Go Ahead, Blame Matt Williams. But Blame Him For The Right Things.
[As I finished writing this, I came across Adam Kilgore’s piece here, which also gets into the differences between regular season and playoff pitcher management.]
This post is not about the Nationals’ terrible, awful, no-good NLDS hitting, where the lineup (not including pitchers, pinch hitters, Rendon, or Harper) hit .112 with an OPS of .316. This post is about Matt Williams’ criticism for his management of the pitching staff.
The critics have missed the moves that were the biggest mistakes. Williams’ errors were related to being stuck in a regular season mindset and/or not thinking far enough ahead.
The regular season is, in the short term, an unending grind. The use of relief pitchers is influenced by their future effectiveness; using a pitcher today and tomorrow reduces his effectiveness for days thereafter. These considerations around effectively managing relief pitching over weeks and months are critical to season-long pitching success.
And it matters almost not at all in the playoffs.
In the postseason, the high number of off-days (5 in a full 12-game NLDS and NLCS) means a relief pitcher can be used in a higher percentage of games without his effectiveness being impacted. A manager can also warm up pitchers more often with less damage to their effectiveness in later games.
There are two primary criticisms of Williams, related to the 2nd and 4th games of the series. The first is that it was a mistake to replace Jordan Zimmermann with Drew Storen in the 9th inning of Game 2 after Zimmermann gave up a walk to Joe Panik. Buster Posey then got a single on an excellent pitch, and Pablo Sandoval hit a decent pitch down the 3rd baseline to tie the game. Critics say Zimmermann should have been left in. They’re wrong.
Pretty much any seamhead will tell you that not only was removing Zimmermann the right move, but it should have been done earlier. The Giants were coming up on their 4th time through the order, and the crazy baseball stats people know that pretty much any relief pitcher will have better results than any starting pitcher on his 4th time through the order — with a large enough sample size.
But baseball’s funny. Clear statistical advantages have high variance. The correct move according to the stats will fail to work out a lot, sometimes spectacularly. Like in >Game 2.
But if replacing Zimmermann wasn’t a mistake, what do I think was?
At the time I thought the numbers favored pinch-hitting for Zimmermann when he came up in the 7th with 2 out and a runner on first. But I thought it was probably a pretty close call (it’s not), given that the chance of scoring at least one in that situation is maybe 15%. Given that low number, I didn’t think leaving him in there was a very bad decision. But leaving him in to start the 9th was a surprise, given that Williams has been very much in the “relievers pitch full inning increments” camp all season. And I worried about the 4th time through the order and figured he’d be out if anybody got on, with Storen warming in the bullpen.
See, that was the problem. Williams was in regular-season-mode, leaving it to his closer to finish rather than asking What If.
If he asked “What if Storen doesn’t get Posey out?” Perhaps Williams would have had Matt Thornton warming up next to Storen to pitch to Sandoval. Sandoval’s a switch hitter but has been terrible from the right side this year: BA/OPS of just .199/.563 from the right, .317/.824 from the left. When Storen entered the game, Williams should have asked “What if Posey and Sandoval load the bases?” and had Tyler Clippard start getting loose.
One hitter and out is not a standard use of a closer. But if Storen started that inning and two runners got on with Sandoval coming up, taking him out wouldn’t be controversial. Essentially, Williams gave part of Storen’s normal work to Zimmermann, and once the situation is two on with Sandoval up, a lefty should have been in.
In game 4, Williams has been criticized for using rookie Aaron Barrett and not Clippard, Storen, or even Stephen Strasburg. This criticism is generally correct (not the part about Strasburg), but once again the critics have missed the root error.
Williams shouldn’t have replaced Jerry Blevins with Thornton to start the 7th inning. Blevins, a lefty, had needed only 14 pitches to get 4 outs and had crushed lefties all year, and would have faced lefties Greg White and Joe Panik. This would have kept Thornton, the Nats’ last lefty, available.
Instead, Williams put Thornton in. Blanco bunted out and Panik singled. At this point, had he still been in the game, Blevins would’ve been replaced by a righty to pitch to Posey and Pence (who had been swapped with Sandoval in the order because the Nats starter was lefty Gio Gonzalez). Then Thornton would have been available to force Sandoval to hit righty and then to pitch to the lefties Belt and Crawford.
As it was, Thornton was left in to pitch to Posey (and presumably would have pitched to Pence had he gotten Posey out). Williams was either in “pitchers are most comfortable thowing whole innings” mode or he was hesitant to bring Barrett in against Posey, one of the league’s best hitters.
When Posey singled, Thornton had to come out and Barrett went in. The criticism of Williams going to Barrett is valid, and is bolstered by his assertion that he used Thornton and Barrett “because those are my 7th inning guys… We’re certainly not going to use our closer in the 7th inning.”
The playoffs are different. You don’t have standard guys for particular innings when the season is on the line. You don’t use Tanner Roark in the middle innings, and you don’t have 7th inning guys. You have match-ups, and the only way to be ready for those is to look far enough ahead.
The failure to look ahead was clear when I learned that Barrett was the only one warming up while Thornton was pitching, and nobody was up in the ‘pen as Barrett pitched to Pence and Sandoval. When the situation became more challenging, or more high-leverage, Barrett couldn’t have still been the correct choice. With two on in the 7th inning of a tie elimination game with the opponent’s #4 hitter up, the correct choice is your best matchup. Period. Maybe it’s your closer. Fine; if you fall behind there’s nothing to close anyway.
There’s not much history, but Pence is 0-5 against Clippard in his career, 1-5 against Storen, 0-4 against Soriano.
Once Barrett walked Pence, he should have come out. Williams surely would have brought in a lefty for Sandoval, Belt, and Crawford if he’d had one. Without one, he made a mistake leaving Barrett in; there can be no higher leverage situation, and you still have Clippard, Storen, and Soriano. And this mistake blew up with the wild pitch.
Incidentally, randomness is a big part of these managerial mistakes; sometimes they blow up, sometimes they don’t. And there are LOTS of mistakes in playoff games.
In Game 2 of the NLCS, Giants manager Bruce Bochy should not have left Jake Peavy in the game to face Matt Carpenter in the 4th inning with the bases loaded and two outs. But it worked out. Then in the 7th, Bochy left righty Jean Machi in to pitch to left-handed pinch-hitter Oscar Taveras, even though he had a lefty ready and the Cardinals had two more lefties due up after Taveras. Taveras homered to tie the game. These days Bruce Bochy is being lauded as a great manager, and he’s made these kinds of mistakes many times this postseason. Matt Williams isn’t the only one making mistakes; he just had a couple blow up on him.
The point is that these mistakes are avoidable. If Williams were to abandon regular season thought and start thinking ahead better, he’d make fewer of them.