Changeup keying Silva's 8-0 record

Second FanGraphs post today ... Jack Moore's postmortem of Carlos silva's eighth victory:

    Silva threw 102 pitches in today’s start, 55 of which were classified by Brooks Baseball’s Pitch F/X Tool as either fastballs or sinkers. Both pitches have very similar spin, movement, and velocity, so I think it would be fair to denote all 55 as “fastballs.” With that clarified, 53.9% of Silva’s pitches today were fastballs, well in line with his 57% fastball rate this season. That’s a stark change from the rest of his career. Silva never had a fastball rate lower than 68% entering this season, and in four separate seasons he used his fastball over 80% of the time, including his disastrous 2009 season.

    It’s not terribly surprising, then, that his strikeouts have risen to above 6 K/9. Even with below average secondary stuff, the addition of more changeups and sliders is certain to increase the amount of whiffs. This change is likely the agent behind a nearly 3 point jump in swinging strike rate from his career mark. In particular, it’s his changeup which has been fantastic this season. It has been a strike over 70% of the time, and even though it’s nothing special in terms of whiff rate (12.4% vs. 12.63% average), it still manages to get outs, as the pitch had a +8.8 pitch type linear weights value entering play today. Brooks Baseball had it as excellent once again today, with a +2.2 mark.

Including this season, Silva's posted excellent ERAs twice: this year, and in 2005.

In 2005, he did throw his fastball more than 80 percent of the time: 84 percent of the time, almost exactly the same as he did last year. I'll also mention that Silva's repertoire (according to Pitchf/x looks almost exactly the same in 2007 as in 2008. His fastball/slider/changeup percentages in those two seasons, with ERA:

68/9/23 - 4.19 (2007)

69/9/21 - 6.46 (2008)

Actually, in 2006 Silva had started throwing fewer fastballs and didn't go back to throwing a million of them until his abbreviated (30 innings) 2009, and that might be a sample-size issue.

Still, his fastball frequency is obviously far lower this season than it's ever been, and he's throwing more changeups than ever before (and slightly more sliders, too). Oddly, his changeup has never been a good pitch; usually it's been a bad pitch for him. And yet this year it's a great pitch. Go figure.

There is (as you might guess) one real similarity between Silva's 2005 season and his 2010 season: luck. In Silva's career, he's given up a .311 batting average on balls in play (batted balls minus home runs). In 2005, though, his BABiP allowed was just .293, which wasn't crazy-lucky but does help explain -- along with his absurdly low walk rate -- his 3.44 ERA.

This year he's given up a .262 batting average on balls in play, which as any schoolboy can tell you just isn't sustainable. Regardless of which pitches he's throwing and how well he's throwing them, Silva's BABiP will probably regress to roughly .300 from now through the end of the season. And his ERA is going to rise concomitantly.

Not that there's anything wrong with that. If he's healthy, Silva's going to finish the season with 14 or 15 wins and an ERA in the low 4s. Pretty good for a guy who wasn't supposed to do anything at all.