This guy has a very interesting argument. In case you are dying to read my knowledge-drop but don’t want to wade through the plebeian blather of "The New York Times,” basically the argument this reader of Freakonomic posits is that closers should start games and hand off to a starter who would finish up the game, thus letting the starter focus on 6-7 innings and not pace himself as much while getting consistent outings from their best short reliever, the “opener”. Where he misfires:
- Using relief pitchers after the starter allows the manager to manage a scarce resource (relief pitchers) within a (generally) known number of innings (i.e. from when the starter depart through the 9th inning with something to spare in case of extras). What if your closer falters? Bring in your starter early? Or you could use a few relievers, then you starter, running the risk of your starter also having a bad day and being short-handed at the end of the game. Let's imagine the Nationals adopt this approach and “opener” Matt Capps gives up 3 runs in 2/3 of an inning, he’s replace by a lefty who closes out the inning, then the next guy struggles to get through the second and you’re down 6-0 starting the third and handing the ball off to your best starter. That’s a waste. Or you hand it off to your 5th starter; he gets shelled but in the 6th inning you’re still in a 9-9 game without your “opener” and lefty specialist available. The conventional approach allows the manager to better manage his pitchers.
- His change of pace argument actually is a better argument for the current pitching structure. As the starter tires and batters have seen the same pitcher 2, 3, or 4 times, that’s when you are most in need of bringing in Daniel Bard to relieve Wake, rather than letting someone in the 9th inning face Wake for the third time when he's getting tired.
- In this case, you’d still need a closer of sorts. When a starter falters but the game is still close, it leaves you with a lack of a good option for the 9th since you’ve already pitched your best non-starter early in the game, or you’re saving him as the next game opener.
- I agree with the writer’s fascination with knuckleballs and the idea that they are underused. If a pitch is so tough to hit – and catch – that major leaguers can know it’s coming and still not hit it, then people should toss it in every now and then if they can pull it off. That said, the talent and effort it takes to throw a knuckleball is also under-appreciated. You couldn’t rely on one to throw two innings every other day all season. Their arm would fall off. Secondly, it’s too difficult to throw a knuckleball for us to think we’ll ever (again?) see every team have one.
- I feel that this is a case of trying to fix something that is not really broken. If a manager doesn’t want his pitchers to aim for complete games, tell them that. Tell them you need six good ones. This is a massive risk as a way to solve a non-problem.
When I try to think about how this might look, I can’t help thinking that a manager wouldn’t be willing to remove an opening pitcher after just two innings if he was pitching particularly well. As any Red Sox fan can attest, there are games– and they may be few and far between, but there are games– when Tim Wakefield’s knuckleball is practically untouchable. In situations where that’s the case, why remove him from the game after two innings or a predefined low pitch count, only to introduce a lesser-known variable– another pitcher’s “on-ness” on this given day– into the situation?
A likely result, then, of a team implementing an opening pitcher strategy would be an inevitable regression to the current starting pitching strategy. Once a pitcher has started pitching, a manager’s decision making becomes inherently more informed: before the game began, you were placing a bet on this guy’s performance based on his recent outings or history against a certain lineup, but after the game begins, you immediately have much more information. Does that knuckler look good today? Are guys connecting with it, even if they’re flying out? Is Wake’s beet-red neck sweating too profusely? And if your guy looks good, then you know you’re more likely to have continued success with him– after all, he’s pitching well now– than you are to have success with someone else.
Well, yes, but only to a point. And there’s the rub. Wakefield’s knuckler will start to hang sooner or later in most cases. If you cut him off early, you are preserving the non-outs you otherwise would have had to give up to find out precisely when that time will be. That seems to be the crux of the argument: that by reducing your odds of seeing an opener fail by limiting him to a couple of innings, you are effectively doing the same thing for the person pitching the middle innings. It’s just a smarter distribution of risk.
So I guess the question really is this: at what point do the odds of your opening pitcher continuing to pitch well become worse than the odds of a “starting” pitcher doing as well?
That, I think, is a question too complicated to be addressed by a strategy that’s defined by holding an opening pitcher to an arbitrary number of innings or pitches. That inflection point would probably come at a very different time for different pitchers (or, more accurately, for different pairings of pitchers) on different days. The opening pitcher idea seems valid, but much more difficult to implement smartly than by using predefined time periods.