Kevin Draper reports an interesting tidbit about Matt Williams, the just-fired Washington Nationals’ manager. Williams was the focus of constant criticism during his two-season tenure at the helm in DC for, among other things, his (mis)handling of his bullpen. Reportedly, San Francisco manager Bruce Bochy was “giddily incredulous” when, during the deciding game of the 2014 Division series between the Nats and the Giants, Williams brought in a rookie relief pitcher during what turned out to be the game-deciding 7th inning. Williams had at his disposal better options, including his ace Stephen Strasburg (available because it was a potentially season-ending game) and his all-star closer, but Williams’ rigidity about bullpen use mandated that he not avail himself of those.
Draper notes that Cardinals’ manager Mike Matheny made a similar mistake last year – failing to use his best relief pitcher in a critical situation because the “book” dictated that he not do so – when his team played the Giants in the League Championship Series.
All of which reminded me of the maddening decision by Yankees’ manager Joe Torre, in Game 4 of the 2003 World Series against the Florida Marlins. The Yankees were up two games to one, and Game 4, played in Florida, entered extra innings. Torre, of course, had his all-world closer Mariano Rivera sitting in the bullpen. But for all sorts of silly reasons, the “book” says that you don’t bring your closer in to tie games on the road. (the “book” apparently thinks that bringing in your designated best reliever when your team is up by three runs in the ninth inning is a more urgent situation than bringing in your best relief pitcher in a tie game whose outcome is the definition of uncertain).
So, Torre tabbed Jeff Weaver, a starter relegated to the bullpen for the postseason because he wasn’t very good, to pitch the 11th inning. When the Yankees failed to score in the top of the 12th, Torre still had the option to bring in Mo for the bottom of the 12th, when Mo could have pitched a minimum of two innings, assuming he got through the 12th unscathed, and afforded the Yankees another opportunity to bat in the 13th. Instead, he ran Weaver back out there and the predictable happened – he gave up a game-winning homerun to Alex Gonzalez leading off the inning, giving the Marlins the win. The Yankees did not win another game in the Series and the next time Rivera got to pitch, the deciding Game 6, the Yankees were already on their way to final defeat.
In other words, Torre saved Rivera for a situation that never came, while passing on using him in the most critical innings of the Yankees’ season.
As Draper says, there’s been some slow progress in applying analytics principles to relief pitching in baseball. But it remains an area oddly resistant to simple common sense (aka using your best players when it matters most).
By a fluke, the save rule was amended four decades ago to include situations in which a reliever who enters at the start of the inning in which the opposition is trailing by three runs and preserves the win by recording the final outs, even though the tying run is not at-bat or on deck at that point (the condition for saves otherwise). So, managers are holding their designated best relievers, among other things, for those situations. Had the rule been written more consistently, only protection of a two-run lead in a game in which the entering pitcher finishes would qualify as a save situation. And were that the case, managers would surely not regularly be using their closers in those circumstances. Just as managers don’t typically bring those closers in to protect four-run leads now.
In other words, whether they recognize it or not, manager are deploying their relievers in part because of a statistic whose application has no bearing on the outcome of the game.
That this is still going on in 2015 is utterly bizarre.
(Torre, I should note, was generally really smart about how he used his bullpen in the postseason, recognizing that October baseball required thinking differently about deployment of relief pitchers than does the regular season. Game 4 in 2003 was a notable, and horrendous, exception).