October 25, 2003

Prospectus Today

Starting Beckett: The Right Call?

by Joe Sheehan

It's official: Josh Beckett will start Game Six for the Marlins on three days' rest, with Carl Pavano scheduled to do the same if the Fish can't put the Yankees away tonight.

I strongly disagree with this decision. It's a move you make when you're down 3-2, not up 3-2. It's a decision you make when the difference between your best pitcher and the rest of the staff so large that going with anyone else in Game Six almost guarantees a Game Seven.

Neither of those apply here. The Marlins need to win just one game to be champions, and they don't get style points for winning in six. The Marlins have at least one pitcher available, in Mark Redman, who was arguably their #3 starter during the season. They certainly have Dontrelle Willis available for at least a few innings, and Willis was lights-out for a good part of 2003 and has been tough on Yankee lefties in this series. Frankly, outside of Game Five starter Brad Penny and Beckett (assuming you hold him back), the Marlins have nine pitchers who can give them at least a couple of innings, and some of those are the better pitchers on the staff. It's Pavano's throw day, so even he can give the team a couple of innings. Deciding that you'd rather start someone--two pitchers, actually--on short rest rather than use those guys is an inexplicable vote of no confidence.

The recent track record of pitchers starting postseason games on short rest, covered by Rob Neyer a couple of weeks ago, is just the most prominent reason to hold Beckett back. The best pitchers in the game are the ones asked to do this, and the record shows that they become, collectively, replacement-level starters. Beckett was a hero out of the bullpen on short rest against the Cubs, and the memory of that outing is no doubt easing McKeon's mind about using him tonight. Nevertheless, the evidence that using pitchers on three days' rest is a self-defeating strategy is considerable, and the Marlins are going to fly in the face of that evidence not just once, but twice in the next two days if they lose Game Six.

While we're focusing on Beckett tonight, this decision really stems from McKeon's desire to undo a key mistake he made last weekend. Starting both pitchers on short rest is the only way McKeon can have both him and Pavano start the last two games of the series. Pavano, who probably should have started one of the Series' first two games, started Game Four and pitched very well, just as he did in Game Six against the Cubs. Had Pavano started one of the earlier games, he would have already gotten two starts in the series, and McKeon wouldn't be trying to get him a second.

Beckett starting Game Six is just the consequence of wanting Pavano to get one more start. It's a symptom, not the disease itself. Given that Pavano would likely be available as part of an "everyone pitches" strategy tonight--it's his throw day, the same day in the cycle on which Beckett made his relief appearance--the gain here is minimal. McKeon is exposing the Marlins to a big risk; from 1995-2002 starters working on short rest went 6-15 in 32 starts, with an ERA of 5.20.

Let me put it this way. Would you rather have Pavano available for 40-60 pitches on two days' rest, and Beckett starting a potential Game Seven on full rest, or both Beckett and Pavano pitching on three days' rest? What the Marlins are basically doing is trading a fully-rested Beckett for the difference between a three-days'-rest Pavano and a two-days'-rest Pavano. Given what we know about pitchers starting on short rest in the postseason, the quality of the other Marlins' pitchers, and, frankly, the fact that it's Carl Pavano, I have to think the decision reduces the chance of the Marlins winning the one game they need to become champions.

Jack McKeon is moving heaven and earth so he can start Carl Pavano on short rest. When you frame the decision that way, the absurdity of it becomes clear.

Let's get something out of the way: this isn't about "abuse." I don't think starting Beckett tonight is a bad idea because of the potential damage to his arm. His workload this season and off-season has been reasonable, and there's no evidence that starting a random game on short rest is problematic. I would be concerned by the fact that Beckett has pitched more, and longer, this year than in any that came before it, and you're never quite sure when that will start affecting his performance. It's another reason to be conservative and to give Beckett the best opportunity to succeed.

This move may work out. Beckett could throw another two-hit shutout tonight. The Marlins may pound Andy Pettitte into submission and make the decision irrelevant. But the outcome won't change the caliber of the decision; it's a poor risk, and the second major mistake--not starting Pavano in Game One or Two was the first--that Jack McKeon has made in this postseason.

So what will happen? The Yankees have looked terrible at the plate in this postseason, but if they're going to be beat anyone, I have to think their chance will come against a right-hander who occasionally has trouble with his command. The Yankees have to treat facing Beckett, who while not on a pitch count, can be expected to tire a little earlier than usual, the way they've treated facing Pedro Martinez over the years: take a lot of pitches, make Beckett throw strikes, and try and keep him to six innings or less.

For the Marlins' part, they get one more crack at a left-handed starter. I've spent a lot of time writing about how well their lineup hit lefties all season, but the fact is that they haven't hit the Yankees' lefties in the World Series. Tonight they face Pettitte and his postseason mystique, which held them to one run in 8 2/3 innings in Game Two.

I genuinely have no idea what will happen tonight, so for the first time this postseason, I will eschew a prediction. The only thing I'm reasonably sure of is that the starters will not get decisions.

(Go Yanks.)

Joe Sheehan is an author of Baseball Prospectus. 
Click here to see Joe's other articles. You can contact Joe by clicking here