Tuesday, April 8, 2008

The Two Most Interesting Pitchers in Baseball, Part 1

There's nothing interesting going on with the Mets right now - the home opener is about two hours away and they'll have the chance to wash the taste of two straight losses to the Braves out of their mouths. Instead, I'd rather write about a couple of pitchers who are among my favorites in baseball right now.

One is a Yankee, the other is a former Met. Both are right-handed starting pitchers who are succeeding in the American League despite being unique statistical outliers. In fact, it is nearly impossible to find more than a handful of pitchers with a similar profile that have had long-term success on the major league level.

They are Chien Ming Wang and Brian Bannister, and every start they make has the potential to change the way pitchers are evaluated in the future.

Wang is more familiar to casual baseball fans - he has won 19 games in each of the last two seasons and is generally considered the ace of the Yankee pitching staff. One of only five major leaguers in the game's history to hail from Taiwan, he features a mid-90s fastball and a devastating sinker that's very hard to elevate. Incredibly, in this age of smaller ballparks and bigger sluggers, Wang has given up only 21 home runs in the last two years - in a span of over 400 innings.

Bannister is less recognizable, but has become a darling of the sabermetric circle for some of his thoughts on pitching and his success. His breakout season was in 2007, when he won 12 games for a bad Royals team and finished third in Rookie of the Year voting. The son of former major leaguer Floyd Bannister, Brian does not have the same pure stuff as Wang. He relies on locatiing a high 80s fastball and pitching to contact - generally considered a disatrous game plan for right-handed pitchers.

In fact, both Wang and Bannister share a statistical trait that often spells doom for pitchers - a low strikeout rate. Strikeouts - especially in terms of rate statistics like K/9 (strikeouts per nine innings and K/BB (strikeout to walk ratio) - are considered by some to be the gold standard for measuring a pitcher's success. Wins and losses are almost completely dependent on run support and are notoriously poor indicators of quality. ERA and WHIP are much more useful, but both are influenced by the fielding ability of the players behind the pitcher and even the ballpark the games are played in.

For a pitcher, the ability to get an out without involving your fielders means that you are leaving nothing to chance. Poor fielding range, errors, even freak occurences - they are all neutralized by the K. For this reason, advanced statistical analysis has placed a great deal of value on the strikeout - and rightfully so. Strikeouts are not only a measure of success, but they are also a measure of dominance.

Which brings us back to Wang and Bannister. Generally, successful pitchers strike out at least six batters per nine innings (the elite pitchers usually get closer to seven in nine) and have a K/BB ratio of at least 2 to 1.

In 2006, a season in which he finished second in Cy Young Award voting, Wang had a K/9 ratio of 3.14. Last year he improved that number somewhat - all the way to 4.69. In each of the last two seasons, Wang's K/BB ratio was less than 2 to 1 (although 2007 was again better than 2006). Now unfortunately, I don't subscribe to Baseball Reference's Play Index - only the single most amazing statistical analysis tool on the Internet - but I can tell you that Wang's previous two seasons are nearly unprecedented, in the sense that he won so many games and had such a good ERA despite his incredibly low strikeout rate.

Bannister is in a similar boat. In 2007, his K/9 ratio was just 4.2 and his K/BB ratio was 1.75. The are generally harbingers of doom, especially for a pitcher whose "stuff" is considered not nearly as good as Wang's.

So how do we explain these anomalies? The simplest answer is luck. Bannister got lucky last year and Wang has been lucky the last two seasons, and any day now they will both end up in the crushing grip of reality, relegated to the back of their respective rotations and muddling along with 5.00 ERAs. A terrific statistic called BABIP (batting average on balls in play) certainly supports this theory.

Last season, both Bannister (.262 BABIP - third lowest in baseball) and Wang (.296 BABIP) had "below average" numbers in this category. Generally, a BABIP under .300 has come to be considered a somewhat "lucky" performance and one can generally expect an increase in ERA and WHIP the next season. Wang has already beaten the system once - his .267 BABIP in 2005 has only risen into the .290 range in each of the last two years while actually reducing his ERA.

If Bannister's BABIP rises from the seemingly unsustainable level he achieved last season, the results could be troublesome to Royals fans. However, in a series of interviews this off-season, Bannister postulated that he could sustain his BABIP levels by following a specific game plan of pitching to contact and controlling the count. This embrace of sabermetrics is incredibly interesting and has the power to sustain Bannister's career, if he sticks to this plan and the rest of the league refuses to adjust accordingly.

Even as late as last season, I would've told you that Chien Ming Wang was destined to be a fourth starter and that Brian Bannister will never pitch regularly on the major league level. Today, I look forward to each time one of them takes the mound and I've changed my thinking on the importance of strikeout rate statistics. More on that tomorrow.

1 comment:

tim said...

Bannister beat the yankees today, doing what he does best. If that team can hit for him, he can win 15 this year, what a terrible move. I said it when Omar traded him and I'll say it now: Omar along with Randolph are on notice. I know, I know it's early. But it was early last May when the Mets were dropping games that would have won them the pennant, so it's never too early. Every game matters and these little decisions cost us games and ultimately pennants.