Matt Williams’ Blues: Consequentialism, Hindsight Bias, And Moral Luck

zimmermann

As I wrote last year about this time, the baseball play-offs make us unethical. Managers make decisions that either work or back-fire, and feed the toxic human tendency toward  consequentialism thusly: when they work, the decisions werecorrect; when they don’t, the manager was an idiot, and the choicee were obviously wrong. As with judging the ethics of an act, what happens after a baseball decision is made is irrelevant to whether it was a good decision when it was made.  This is almost impossible to keep firmly in mind. Our logic rebels at the idea that an ethical act can have horrendous consequences, or that the right tactical decision can result in defeat. But that’s life, as my father was fond of saying.

Hindsight bias further pushes us to confuse the making of a decision with its consequences. It is, not surprisingly, much easier to make a strong case that a decision was the wrong one after all the results are in. This, of course, is unfair to the decision-maker, who didn’t have the data the critics do when he or she acted. On the other hand, sometimes the reason the decision was the wrong one is that it was wrong, and the fact that the results were bad just support that verdict.

This morning, indeed since last night, Washington D.C. baseball fans and sportswriters have been wrestling this conundrum. The Washington Nationals, widely believed to be the strongest National League team in the post season, and quite possibly the favorites to win the World Series, find themselves down 0-2 in the best of five National League Division Series after a grueling, 18 inning loss to the San Francisco Giants, who didn’t even win their own division. The way the game went into extra innings will be debated for months if the Nats fail to rally and win the series. Nats starting pitcher Justin Zimmermann, who had pitched a no-hitter in his last outing, had been almost as good this time, pitching his team within one out of a 1-0 win that would have evened the series. He had dominated Giants hitters in every way, and had not shown any signs of weakening or, as they say in the game, “losing his stuff.” In the old days, that is, as recently as 20 years ago, a pitcher on a run like this would finish the game unless he had a stroke on the mound. Now, MLB managers are trained to be ready to go to their ninth inning specialist, the so-called closer, at any hint of trouble or even without it, and they almost always do.

As a reflex action, it makes no sense a lot of the time, other than “everybody does it.” A pitcher whom you know is pitching well is a known quantity, while a pitcher newly arrived to the game, whatever his skills, is not.  If the choice is between a starter who is not just doing OK but rather mowing down batters like Samson jaw-boning the Philistines, and bringing in a new arm, logic would dictate that the latter is the greater risk.

Nats manager Matt Williams acknowledged that Zimmerman was “in the zone” by not lifting him to begin the ninth, and was rewarded with two quick outs. When he walked his first batter of the game, however, on his magic 100th pitch (they count pitches now, and 100 is the number at which pitchers supposedly turn into pumpkins), and Williams lifted him, calling on closer Drew Storen.

In the stands, in front of TV sets, baseball fans breathed a collective, “Oh-oh.” Zimmerman didn’t look tired. He had given up just three hits the whole game. Storen, though he had been excellent lately, is still remembered as the relief pitcher whose failure lost the crucial game the last time the Nats were in the play-offs two years ago.

Was it the right decision? Well, this is exactly what Williams has done in similar situations all season, presumably because he thinks it’s the right thing to do. His team won its division with him using his pitchers this way. If it’s the right tactic, it’s the right tactic, whether the game is an important one, like a play-off game, or not. Essentially, Williams, having made the decision to remove the starter, is at the mercy of events beyond his control….moral luck.  If Storen comes in and gets the final out, well, the manager was “right.” If the closer blows the game, the manager must have been wrong, or, as my sister termed it last night, “an idiot.”

Storen gave up two hits to let the player Zimmerman walked score the tying run, sending the game to extra innings, and a devastating, season-threatening loss.

So Williams is an idiot, and everyone on Facebook, the Washington Post and elsewhere is saying so. Williams is legitimately wondering: if this was such a bad way to handle pitchers this season, why didn’t I hear about it before?

