Thursday, April 7, 2011

HIGH FALUTIN' THE FIGURES FOR THE "LOW FIVES"

As the world keeps to its lockstep schedule of falling apart, the Boston Red Sox and the Tampa Bay Rays are doing their part to penetrate deep into the conga line of pandemonium. Both of these estimable organizations and lightning rods for the consecration of post-Moneyball nerdelicious hegemony have taken a serious tumble out of the gate, starting the season with five [semi-late update: make that six...] consecutive losses to begin the season. (The third member of this group--the Houston Astros, who also started 0-5, marking the first time since 1988 that three teams began the season in such fashion--has no cachet in the burgeoning little world of post-neo-sabermetrics, so they've been mostly ignored in the midst of this early-season confuffulation.)

The partisanship that both begins and ends even the most sophisticated engagement with baseball becomes especially strained when these anointed franchises show any potential sign of weakness--which is where penis size and IQ suddenly find themselves to be the strangest of all possible bunk-bedmates. A lot of ink has been spilled about the "long odds" facing both of these teams as they have stumbled out of the 2011 starting gate, but as usual there is little nuance or real historical perspective being provided.

Fortunately (or otherwise, as your particular mien dictates...) we are, as usual, here to dig into the data in ways that others deign to do (though no aspersions cast at you, Dayn Perry!). What we propose to do is look at the context of these "low five" (0-5 starts to the season) and put some more meat on the bones. Truth told, it wouldn't be hard to do.

We can all start with the BB-ref "streaks" data, but only a few of us (as in one more than zero...) will bring it to you with the full context of baseball history. The chart over on the right side shows the full breakout of year-end results for the first five games of the season from 1901-2011. (We've greyed out the results where teams have less than five decisions, mostly artifacts from baseball's primordial past). At the extreme right, you will find the aggregate end-of-season winning percentage (TW%) for all teams in each of the first-five-game won-loss categories. It is, as you'd probably suspect even if you were a liberal arts major, a linear ascent from 0-5 to 5-0.

Now this does not mean that any particular team who happens to start the season 0-5, or 3-2, or 2-3, or 5-0, is going to wind up with precisely the WPCT that's shown in the TW% column. Of course, there is a great deal of variation within those aggregate values. (A great deal of energy has been expended, primarily by Red Sox fans, in venting this fact--and for sound reasons. Just because the Red Sox began 2011 with an 0-5 record--er, make that 0-6--doesn't mean they will wind up with a sub-.500 record.)

No, the way we should look at this information is a bit more involved than what we've seen in the above chart. To get great context, we should identify each of the 84 teams in baseball history who have begun the year 0-5 (or 3.7% of all team-years in baseball since 1901) and look at some additional performance data. We should be looking at the team's record in the season prior to the one in which they began the year 0-5 (let's call that W-1), the team's record in the year sullied by the 0-5 start (W0), and the team's record in the following year (W1).

We'll also record whether the team has gone to the postseason in the previous year, in the year with the 0-5 start (the mainstream and sabermetric media has already told us in their semi-interchangeable way that only two teams have actually made it to the post-season after such a stumble...), and in the year afterwards. After we've looked at the master data, we can do some additional summarization and look for  patterns that exist below the superficial "bottom line" that's already been presented.

As you'll see, we organized the data by franchise and by year. Sort of in the manner of Charles Foster Kane, who thought it would be "fun to run a newspaper," we think it's fun to see how often various franchises have managed to stumble out of the starting gate. It turns out that the Detroit Tigers and the Los Angeles Dodgers have done the most such stumbling, with eight each (though the Dodgers had five of these 0-for-5's while they were in Brooklyn).

The Red Sox are next, with six. But you'll notice that the Sox have a lot of bold type in their entries. We can see that the 1905 Red Sox, when they started 0-5, were coming off a previous year that was simply sensational (105-47 and a pennant). Similarly, in 1996, when they lost their first five, their previous year had been a division winning affair.

As you look down at the bottom of the chart (way, way down), you'll see that there aren't too teams with really good records (read: post-season appearances) in the previous year who start out the next year 0-5. As a matter of fact, there are only four: the 1904 and 1995 Boston Red Sox, the 1994 Cincinnati Reds, and--how about that?--the 2010 Tampa Bay Rays.

And we already know that only two teams (1995 Reds and 1974 Pittsburgh Pirates) have come back from an 0-5 start to make the post-season. This fact has been sliced dozens of different ways in order to confirm or deny whatever theory or posture is in need of being expressed by a plenitude of pundits.

What's kind of interesting, however, is the fact that out of the 81 teams who've had an 0-5 start for whom we have following year data, 13 of them have gone to the post-season in the next season. That's a surprisingly robust 16% of all teams, about four times what you'd expect to be the case.

Another tidbit available via this historical survey is the fact that on six occasions, a team has stumbled out of the gate 0-5 in two consecutive seasons. The 2010-11 Astros are the sixth such team, following in the footsteps of the 2010 squad. The others are the 1905-06 Dodgers, the 1962-63 New York Mets (something poetically correct about that one, don't you think?), the 1987-88 San Diego Padres, the 1992-93 Kansas City Royals, and the 2002-03 Detroit Tigers.

It's also kind of amusing to discover that, when you do a bit of rounding in the data, the difference between the records of the teams in the year they start 0-5 and the previous year averages out to--yes, that's right, exactly five games! So they end up where they would have been, except for the 0-5 start. There are some overtones to such a result that, if not overtly existential in nature, can at least be seen as Sisyphean.

In the year previous to the one in which they start the year 0-5, teams have an average WPCT of .464. In the year with the 0-5 start, that aggregate WPCT drops to .431. What's a bit scary for teams like the Red Sox (89-73 in 2010) and the Rays (96-66) is that the better the team is in the previous year, the more on average they lose when they start 0-5 the next season: the 26 teams who were above .500 in the "-1 year" lose more than twice as much ground as measured by WPCT (the dip goes from -.033 to -.084).

That works out to about a thirteen-game downturn on average (from 86-69 to 73-82: remember that many of these seasons occurred before expansion, so the average number of games is closer to 154 than 162). If we apply the aggregate to the Red Sox, they would wind up 76-86 this year, while the Rays would wind up 83-79. Of course, that is not anything close to a scientific prediction, but merely a very rough indicator of what might happen based on a small sample size pattern from the past.

The last little piece of fun you can have from the large chart is to examine it with the idea of determining which franchises have never had an 0-5 start in their entire history. One of them is rather surprising when you consider that a famous slogan of baseball futility is associated with them. Have fun figuring it out, and if you're a fan of the Rays, Red Sox or Astros, feel free to avoid walking under ladders and other similar market inefficiencies, at least until the coin flips seem to be going your way again...