What's eating Carlos Silva's ERA?

Wow. A post by me that has nothing to do with Pittsburgh. Or North London.

However, it does have something to do with a team I am somewhat partial to, that being the Minnesota Twins. I've had family in the Twin Cities since I was born, and semi-regular visits expanded to as many as three trips a year from when my grandparents moved there in 1995 or so until my grandfather passed away on New Years Day, 2003. Over the years, I've attended many a game at the Metrodome, starting in about 1997. The Twins, being in the American League, are never a threat to the Pirates, plus they were the underdog for the first few years I watched them. Even now, they're exactly what I wish the Pirates were, being that they've got payroll constraints but still compete every single year because of competent management. Because of all this, I tend to root for the Twins whenever possible.

The big issue of late amongst Twins fans has been the exercising of starting pitcher Carlos Silva's $4.35 million option after a season in which Silva lost 15 games and had an ERA of 5.94. Even though he was a little better than league average in 2004 (14-8, 4.21) and much better in 2005 (9-8, 3.44), I've been told by multiple Twins fans that they are upset about the move because they're convinced Silva, like some pitchers who live or die with sharp control and the ability to summon ground balls seemingly at all, has finally succumbed to his low K rate and been figured out by American League hitters. While I don't know if this extremely small sample speaks for a much larger contingent of Twins fans, I am here regardless to examine whether or not Silva is worth the financial commitment the Twins have made to him for 2007.

The main basis for exercising Silva's option being a wise decision lies within the assumption that he'll come much closer to repeating his success of 2004 and 2005 than his disastrous 2006. In order to do that, we'd have to discern what exactly caused Silva's implosion this season. We'll start by first debunking the myth that Silva's regression was caused by the league "figuring him out." The easiest way to do that is to calculate Silva's batting average on balls in play (BABIP) over the past three years, and compare it to the team's overall BABIP during that span.

BABIP is calculated by first determining how many times a batter ended an at-bat by hitting the ball within the field of play with the intent of reaching base. Balls in play = batters faced - walks - hit batsmen - sacrifice bunts - strikeouts - home runs. BABIP, then is simply (H-HR)/BIP.

Here's the math, using stats from Baseball-Reference:

2004

Carlos Silva

869 batters faced
- 35 walks
- 5 hit batsmen
- 7 sacrifice bunts (I can't find numbers of sacrifice bunts allowed anywhere, so 7 is a very reasonable educated guess)
- 76 strikeouts
- 23 home runs
= 723 balls in play

255 hits allowed
- 23 homers
= 232 hits on balls in play

BABIP = 232/723 = .321

Minnesota Twins

6269 batters faced
- 431 walks
- 54 hit batsmen
- 40 sacrifice bunts (obtained by averaging the amount of sacrifice bunts executed by the other 13 American League teams)
- 1123 strikeouts
- 167 home runs
= 4454 balls in play

1523 hits allowed
- 167 home runs
= 1356 hits on balls in play

BABIP = 1356/4454 = .304


2005

Carlos Silva

749 batters faced
- 9 walks
- 3 hit batsmen
- 7 sacrifice bunts
- 71 strikeouts
- 25 home runs
= 634 balls in play

212 hits allowed
- 25 home runs
= 187 hits on balls in play

BABIP = 187/634 = .295

Minnesota Twins

6072 batters faced
- 348 walks
- 43 hit batsmen
- 40 sacrifice bunts (I didn't bother recalculating the average because I can't imagine it being so drastically different as to make any sort of impact in the final numbers)
- 965 strikeouts
- 169 home runs
= 4507 balls in play

1458 hits allowed
- 169 home runs
= 1289 hits on balls in play

BABIP = 1289/4507 = .286


2006

Carlos Silva

811 batters faced
- 32 walks
- 7 hit batsmen
- 7 sacrifice bunts
- 70 strikeouts
- 38 home runs
= 657 balls in play

246 hits allowed
- 38 home runs
= 208 hits on balls in play

BABIP = 208/657 = .317

Minnesota Twins

6066 batters faced
- 356 walks
- 36 hit batsmen
- 40 sacrifice bunts
- 1164 strikeouts
- 182 home runs
= 4288 balls in play

1490 hits allowed
- 182 home runs
= 1308 hits on balls in play

BABIP = 1308/4288 = .305


In each of the three seasons, regardless of his ERA, his BABIP has been between .009 and .017 higher than that of his team's, which shows nothing anomalous whatsoever. On the contrary, it proves that Silva was about consistent in giving up base hits in his awful 2006 as he was in 2004 and 2005. The fact that his BABIP is always higher than the team's is probably due to the fact that he, as an extreme ground-ball pitcher, has been victimized more than the other pitchers on the team by the fast infield caused by the astroturf at the Metrodome.

Furthermore, there haven't been any ominous trends with his strikeout or walk rates. If his K/IP rate was significantly decreasing while at the same time his BB/9 significantly increased, it would send up a red flag that Silva's collapse was indeed a case of simply no longer fooling hitters, but that isn't the case.

However, the one key statistic from Silva's 2006 debacle that was totally incongruous with his previous two seasons is home runs allowed. In 2004, he allowed 23 in 203 innings. In 2005, he yielded 25 in 188 1/3 innings. In 2006, he served up a whopping 38 home runs in only 180 1/3 innings. This alarming habit, when coupled with his declining GB/FB ratios (1.58 and 1.55 in '04 and '05; 1.29 this year), indicates that for some reason, Silva's pitches are not staying down in the strike zone. Either he's locating up in the zone to begin with, or the pitches are simply not moving the way they should. I'd have to watch game tape to have any idea for sure, but it's a solid theory.

Why that would be, I have no clue, to be quite honest. It could be mechanical. It could be medical. It could be both, i.e. lingering side effects from the knee surgery that cut short his 2005 campaign that are forcing him to alter his mechanics. Who knows. But Silva's still only 27, and if he's able to get rid of the home run bug and regain his excessive ground-ball tendencies (and achieving one will more than likely mean achieving the other), there's no reason not to expect him to perform at his 2004 and 2005 levels once again.

The mere likelihood of that happening should be enough by itself to warrant bringing Silva back. The main gripe is that a team as cash-strapped as the Twins needs to be making more shrewd moves financially, but the truth is the Twins would certainly not be able to find a pitcher at that price who would even have much of a shot at performing like Silva did in 2004/2005, let alone one who would only sign for one or two years. A pitcher like Gil Meche, who has barely even approached Silva's previous levels of success, is poised to receive a contract worth $7 or 8 million per season.

There isn't really too much in the way of interior options, either. Johan is Johan, but with Radke hitting the links full-time, the rest of the rotation would stand to be Francisco Liriano coming off a severe injury, Boof Bonser, and who-knows-what after them. Scott Baker? Matt Guerrier? Matt Garza? A team with pennant aspirations really can't afford that many question marks in the rotation. The Twins at least are aware of what Silva's capable of.

The verdict: Silva is a wise investment at $4.35 million. Terry Ryan makes yet another good choice. I wish I could say the same for Dave Littlefield.