Wednesday, June 2, 2010

The Advantages of Being Unique

Tim Wakefield is just about as unique as you can get in terms of a career path. He's been waived by the Pirates in his 20's, posted solid seasons starting and in the bullpen for the Sox, including one extraordinarily nerve wracking stretch as closer, and made his first All Star Game in his 40's. Throughout most of his career, Wakefield has been one of the only knuckleballers in the league. In the first half, you did have Steve Sparks in Detroit, and in the 2000's you had the occasional R.A. Dickey sighting, but hasn't been another established knuckleballer, particularly in the National League. At this point, Wakefield has been in the league long enough that most players in the AL have faced him a few times and at least have an idea of what they're getting into. However, because Wakefield was the only knuckleballer National Leaguers were facing, before this year with Dickey's resurgence and the rough start that Charlie Haeger has had in LA, I wondered - did National Leaguers fare worse against Wakefield because of the rarity of his pitch selection?

To start out, I limited the search to 2005 to 2009. I didn't include this year because I wanted to avoid confirmation bias by including Wakefield's marvelous start against the Phillies. Over that stretch, Wakefield made 138 starts - 119 against the AL and 19 against the NL. Against AL opponents, Wakefield was slightly worse than the average AL starter in every single category.

Wakefield Avg AL Starter
K/9 5.62 6.16
BB/9 3.13 3.00
K:BB 1.80 2.05
HR/9 1.23 1.12

Except, of course, in ERA and WHIP. Wakefield's ERA for 2005 to 2009 was slightly better than the AL average starter (4.61), as was his WHIP. The reason for Wakefield's success is quite simple, while most pitchers give up a batting average on balls in play (BABIP) of about .300 and have very little control over that number, knuckleballers give up a lower BABIP, presumably due to weaker contact. Unsurprisingly, Wakefield's BABIP for this five year stretch was .258, compared to .307 in the AL for the same time frame. Those extra few outs make a difference, allowing Wake to succeed despite unimpressive underlying numbers. Now that we understand how Wakefield succeeds, how does he do against the NL? The answer? He does awfully well. For comparison's sake, I've left Wakefield's AL numbers in the table.

Wakefield vs AL Wakefield vs NL Avg NL Starter
ERA 4.54 3.66 4.46
WHIP 1.31 1.17 1.39

In 19 starts against the NL, Wakefield was transformed from a league average pitcher to a great #2. It is always nice to see your observations borne out by the data, but why did Wakefield do so much better against NL opponents? Were they just making weaker contact, resulting in an even lower BABIP? Striking out more often?

Wakefield vs AL Wakefield vs NL Avg NL Starter
K/9 5.62 5.61 6.47
BB/9 3.13 2.54 3.19
K:BB 1.80 2.21 2.03
HR/9 1.23 0.82 1.09
BABIP 0.259 0.251 0.303

As you can see from above, the answer is neither. Both his K rate and BABIP are essentially the same, instead, National League batters are walking less often against Wakefield and hitting fewer homeruns. Some of this might be attributable to facing the pitcher or small sample sizes (a pitcher can definitely get lucky home run wise in half a season's worth of starts), but I would guess it is a change in approach. NL players who are unfamiliar with the knuckleball may be more aggressive, causing a decrease in walk rates. Looking at pitch per plate appearance data would be one way to nail this down, but unfortunately that isn't readily available to me.

Others have done some really cool work on the knuckleball using PitchFX data, such as this article by John Walsh at Hardball Times from a few years ago.


  1. Dear Blogger,
    Why do you eat my God damned table formatting?

    Mr. Bill

  2. Dear Chrome,
    Why do you, the Google browser, strip the formatting out of my tables in Blogger, the Google blogging website? Firefox and Blogger get along just fine.

    Mr. Bill