Darvish threw cutters just 12 percent of the time to that point but turned it into his second most popular pitch for Soto, thrown 28.6 percent of the time. The difference was even starker to left-handers — at 16.4 percent, the cutter was his third-most frequent pitch against lefties prior to Soto’s arrival. After, he threw it 36.4 percent, a plurality, over 11 percent more often than his fastball.
The difference is in the horizontal movement:
The cutter (orange) has a tendency, however slight, to break inside to left-handed batters, whereas the fastball strays outside at the last moment. With that in mind, observe the difference in Darvish’s location before Soto and with him:
The darker the square, the more frequently a pitch was located in that zone; views are from the catcher’s perspective.
Generally, Darvish stayed within the confines of the strike zone more often with Soto; specifically, he was more willing to challenge left-handed batters inside.
The impact? Left-handed hitters recorded a paltry .184/.273/.296 line against Darvish in August and September striking out 42 times in 143 plate appearances (29.3 percent) and walking 15 times (10.4 percent). The extra base hits were largely swept away — no home runs, 10 doubles and two triples for a .112 ISO.
Darvish avoided the inside corner against lefties like a plague before Soto came in. If Darvish was trying to avoid extra base hits, it makes sense — he allowed a .505 slugging on contact on pitches on the inner third all season long, a number that actually rose to .546 after Soto came on. But the key there is “on contact.”
When Darvish induces a swing on the inner-third, he gets a whiff over 26 percent of the time, compared to just 14 percent on the middle third and 15 percent on the outer third. Darvish is especially dangerous when he spots the ball in the lower inside corner (the lower-inside-ninth?), drawing whiffs on an absurd 40.3 percent of swings. This is the power zone Darvish (and many other pitchers) are afraid of — lefties league-wide slug .614 on contact in that zone. But it doesn’t matter against Darvish — his arsenal, particularly his cutter, is just too tough to be hit with any regularity.
It would seem Darvish’s passiveness — his desire to nibble — was the main driver behind his issues to start the season. By becoming more aggressive, Darvish showed why the Rangers were willing to toss nine figures at him this offseason: in his final nine starts (not including his Wild Card gem), Darvish struck out 76 against 19 walks in 64 innings, allowing a sharp 2.95 ERA.
Darvish’s stuff was always this good. With Soto guiding him, he showed trust in that stuff, using it to attack rather than nibble. The Rangers may have been bounced from the postseason in his last start, but they should feel confident they have a legitimate staff ace in Yu Darvish heading into 2013.
PITCHf/x data and first diagram via BrooksBaseball.net.