That’s simple, Matt: because it didn’t lead to a horrible, disappointing, six-and-a-half hour loss before. Hey, I would have told you, if you asked, that automatically taking out a successful starter because of an arbitrary pitch count and because closers are the current fad in baseball is lazy, risky and dumb, but you didn’t ask. Now you’re in the maw of hindsight bias. Sorry.

Over at the Post, the most over-rated baseball writer in the world made a jaw-dropping argument against Williams’ move, refreshing that it did not relay on consequentialism, outrageous because it was, well, stupid. I’ll let Thomas Boswell explain:

“There is strategy and there is theater. As strategy, there is some justification for taking out Zimmermann, who had just thrown his 100th pitch — a ball four to Joe Panik. Perhaps a fresh arm — a right-hander Storen with his 1.18 ERA this season, was a proper by-the-book move. But, sometimes, there is also context and theater — and even the danger of messing with the lower-case but still haughty baseball gods….Why do we revere sport in this culture? Why do we pass stories down through generations? Surely, at some level, it is about seeing the best-of-the-best test each other in the supreme moments of their sport. Nobody, it’s said, comes to ballgames to watch the umpires. To a lesser degree, that applies to managers. There are times, and Saturday night was one of them, when we don’t come to see them.What we arrive to see, then stay to the end to witness, is performances like Zimmermann’s…”

Ridiculous:

1. There are no “baseball gods,’ Tom.

2. Williams’ job is to do what is most likely to win the game, not what will make the best movie.

3. There is no time, ever, when avoiding what would normally be the best strategy should be ignored in the interests of creating more compelling drama. The best drama ends with a happy ending.

Unfortunately, there is no way for a manager, like Williams, to make the “right decision” from the outset the way most people evaluate such things. You can only know the right decision after all the incalculable, chaos-driven and random events have delivered a final result. This is an illogical, unfair and ultimately destructive way to think, basing  your comprehension of right on the vagueries of moral luck, but that’s how we are wired, and to some extent, that’s why we have such a hard time understanding right from wrong.

.

 

14 thoughts on “Matt Williams’ Blues: Consequentialism, Hindsight Bias, And Moral Luck

  1. You confused me, Jack. You’re writing about Matt Williams but the title says Mitch Williams (no doubt your brain is still pondering the death of your friend, Mitch Dale). So, as a lifelong suffering Phillies fan, my brain went to relief pitcher/closer Mitch Williams (aka “The Wild Thing”), who some (many?) believe was responsible for the Phillies losing the 1993 World Series to Toronto. (Oy, that series broke my heart.) Oddly though, the Mitch Williams disaster (and Curt Schilling’s bad behavior) are pertinent to this discussion, dontcha think?

  2. I agree with the points about consequentialism and the manager being judged by factors outside of his ken. I can create a very solid argument for why, regardless of whether it is SOP or not, the move was the right one to make. As you note, it just didn’t work. Both in his career and in 2014, once you get past 75 pitches his opponent’s OPS goes up (past 100 it actually dropped slightly in 2014, but only based on 18 plate appearances it’s hard to draw any meaningful conclusions). What is significant is that the only time you get to those pitches is when you throw well early on, i.e., when you’re on your stuff. There’s a reason that sabremetric types say that teams with a good bullpen (i.e., not the 2014 Tigers) should NOT let their pitcher go through the lineup a 4th time if the game is in balance. All pitchers tire – the argument is whether you’d rather have 75% of JZ or 100% of Storen to get one out. And Storen had been excellent all season, particularly since taking over the closer role.

    Further, in last night’s game, JZ had not only walked a fairly light hitting middle infielder (Panik) – never a particularly good sign. Before that, JZ came within a couple of millimeters of giving up the game tying HR to him. A guy with 1 HR in 287 plate appearances turned on a pitch and drove it into the second deck … just foul. If Williams leaves JZ in that game and Posey hits it out, then (because of consequentialism, etc) he gets roasted for leaving JZ in the game. Grady Little II, if you will.

