Judge Matt Williams by his decisions, not his results

Joe Ross began the bottom of the 7th inning Saturday night with a 2-1 lead. After giving up a game-tying home run to Lucas Duda–who had already homered that game–Matt Williams’ critics pounced. He should have gone to the bullpen. Why is Ross still out there? What was he thinking???

Well. Sending Ross back out for the 7th inning wasn’t an entirely unreasonable decision. Sure, this was his third time through the batting order, but Ross’ pitch count was low at 84 pitches. He was cruising. He had only given up two hits all game. Of course, Ross was set to face Lucas Duda, and Duda had already homered earlier in game. But the bases were empty. It was reasonable to trust Ross just as much, if not more, than a reliever in that situation. 

Had Williams gone to the bullpen to start the 7th inning, and the reliever gave up the tying run, Williams would have been crushed by his critics for pulling Ross too early.

On Saturday night, detractors of Matt Williams criticized the results, not the decision-making. Of course, if Williams had a time machine he would go back and choose a different pitcher to start the 7th inning Saturday night (we hope). But based on the information available at the time, Williams didn’t make a mistake, at least not a grievous one.

Using this same standard, though, Williams did make some pretty huge mistakes in the 8th inning Saturday night. Williams decided to start the inning with Matt Thornton, when Drew Storen was rested and available (I wrote a little bit about that last night). While Thornton might have been a better matchup for the first two hitters that inning, he certainly wasn’t once Yoenis Cespedes came to the plate. With a runner on second and one out with Cespedes batting, there’s little doubt Storen would have been the better option rather than leaving Thornton in the game. Instead, Williams opted to intentionally walk Cespedes to set a lefty-lefty matchup with Thornton and Duda. Ordinarily, this would playing the percentages and therefore defensible, but there was information available to Williams at the time he made the decision which made it completely bat shit crazy.

1. Storen was still available. Williams chose not to use Storen at all that inning (indeed he didn’t use him all series). After Thornton grooved a fastball to Curtis Granderson to put a runner in scoring position, Storen should have been off his ass and throwing. If the Mets take the lead in the bottom of the 8th, the game is likely over. With the Mets two best hitters coming up with a runner in scoring position and the game on the line, that’s the time to put your best reliever on the mound.

2. Duda isn’t any less dangerous in that situation. Cespedes is a legitimate power threat, slugging close to .500 on the season. But Duda isn’t far behind (near .480 with a higher on base percentage than Cespedes). Moreover, Duda already had two home runs that game. Why would any manager choose to have him batting again–with two runners on base? Also, it wasn’t just Saturday. Duda hit six home runs that week. Duda is the one Met you don’t want batting with the game on the line right now, yet that’s the one Williams chose. 

3. There was no platoon advantage. This is the final nail in coffin. Williams presumably walked Cespedes to “play the percentages” but Cespedes actually has a significant platooon disadvatage against lefty pitchers this season (he was batting .183 to .318 vs. righties). His career numbers aren’t as dramatic, but the platoon split is still there. Duda, meanwhile has a platoon advantage against lefties this season (almost 75 batting average points). Duda has better numbers against righties for his career, but in this situation its wiser to use the more recent numbers. Duda, like a lot of lefty sluggers, struggled against lefties early in his career. The more recent numbers suggest he’s figured it out. Thornton versus Duda in that situation wasn’t the advantage Williams thought it was. 

The Nationals lost a close game on Saturday. That by itself doesn’t mean Matt Williams blew it. The decisions made by Williams, however, showed bad judgment. Yes, the players have to perform. But a manager’s job is to put his players in the best position to succeed. That fact that Williams didn’t do that on Saturday night is more than troubling. 

What in the world is Matt Williams doing with his bullpen?

It’s not news that I am not a fan of Matt Williams’ bullpen management. I wrote this last April

I wish I could say Matt Williams’ bullpen management was a mystery. If it was a mystery, there might be a revolutionary and ultimately satisfying explanation for how he’s managing this team. Unfortunately we know exactly what’s happening. Williams assigns his relievers certain roles and he refuses to deviate from his plan. In series like this weekend, it causes problems, and likely loses ballgames.

Three months later, Williams hasn’t changed. He’ll probably never change. He still assigns roles his relievers–sometimes roles that don’t make sense–and then refuses to deviate. Friday and Saturday against the Mets, neither Drew Storen nor Jonathan Papelbon pitched. Each night, the Nationals surrendered a late lead to the 2nd place team in their division while Williams’ two best relievers sat in the bullpen, unused.

I could complain about this for the 589th time, but there’s no point. Matt Williams is Matt Williams. If he cared what other people thought, he would have changed by now. In the wreckage of the 2014 NLDS, I tried to explain Matt Williams’ Mattiness:

Unlike many MLB managers, Matt Williams was a star player. He didn’t bounce from level to level and team to team. He played 16 of his 17 MLB seasons with only two teams. Over half of his nearly 7600 MLB plate appearances were from the cleanup spot. Matt Williams was a player who didn’t see a lot of change in his career. There’s a good chance he believes a chunk of his success came from this consistency. Every day, same team, same place in the order. Stability led to success over time.

So instead of complaining about Williams, which is entirely unproductive, I want to figure out what he’s doing. If we can predict what Matty’s behavior–and it’s obvious by now he is predictable–maybe we can save ourselves a little bit of stress on gameday.

1. Papelbon will not be used on the road unless it’s a save situation. Papelbon reportedly asked for assurances he would be the closer in DC before agreeing to waive his no trade clause. Once arriving, Papelbon told reporters he wants to break Mariano Rivera’s career saves record. Now, it would be nice if nobody cared about meaningless statistics like “saves” but that doesn’t appear to be the case. On Friday night, Williams had plenty of opportunities in the 9th, 10th, 11th, and 12th innings to use Papelbon, but chose not to, presumably saving him for a save situation (that never arrived). During home games, it remains to be seen how Williams will use Papelbon when a game goes to extra innings.

2. Storen won’t be used before the 9th inning unless it’s a save situation. Papelbon’s arrival bumped Storen to a setup role in the 8th inning. This is exactly how he was used in Thursday’s 1-0 win against the Marlins. On Saturday against the Mets however, Williams turned to Matt Thornton and Aaron Barrett in the 8th, despite Storen being rested and ready to go. In the top of the 9th, with the Nats losing 3-2, Storen started to warm up. This likely means he would have been used in the bottom of the 9th if the Nats tied the game in the top of the inning. If the Nats took the lead, Papelbon would have pitched…of course.

3. Everything else is a mystery. We know Tanner Roark is the long relief guy. If the starter can’t make it to the 6th inning, he’s probably pitching. If it’s not time for Storen yet, Williams will try to get by with some combination of Thornton, Janssen, Rivero, and Barrett. Whether Williams uses the right guy at the right time, we’ll just have to put our faith the reigning National League Manager of the Year.