Tuesday, October 20, 2009

Bad Ideas Even If They Work

There have been a lot of things to second-guess in the playoffs already this year, but I want to focus on two of them here. In the aftermath of Joe Girardi's disastrous decision last night to bring in a reliever--with two outs and nobody on base--who subsequently gave up two straight hits to lose the game, Girardi's defenders (starting with the manager himself in his post-game remarks) simply state that "it was the right move, but it just didn't work out." His critics note that the pitcher Girardi removed, David Robertson, had just retired two batters with ease and appeared to be throwing very well, so with no crisis looming there was no reason to think that he needed to be replaced by another right-hander.

I'd like to take the criticism one step further. I've done a ton of research on relief pitching, and one conclusion that keeps jumping out from the data is that it is pointless to remove a pitcher who is effective. One study I've done looks at relievers who enter in the 8th inning and don't give up any runs (either their own or inherited runners). In other words, they have worked successfully. Until the mid-1980s, such a pitcher was allowed to start the 9th inning more than 90% of the time. Today, that figure is down around 12%. Managers have developed a mind-set that certain pitchers are to be used in certain roles and situations regardless of what has come before. But my study shows that the success (i.e. save) rate was a little better for the pre-1990s relievers who pitched the 8th and 9th innings than it is for the bullpen committees so popular today.

The logic behind this is simple. Pitcher A demonstrates that he is throwing well by getting batters out--in last night's case, the first two batters Robertson faced. Pitcher B is an unknown quantity. He might or might not have his good stuff. If he does, he is likely to be as effective as the pitcher he replaced. If he doesn't, he is likely to blow the game, as Alfredo Aceves did last night, surrendering two sharp hits and the ballgame.

This is hardly the first time this month that a manager has jettisoned an effective pitcher and subsequently lost the game. In Game 2 between the Dodgers and Cardinals, the Cardinals led 2-1 heading to the bottom of the 9th inning. Adam Wainwright pitched a terrific game, allowing only one run on three hits in eight innings. In the 8th, he loaded the bases with two outs but Tony LaRussa left him in, and he responded by sawing off Matt Kemp and getting an easy out to end the threat. Somewhere between being sharp enough to retire Kemp and walking to the dugout, Wainwright evidently ran out of gas, because LaRussa replaced him to start the bottom of the 9th. Trever Miller retired one batter and hit the pines in favor of a slumping Ryan Franklin. Minutes later, the Dodgers scored two runs off Franklin to win the game, helping send the Cardinals home early and leaving Wainwright to think he could have done better if given the chance.

Watching last night's game, I marveled at how announcer Tim McCarver continues to live up to the worst expectations of baseball fans who soured on his over-analysis years ago. He made a big deal of Girardi's visit to the mound with two outs and two strikes, trying very hard to figure out what Girardi was telling Andy Pettitte. His conclusion, stated several times initially and repeated a couple of innings later when he revisited his point, was that Girardi was reminding Pettitte that he should "expand the strike zone" against the batter, Vlad Guerrero. This is the same Vlad Guerrero they made sure to show footage of when he hit a pitch on one hop. It's hard to think of a batter from the past 20 years who swings at more pitches outside the strike zone than Guerrero does, with quite a bit of success. Did McCarver really think that Pettitte (a borderline Hall of Fame candidate) needed to be reminded that it wasn't essential to throw the ball in the strike zone in order to get Guerrero to swing?

Andy Pettitte has been in this kind of situation more times than Girardi has. Whatever Girardi said to him, I can't believe he made a trip to the mound (with his pitcher one strike from being out of the inning) to point out Guerrero's habits. Whatever Girardi said, I think all the visit amounted to was a big distraction of Pettitte's concentration. He was focused on getting that last strike. Here came Girardi to complicate things. A moment later, Guerrero took a fastball just off (or just on) the inside corner and drilled it into the stands for a game-tying two-run home run.

On the other hand, when Girardi brought Aceves in to replace Robertson, both McCarver and Joe Buck criticized the move. As I did at home. It isn't second-guessing when you criticize the move before you see the result. McCarver echoed my logic above: why remove a pitcher who's working effectively? But I have a bigger problem with the move: even if it worked, it was pointless. This was the 11th inning of a game following a 13-inning battle. Girardi had only two pitchers left in the bullpen, Aceves and Chad Gaudin, primarily a starter this season. Because of an earlier maneuver, the Yankees had no designated hitter, meaning their pitchers would have to bat. He had already used one pinch-hitter to bat for Mariano Rivera.

My point is that even if Aceves had retired Howie Kendrick instead of giving up a single, Girardi would still have been in worse shape than if he had let Robertson pitch to him. He needed to conserve his pitchers in that spot. Giving himself one less pitcher gave him one less option as the game went on and put him that much closer to using his last pitcher and having no flexibility at all. It's one thing to match up pitchers (lefties against lefties, as in the Trever Miller move discussed above) in a crucial situation. But two outs and nobody on base is not a crucial situation, and neither reliever had a significant enough history against either of the Angels hitters due up to make a clear case for thinking Robertson was a liability. There was going to be a more dangerous spot further on, when Robertson was either tiring or due to face a batter who could win the game with a hit, when Aceves (or Gaudin) was an obvious upgrade on the mound. Or the innings would pass smoothly, Robertson would need to come out for a pinch-hitter, Aceves would push as far as he could, and then Gaudin could come in.

If it ain't broke, don't fix it. Last night's game in Anaheim was far from broke when Girardi decided to fix it. Then it broke into little pieces. But even if Girardi's fix had worked at the moment, it still would've left his team in a weaker position in the game.

Another kind of ill-advised decision occurred in Game 2 between the Yankees and Angels, and it probably cost the Angels the game. Situation: bottom of the 13th, tie game, one out, Yankees on first and second. Melky Cabrera hit a sharp grounder in the hole between first and second. Angels second baseman Maicer Izturis raced to his left and made a nice grab of the grounder, then pivoted quickly to throw to second for the force. The throw sailed wide, rolled away, and the runner scooted around from second with the winning run.

The throw stank, but not as much as the decision to make it. Suppose Izturis had gotten the force at second. What would've happened. With runners on first and third and two outs, the Angels would not have held Cabrera on first. His run didn't matter, and the defense would have let him take second base. It happens all the time. In other words, the Yankees would have had runners on second and third within a pitch or two--exactly what they would have had if Izturis had been sensible enough to throw to first.

There was absolutely no reason for Izturis to attempt the force play. That might be important in the early innings of a game, when you want to keep runners out of scoring position. That might also be more acceptable if there were no outs instead of one, because the force would keep open the possibility of a double play. But Izturis was moving toward first base when he fielded the grounder. It would have taken little effort to get the out at first. Instead, he made a pivoting (wild) throw to second--when there was nothing at all to be gained by a force play.

My inclination is to say that Izturis hadn't thought things out in advance and that it was a bad decision in the heat of a panicky moment. On the other hand, thinking things out in advance didn't do Girardi any good. Of course we don't know what would've happened if Izturis had gotten the out at first or if Robertson had stayed in the game. The outcomes might have been the same. But that won't change my view: the wisdom of a decision sometimes hangs on the issue of whether a successful short-term result actually gives you a better chance of winning.

No comments: