What Is Going On?
On curses, probabilities, and the role of luck in outcomes
slate grey cloud cover copperized as it absorbed the setting sun’s rays. Some in the crowd thought this was an omen, the sky a shade of Padres brown. Others saw the menacing hue of Giants orange. 9 innings later the Padres lost their 5th straight game, which included one of the most controversial and consequential calls of the year, and as a result, were shutout for the second time in as many days.
The Padres offense has not performed as well as expected since the biggest trade in MLB history.
That’s actually an enormous understatement.
Since a magical performance on Opening Day 2.0 this revamped Padres offense featuring sluggers Juan Soto, Josh Bell, and Brandon Drury have scored 7 runs over 5 games, losing all 5 including, as of the moment of the writing, a 23 inning scoreless streak. That’s 1.4 runs per game. To put that in perspective the 1969 Padres expansion season offense, considered one of the worst offenses in MLB history, still averaged 2.89 runs per game.
What to make of this?
When results are as bad as the Padres’ offense has been the past 5 games we tend to look for intuitive explanations. We believe there must be a reason for the offense to be so bad. Are the Padres cursed? Are Juan Soto and Josh Bell actually worse than Mazara and Hosmer? Is this lineup not as good as we all thought? Something we don’t intuitively consider is whether the lineup could be just as good as we expected, and still produce these bad results because they don’t fully control the outcome of their efforts.
As humans we rely on heuristics, mental shortcuts, to understand the world around us. Complex calculations consume time and resources and are impractical in day to day life, so we rely on rules of thumb which are usually good enough. But one of the mistakes we make is to stick with heuristics in complex situations where a probabilistic approach would yield better understanding. One of the most common heuristic errors is assuming that outcomes directly reflect the processes that went into those outcomes, forgetting the role that randomness plays. That is to say, if an outcome is bad we assume the process that went into that outcome must have been bad and vice versa. This assumption is true for deterministic systems where randomness plays no role in the development of future states. But this just isn’t true of probabilistic systems, where the same input can yield wildly different outputs. Baseball is the ultimate probabilistic system. Juan Soto is unquestionably a top tier hitter. Juan Soto at the plate can yield deep home runs, four pitch walks, groundouts, or even occasional strike outs. These are extremely different outcomes but the process leading to the outcome is the same: Juan Soto walks up to the plate and tries his hardest. Regardless of the outcome the process is the same, he is still Juan Soto, arguably the greatest hitter in baseball.
O = P + L
Outcomes are a composite of Process and Luck. We are capable of thinking probabilistically, but we do not think probabilistically by default, it takes a lot of effort. We instinctively see this terrible 5 game performance and assume the lineup must not be as good as we thought because if they were so good we wouldn’t get such bad outcomes. Our gut reactions are based on the heuristic O = P, Outcomes are direct results of Process. But that assumption ignores the role of randomness. A probabilistic framing would be O = P + L, Outcomes are the result of Process plus Luck:
The Process here is assembling a lineup of productive hitters. The Outcome is runs being scored. There is overwhelming evidence that the lineup of Profar, Soto, Machado, Bell, Drury, Kim, Cronenworth, Myers/Grisham, and Alfaro/Nola is a major improvement on the lineups that the Padres put forth in the first half of the season which generated a roughly league average offense. That is to say, the Padres have done a good job on the Process part of the equation. The Outcome over the past 5 games has been extremely bad, much worse than the all-time worst offenses in MLB history. But it’s a mistake to assume that the bad performance over those games suggests that the lineup (Process) is bad.
Where on the Process/Outcome matrix do these results lie?
Using the matrix above, the 1969 Padres were a Deserved Failure: an expansion roster put together from the dregs of league. They were expected to be terrible because the Process was terrible: they assembled a group of very bad players. The 2022 revamped Padres’ performance these past 5 games strikes us as a Bad Break; a very good process yielding a poor outcome due to randomness (luck). This offense is unquestionably better than its performance the past 5 games, and there will be positive regression in the future.
The best hitters in baseball make an out in 60% of their at bats. Thus, even an excellent Process, assembling a lineup of very good hitters, will yield failure in about 60% of their at bats. The failures and successes will be distributed with some degree of randomness which will yield large variations in short term performance that will tend to normalize over increasingly large samples. We are almost unquestionably living through the nadir of this offense’s performance. We are also almost certainly going to see the revamped Padres offense outperform the first half Padres over the sample of the entire second half.
Can we rule out a curse?
It’s hard to define a curse, but generally it can be thought of as performing far worse than would be expected given normative distributions of luck. It is thus very difficult to ever rule out a curse as long as a team is performing poorly. We feel that a curse is not the most likely explanation for the team’s struggles, but also that it is impossible to rule out. In exploring this topic two of our staff writers realized that they vowed to buy Juan Soto City Connect jerseys if the Padres traded for Juan Soto, but somewhat negligently, let a week pass without fulfilling this holy vow due to the unavailability of the Juan Soto City Connect jerseys in the Padres store. Realizing that this transgression may have invited scorn from the baseball God(s?), the two staff took steps to remedy the sacrilege after the game tonight:
Lest it appear that we are exploiting a loophole in this holy vow, both writers intend, out of an abundance of caution (and at great peril to their pocket books), to also purchase the Juan Soto City Connect jersey as soon as they become available.
Nothing is guaranteed
We’ve covered what a World Series contender looks like, and the fact that it is often not the best team in baseball that wins the World Series. This is due to the randomness that permeates probabilistic endeavors like baseball. No team in baseball, not even the Dodgers, can fully control their destiny. What teams can control is the Process. Evidence strongly suggests that the Padres have done a good job controlling the Process, assembling a very good offense to support a very good defense and pitching staff. The Outcome so far has not been what we hoped. But there is every reason to think that this is due to the Luck portion of the O = P+L equation. It can be hard to believe that an offense producing such bad outcomes can be elite, but if you understand the Process/Outcomes matrix it is rational to have faith that elite performances are coming. We have to keep the faith.
Please share this post with any Padre faithful that may need some encouragement.