    Heck, none of the three pitches Storen threw were fat pitches/mistakes. If you look at pitch track, they were all on the edge – and different edges of the plate (left, right, low). They just got hit and found holes. Posey went up the middle, and Sandoval stuck his bat out and got it down the left field line. Good hitters putting the ball in play; it’s what they do.

    So I take issue with your “automatically taking out a successful starter because of an arbitrary pitch count and because closers are the current fad in baseball is lazy, risky and dumb, but you didn’t ask.” Your assumption is that the decision was automatic, lazy, risky and dumb. Williams didn’t have any non-risky decisions there. And it’s one helluva leap to imply, as you seem to, that Williams had not thought through and made a reasoned, considered decision instead of thinking a lazy reflex reaction. If that is your iimplication, that is also unjust.

    • And obviously, the sentence should have said “should never let their pitcher go through the lineup a fourth time”

    • No, it was lazy. Williams said that he’s always done this before, but in fact each situation is unique–Boswell is right to that extent. Context matters: the game has to be won. Personnel matters. Which is more relevant—that Storen blew the last game of the last playoffs he was in, or how he did this season? I think that needs to be part of the calculation. (In the 1986 Game 6, John McNamara lazily chose Schiraldi to close. He had been lights out in the season, but in the post-season had looked like a deer in the headlights. AT THE TIME I thought it was lazy, rote call.) The walk—it was very close, and could have easily been a K. JZ wasn’t losing command. He looked as strong as ever. A foul homer? You know that such things occur on pitches that can ONLY be hit foul—they don’t mean anything. And he had the hottest pitcher in the universe, with a manageable pitch count. It wasn’t like Lester, in the wild card game, who was fried, or Pedro in that 2003 horror, who was obviously out of gas, and thank YOU for reminding me.

      Williams did the book move, and based on his “I always have done that” excuse, I have to assume he was on autopilot: lazy.

  3. Patrice, the manager’s dilemma is that he (and so far it’s always a he at the MLB level) has to make a move that works or he gets roasted. Take the guy out? You’re Matt Williams last night, or anyone who relied on Matt Williams. Leave him in? Congratulations, you’re Grady Little. Helluva choice.

  4. Jack, you’re simply wrong that JZ “had not showed any sign of weakening.” I was at the game, all six+ hours of it, with seats in the infield gallery. The trope that the Giants were helpless before him throughout is not accurate. The third time through the order the Giants were making solid contact off him. No hits yet, but as much due to good BABIP fortune (including nice defensive plays by Rendon & LaRoche) as to anything else. The Panik at bat was simply the latest in a series of signs for anyone who cared to look. Given that the next hitter was not only one of the best hitters in baseball but had also been responsible for one of those loud outs in his previous at bat, the hook was not shocking to those paying attention to the game and not the story.

    And Jack, there certainly are pitches that are intended to only be able to be hit foul. If it comes THAT close to be a second deck HR, one is doing it wrong. And doing it wrong where before it had been done right is another sign of weakening

    To me, Williams was neither reflexively “new school” (Storen would have started the 9th) or “old school” (JZ would have gotten a longer leash). Nor was the pattern completely consistent with his usual pattern through the season. Typically he brings in relievers to start an inning. You could look up the pattern, if so inclined.

    All these are circumstantial of course. Unlike you, I can’t read the man’s mind. But there is a solid basis for concluding that he went outside of his usual thinking. So I find your calm assumption that you have divined his reasoning and found it wanting to be a house of cards. And somewhat ironic, given the rest of your post.

    • I saw all six hours too. Glad you could tell what none of the trained “experts” could from the booth. I have never seen a pitcher yanked because of a hard hit foul—if you use foul home runs, why not foul line drives? Hard foul grounders?—and using a BB simply because of its timing—a base on balls in the 6th, “Eh.”, a base on balls in the 9th, “He’s weakening!”…come on. You don’t call that reflex managing? When the pitches were sharp and close calls? When he wasn’t missing by much at all?

