Wednesday's 13-inning game was just one of many times when the A's high leverage bullpen management has been tested. The A's, by virtue of neither scoring nor giving up a lot of runs, have played a lot of close games. The question of using your closer on the road in a tie game has been much discussed, and so I will not rehash it here. Rather, I want to focus on another, more damaging, practice to which both Bobs have repeatedly succumbed.
On Wednesday, Devine pitched the 8th, Breslow the 9th, Fuentes the 10th, Wuertz the 11th, Ziegler the 12th. The practice of picking relievers to adopt certain innings is exactly the problem. As a result, Breslow wound up facing a RH batter with the game on the line, Fuentes wound up facing a RH batter with the game on the line, and later Ziegler wound up facing LH after LH with runners on base. How could the A's get far more from these guys?
Personally, I like the "4-out appearance" where a pitcher is summoned to finish an inning because the situation renders him most useful (that might be Ziegler to get a RH hitter, Fuentes to get a LH hitter, Bailey or Balfour to get a key out, period) and then pitches the following inning. A typical 4-out appearance might require a pitcher to face 5 hitters while a typical inning might require the pitcher to face 4 hitters. The difference is roughly 25 pitches instead of 20 and for most relievers, if they're fresh enough to throw 20 they can throw 25, and if they're too spent to throw 25 they shouldn't throw 20 anyway. It's not a huge difference. What is huge is to get the platoon advantage, or strikeout pitcher, or sinkerball pitcher, or simply your best pitcher, into a situation like "two on, two out" where he is most desperately needed.
Key principle: The last out of an inning is often a lot higher leverage than the start of the inning.
Now let's say you had a group of relievers who simply couldn't be asked to get 4 outs in their next appearance. You should STILL be putting these guys in to squash rallies rather than to start and finish innings. Why? Because then you can leverage your relievers best, by choosing them to best fit the current crisis. Bring in Fuentes to get a LH hitter with a couple runners on base and two out, then let him start the next inning, and with two outs, if the batter is RH you can go to Ziegler, or if the bases are loaded you can go to Bailey or Balfour, and so on.
In other words, having your "one inning" (three out) relievers get outs 3, 1, and 2 would allow you to select relievers to match the highest possible leverage. Instead, by picking innings you wind up throwing a Fuentes or Ziegler to "whoever happens to come up" and that's not a good gamble -- especially once you're in trouble and a single will beat you, and the matchup isn't a good one for you.
By having all these "one inning relievers" instead of "three or "four outs" relievers, the A's are consistently forcing their relievers to face bad matchups with runners in scoring position. On Wednesday, Breslow dodged the bullet, as did Fuentes, as did Ziegler once. Then the parade of LHs caught up to Ziggy -- though had Bailey come on with the bases loaded in the 13th, likely the game goes into the 14th and if he throws a scoreless 14th it goes into the 15th. Two extra innings to hit, all for the price of going to your best reliever for one out, because that's when he's most desperately needed.
My reminder to Bob Melvin: It's less important who starts a high leverage inning than who finishes it. Eventually, the inning won't finish at all.