Sunday, October 18, 2020

In Defense of the Astros

I am glad to be writing this post as an irrelevant digression rather than a timely attempt to wrap one’s mind around the possibility of a team with a 29-31 regular season record winning the World Series. My standard outlook when watching the playoffs is to pull for the better team. This is not an ironclad rule, as I do allow my various prejudices (fandom, a favorite player, a least favorite player, style of play, etc.) to override the general principle. I also do this despite being unable to be truly surprised by any outcome of a seven-game series between any two legitimate major league baseball teams. That it does not at all surprise me when inferior teams win series against superior teams, and while it can occasionally be fun to see this happen just to see the egg on the face of people who wildly overestimate the probability of the better team winning (for me, the 2010-11 Phillies were the archetype of this phenomenon), I generally find it disappointing.

After Houston tied the series at three, though, and as I read some online reactions to the possibility of them appearing in the World Series, I decided it would be worthwhile to explain at length why I do not feel that they would be an illegitimate champion should it come to pass. The fact that it did not come to pass makes this discussion academic, but look at where you are. One thing that should be said upfront is that given the emotion generated by the Astros, it is impossible to disentangle how much of the backlash to the possibility of them winning was rooted in genuine concern about the legitimacy of the outcome of the MLB season (concerns that, I have argued, would have been much better channeled towards the possibility of Miami advancing deep into the playoffs), and how much is simply rationalization for the often disproportionate reaction to the sign stealing scandal. There are also people who are looking for any reason to discount the 2020 season, whether out of allowing the perfect to be the enemy of the good (I just cannot relate to people who would rather see no season than 60 games; neither can the bank accounts of the players), or out of darker authoritarian and totalitarian impulses. Arguments borne from there are outside the scope of this piece – this is about the implications of a 29-31 team appearing in or winning the World Series.

The first question that should be asked is “How good are the Astros, really?” My answer is – almost certainly better than 29-31. Why do I say that?

1. The Astros underlying performance was better than their win-loss record

The Astros scored 4.80 and allowed 4.72 runs per nine innings, which in a world of nine-inning games would be expected to result in a winning percentage of .508. The Astros tallied 4.58 runs created and 4.50 runs created allowed per nine innings, which in that same world would also be expected to be allowed in a winning percentage of .508.

Of course, this is not significantly or in any way meaningfully different than their actual W% of .483, but if you have a psychological hangup on a sub-.500 team winning a pennant, their underlying performance was ever so slightly better than that of a .500 team, and to the extent that these tiny differences mean anything, they should mean more the smaller the sample size of games actually is.

2. The Astros win-loss record would likely have been over .500, even at their same level of performance, had a normal schedule been played

Last year, the AL West was the best division in MLB, with an average schedule-adjusted W% of .527 (of course, the Astros as the best team in baseball contributed much to this). The NL West was third at .510. While such calculations cannot be made for 2020 because of the lack of inter-geographic games (we can add the playoffs in, but they don’t provide a sufficient sample to be relied upon, although it is hard not to notice the complete playoff decimation of the Central), my guess is that playing a western-only schedule was tougher than the average schedule faced by a MLB team. This also suggests a slightly better than .500 true performance.

3. Most importantly, the Astros’ true talent was significantly better than that of a .483 team

Some time before the season, I grabbed the win projections from three sources that I think do quality work on these sorts of questions: Baseball Prospectus, Clay Davenport, and Fangraphs. Their win projections for the Astros were 36, 36, and 35. Respectively, that put them second, first, and tied for first in the AL (I should note, I didn’t record the date I jotted down these numbers, so they may not have represented the final estimate from each source).

Of course, things changed, most notably the injuries to Justin Verlander and Yordan Alvarez, but these are insufficient to take a 35 win team to under .500. Some Houston hitters (particularly Jose Altuve) had poor seasons that might have recalibrated estimates of their true talent, but how much stock can you put in sixty games of additional data, and there are players going the other direction too (Framber Valdez is a big one). It seems highly likely that the Astros underlying true talent was that of a winning team.