      Of course we don’t know how Williams made the decision (he SAID that he played it the same way he always had, though…doesn’t that suggest rote?), but it sure looked like the lamest of half-decisions: “I think the starter has enough in the tank to finish, but not enough faith in him to get the last out before he gives up a hit.” Huh?

      If you are going to lift him, you wait to see what the next batter does.

      I think you are gripped by a a different bias, confirmation bias. You want Williams’ to have made the right move, so you are seeing things in the most favorable light.

      • I’m glad you used quotes around “experts” in the booth, because it’s rare I hear anything insightful from a broadcaster. There are plenty of experts that have noted the same things I did (C. J. Nitkowski, etc). I go by what I saw:

        (1) increasing hard contact the third time through the lineup, navigated by a combination of good fortune & good defense;

        (2) the fact that walking a slap hitting middle infielder to give Buster Posey a chance to bat is a mortal sin, not a venal one; and

        (3) the fact that Jordan put ONE pitch in the zone to Panik, and damn near lost the lead on it. Jack, I went back and looked at the pitch f/x data. That ball wasn’t off the plate in, where it could only be barreled up foul; it was thigh high and about 1/3 of the way in from the inside corner. In other words it was in classic “crush me” position, so bad that Joe Panik (of all people) damn near did.

        It’s funny that you whack me with confirmation bias. I don’t loathe Williams the way some do, but I find his “Big Marine” schtick tiresome and was pushing for Knorr to get the job. I don’t like his lineups, but recognize that optimizing lineups has little real impact. I don’t think he should have gotten himself tossed either (Cabrera letting himself get rattled by the ump’s wandering zone & getting tossed was much worse, though).

        • And when I say 1/3 of the way in, I mean towards the heart of the plate of course.

          Jack, when I agree with that one decision by Williams, I’m basing it on personal observations underscored by pitch f/x data, coupled with statistical analysis both of pitchers in general (batter OBP goes up with more pitches and more times through the order), Zimmermann in particular (he’s not immune to this trend – although for 2014 batter OBP goes down slightly over 100 pitches, that is based on only 18 plate appearances and thus more likely simply random noise) and Storen’s stats. I’m not immune to confirmation bias. But then again, who is?

          • The answer to that is “Nobody.”
            I love sabermetrics, I love what it has brought to the understanding of the game. But human athletes are still not Strat-O-Matic cards, and can’t be reduced to that. If Williams was basing his judgement on data showing how well Zimmermann’s stats show he gets through the line-up the third time around without taking into consideration what he was seeing, feeling and actively assessing in the game itself, then he was managing exactly as I suggested he was, right? Why not have a computer manage the game decisions: number of pitches, score, pitch f/x data–DING! Take him out!

            • And if he was using it (pitch counts, not pitch f/x, which he doesn’t have access to during the game) as a tool to supplement what he was observing, then – ding! – he’s not.

              Just because he didn’t make the choice you would have made doesn’t mean that he was wrong, foolish or unthinking.

              FWIW, Rizzo has a reputation as a scout who has invested heavily in numbers as tools, not deciders. They have highly paid numbers crunchers, but they feed data to the scouts instead of dictating to them. MW is Rizzo’s hand picked man for the job, someone Rizzo has observed for years. The chances that MW made a reflexive numbers based decision is not nil, but it’s pretty close.

              • “Just because he didn’t make the choice you would have made doesn’t mean that he was wrong, foolish or unthinking.”

                I remind you what HE said to justify his call. If he had cited what you have, that would be different (a little). Williams said, rather, that he made the same move he made all season. You don’t think that’s a “Ding!”?? In fact, he never had lifted a pitcher who had not given up a hit for 21 straight batters, and if he thought what he was doing was the obvious thing to do, I submit he hadn’t thought very hard, or wasn’t paying attention. Nobody lifts a pitcher who hasn’t given up a hit in 21 batters, though you seem to be saying, if the metrics say lift him, that shouldn’t matter.

                And this is where the stat-heads go off the rails.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.