March 4, 2016
Tools of Ignorance
The Dodgers' Breakable Rotation
Three, four, maybe even five years ago, Jimmy Rollins took a healthy lead off second base. A teammate of Rollins was taking a lead off third base and another was leading off first. When watching Jimmy Rollins, or really any player, take a lead off second base with the bases loaded, I rarely take notice. This time was different though; this time I was thinking that Rollins should be taking a much smaller lead. I was screaming, internally. I considered tweeting.
Why was this lead different than almost all other leads? It was different because the Phillies were tied in the bottom half of an extra-inning game with two outs. Rollins’ lead would therefore only help if it helped avoid a force out at third base. At the time, I remember thinking that there was a bigger chance that this lead would backfire (Rollins could be picked off) than there was a chance it would make a difference for the better. A bunch of pitches later, John Mayberry Jr. hit a single up the middle, Rollins’ teammate on third scored, and the Phillies won.
If I had to bet, this type of lead has been taken thousands of times. I would also bet that this type of lead (or any lead at all) has been taken when there was absolutely nothing to gain (runners on second and third with two outs in a walk-off situation). If any of these players had been picked off, things would have been said, calls would have been made into radio stations, tweets would have been tweeted, and GIFs would have been created; and that all would have happened before the analysis started. Articles would probably be written about the costs versus benefits to determine whether the lead would truly be worth it or not. Articles in response to such an article might have been written about how asking a player to do something he is unfamiliar with might do more harm than good—what if Rollins got injured because of his not-lead? That would certainly tip the cost-benefit analysis. Teams might have practiced situational leads with the new analysis considered. This all seems like a waste of words, though, because none of this happened, because no one has been picked off. No teams have spent time practicing what to do in this situation this spring training. So, what’s the point?
The point is that we (people) struggle to consider things that did not happen in the past when making decisions about the future. As Daniel Kahneman writes in Thinking, Fast and Slow, “The human mind does not deal well with nonevents.”
Baseball is a game of events and nonevents. The pitcher either throws a strike or a ball. The hitter either swings or takes. The swing either connects or it does not. The contacted ball either ends up in play or out of play. The ball in play is either converted into an out or it is not. And on and on.
The biggest event in baseball (where something either happens or it does not) from a production perspective is injury, particularly pitcher injuries. A pitcher who gets injured is much less productive than a pitcher who does not get injured.
After the previous Dodgers regime spent tons of money on baseball player contracts, the new regime has not spent tons of money on baseball player contracts. This—the nonspending of money—has been the main topic of discussion when it comes to the Dodgers recent decisions. Is there a budget? Do they care more about profits than titles? Are they rebuilding in disguise?
Once again, Los Angeles has acquired a player with an uncertain future. They did it with Brett Anderson and Brandon McCarthy last year, they’re doing it with Maeda this season. Instead of shelling out the biggest money for the most premium players (Greinke, David Price), they’re signing Chase Utley to round out their infield and investing in Maeda and Scott Kazmir to shore up the rotation on funky, long-term, low-yearly-payout deals.
When we also consider Alex Wood, the trend of investing in uncertainty and perceived injury risk remains intact. There is a chance that the Dodgers are investing in such players because they are cheap and untied to draft picks. But there is a chance that it is more than just treading water while they wait to build a monster farm system and shed bad contracts. There is a chance that the Dodgers feel that the market is overrating consistent pitchers and underrating inconsistent pitchers.
There might just be something to this. Which of the enormous starting pitcher contracts has ever panned out as the team has hoped? Pretty much, not a single one. In Max Scherzer’s Transaction Analysis, Sammy Miller writes,
If we don’t want to give teams the benefit of the doubt, we’d explain the repeated signings of huge-contract pitchers as an example of Base Rate Fallacy, or Base Rate Neglect. As Jeff Quinton put it to me: “We say ‘this guy's different’ instead of saying ‘this guy is a pitcher.’” The fallacy comes in not because nobody is different but because the exceptions are much rarer than the non-exceptions; the best presumption is that you’re not looking at an exception. As doctors say: Think horses, not zebras. Or as that defense lawyer put it to Sarah Koenig: “The odds of you getting a charming sociopath? You're just not that lucky.”
Hey that’s me! (The person Miller quotes, not the person from Koenig quote.)
These enormous contracts do not get handed out to risky pitchers or pitchers with high injury risk. They get handed out to consistent, seemingly safe, “workhorse”-type pitchers. As Miller writes about Szherzer, “He’s so good. He just needs to not get worse.” But let us not get caught up in the big contracts and rather focus on applying the same analysis to the other side of the player-type coin: the inconsistent or injury prone pitcher—the risky pitcher. With these pitchers we say “this guy’s different” instead of saying “this guy is a pitcher,” but “different” means something completely different in this situation. Instead of meaning that they will not break, it means that they will certainly break or that they will certainly disappoint at some point.
This way of seeing the world, that things will either happen or they will not happen, is helpful, especially when facing complex decisions. It is easier to figure out what to do when we have eliminated some options or decided some things are certain. Certainty appeals to us because it makes decisions easier. Pretending things are certain when they are not, though, does not make us better decision-makers.
During any offseason, many teams will decide that they need to improve their starting pitching. All will decide which amounts of risk they are willing to take on when adding starting pitching. Typically, teams that are merely rounding out their rotations will take on the riskier options (the Astros and Doug Fister, or the Pirates and Jon Niese), while teams with bigger holes or question marks target the more consistent pitchers (the Royals and Ian Kennedy, the Red Sox and David Price, and the Diamondbacks and Zack Greinke). Conversely, the Dodgers, who have nothing but question marks after they pencil in Clayton Kershaw as their Opening Day starter, did not target a consistent option.
We have tested a lot of hypotheses in trying to figure out why the Dodgers have been acquiring these types of pitchers, but the reason might be that they are focusing on the second half of the previous quote. That while all other teams are saying that “this guy is different,” that this pitcher had a terrible second half, that this guy is an injury risk because of medicals or past injuries, that this pitcher will definitely or almost certainly not do as well as expected, the Dodgers are saying “this guy is a pitcher.”
Because people do not deal well with nonevents, they tend to place too much emphasis on what happened in the past. The consequence of this is not placing enough emphasis on what has not happened. For the same reason no one worried when Jimmy Rollins took a healthy lead when he could not be the winning run, for the same reason teams continue to hand out large contracts to players who have yet to not be prolific with expectation they will always to be prolific, teams will continue to shy away from risky pitchers. If this is in fact the case, then the Dodgers’ strategy of targeting these players will likely be effective (at least in theory) until enough teams catch on and follow suit in order to affect (drive up) the market prices of such players.
Or until Brett Anderson hurts his back and has to miss five months. Either way.
This plan might work and it might fail, but no matter what happens, there was always a chance for a different outcome. When facing a probabilistic future, the best action is to get the best odds possible, and maybe the way to get the best odds is to take advantage of the competition miscalculating the odds. Again, we have no way of knowing, but it will be interesting to see what happens and what does not happen.