I note this not to self-aggrandize (heavens knows I’m wrong about these things more than I’m right), but rather to cover myself in case anyone looks back at my 2020 Predictions post, I actually did not pick the Astros to make the playoffs under the traditional format, but did once the field was expanded. While I did not expound on my logic, I did think the Astros were more vulnerable than the purely objective approaches did, mainly because I thought that a starting staff anchored by two pitchers older than me was not a foolproof plan (even if they are Hall of Famers to be). I actually picked the Rays to win the AL pennant, which didn’t take any particular insight but viewed in the wrong light could discredit the opinions I’m expressing here.

If you now will accept the premise that the Astros 1) performed like a winning team 2) would have been a winning team against a normal schedule and 3) could expect to perform better over any stretch of future games due to their underlying talent, are the Astros in fact a beneficiary of a sixty-game season – or a victim? I would argue the latter – the 2020 Astros were a good team that happened to have a mediocre record, largely caused by a season with a schedule that was too short and too unbalanced.

I don’t know what the purpose of a playoff system is (actually I do – it’s to make money). Is it an attempt to reward the teams that performed the best in the regular season? (I’m pretty sure it’s not). Is it an attempt to identify the best team, treating the regular season only as a means of culling the herd? (I think many fans feel this way, which I think is crazy). In reality it is likely some combination of all three. But if the point of a playoff system is to use the regular season to cull the herd, then attempt to identify the best team, I am far from convinced that the Astros should not have been a beneficiary of such a system. And while I would generally consider such a system foolhardy, it is certainly more defensible when the regular season used to do the culling is only sixty games – we should have much less confidence that the results of the sixty games should count for more than the results of the playoffs than we normally would be.

To put this into quantifiable terms, let’s suppose that we decide that in order for a champion to be legitimate, they have to meet some minimal standard of competence. We could then test the worthiness of a champion by calculating the probability that a team that met that minimum standard of competence would have compiled a record equal or worse than the team in question.

For example, let’s say that we decree that in order to be a worthy champion, a team should be at least a true .556 W% team. This is of course an arbitrary value; I’ve chosen it because it corresponds to a 90 win team, which is a value I have always subjectively thought of as marking a legitimate contender. So under this argument, we will accept the playoff result as legitimate, to the extent that the champion it produces was in fact a 90 win quality team.

Using the binomial distribution, it is simple to calculate the probability that a team with a given W% would have a record of 29-31 or worse. The probability of a .556 team going 29-31 or worse over sixty games is 15.8%. Contrast this with the Rays, who were 40-20, which would happen with probability 97.0%.

What would an equivalent performance to Houston’s 29-31 be over the course of 162 games? A .556 team would win 83 or fewer with a probability of 14.9% and 84 or fewer with a probability of 18.9%. The 2006 Cardinals won the World Series after going 83-78 (darn that missing game; the probability for 83-78 is 17.0%). So if you demand that your world champion be at least a .556 team, you could plausibly argue that the Astros would have been the least worthy champion in history. (I would in fact argue that the Astros were more worthy under this criteria based on the considerations discussed above, particularly strength of schedule since it remains grounded in actual wins and losses rather than component performance or true talent).

But .556 may well be setting the bar too low. Is it too much to demand that a world champion be a .600 team (equivalent to about 97 wins)? One could certainly argue that it is not – the champion should be a team that has demonstrated excellence, not simply a contender.

The probability that a .600 team would win 29 or fewer out of 60 is 4.4%. Conveniently, the probability that a .600 team would win 86 or fewer out of 162 is 4.4%. So under this formulation, the Astros could be seen as an equally worthy champion to a team that went 86-76 over a full season. The 2006 Cardinals are joined by the 1987 Twins (85-77) in not clearing this bar, and the 2014 Giants (88-74) were close as well. Of course, in all cases I would argue that we should at least adjust actual wins for strength of schedule, but I think this suffices to make the point. One can make a very logical case that a 29-31 team would not in fact have been the least worthy World Series winner in history. (I am only going to hint at the low hanging fruit offered by the sixty-game record of the reigning world champions who denied a much more worthy team called the Astros in 2019).

No comments:

Post a Comment

I reserve the right to reject any comment for any reason.