April ciphersby Brandon Isleib
April 17, 2008
Don't worry if the sun don't shine. You've seen it before; you don't need to worry.—"Don't Give Up" (Chicane featuring Bryan Adams)
No one wants you when you lose... it is so strange the way things turn.—"Don't Give Up" (Peter Gabriel featuring Kate Bush)
All but the most pessimistic fan knows that what happens in April doesn't mean a whole lot (individual performance, W-L record, how much time you'll spend doing a 1040 Form each month); still, it's hard not to feel a bit leery when your team underperforms. The feeling of not knowing when to push the panic button is a gut-wrenching one, and it's presumably as hard on general managers as it is fantasy owners (well, okay, maybe harder, given the whole job-at-stake thing). Seeing as how I not only picked the Tigers to win the World Series but also did so quite publicly, I'm feeling the early pressure and hoping that my forecasting abilities don't get dashed on the rocks after only one month.
With so little season down, there are far more questions than answers. Perhaps the biggest question is, Exactly how much should we be questioning? What matters and what doesn't? Hopefully, at least in terms of how your favorite team is doing, and whether or not your slumping team might be a good team this year, this article will answer some of that.
The number 162 is a wonderful amount of games for a baseball season, for the reason that it's divisible in handy ways. You can split it into halves, thirds, ninths, etc., and get nice packets of information. To see how much a big slump at any point affects a team's chances of being a good team, I divided historical 162-game seasons into nine periods (or, if you prefer, innings). Each period consisted of 18 games: games 1-18, 19-36, 37-54, and so on. I then sorted out the records in each 18-game section—whether they broke even at 9-9, won two-thirds of their games at 12-6, etc.—and tabulated the results for all teams that finished the season with at least a .550 winning percentage. (I left out the strike-shortened seasons of 1972, 1981 and 1994, which could not be broken into 18-game pieces, but the strike of 1995 cut out exactly 18 games, so that year was still useful. I also left out the final period for teams that ended up playing fewer or more than 182 games due to rainouts, playoff games, etc.)
I was left with 268 teams and 2,364 eighteen-game samples. For all the teams in the study, here's the world's first BLECH—Brandon's Lamely Expressed CHart—showing the distribution of wins in the 18-game samples across the data:
Since there are nine periods from teams pretty much hitting at least 90 wins, you would expect the periods to hover around 10-8, and the near-perfect bell curve says exactly that. There are a few more 12-6's than 9-9's, but since some of our teams won 100 games or more, that's to be expected.
The point is that the distribution of wins across bite-size samples is entirely normal—good teams have no special ability to either avoid slumps or go on tremendous winning streaks. The extreme outliers on the chart are the 1996 Padres (who went 3-15 in period 4) and the 1987 Brewers (who started 17-1 and then followed up by going 4-14).
Although the above chart shows that good teams have an entirely normal distribution of slumping and streaking, is there any difference as to when they do it? In particular, should we trust the early April sample? Consider the following chart, which tracks exceptional 18-game samples (12-6 or better), average ones (9-9 to 11-7), and losing ones (8-10 or worse) for all nine periods:
You can see that the first half of April is not particularly trustworthy and is perhaps slightly misleading. The distribution in the first period is 33 percent exceptional, 49 percent average, and a whopping 18 percent losing. In other words, early April undervalues the ultimate performance of our teams.
Apparently, good teams get their losing out of the way primarily in the first third of the season; the figures are 17-19 percent in those periods, and with the exception of period 5, they're 15 percent and lower from there, going all the way down to 11 percent in the final two periods. The hot streaks are generally distributed evenly, all at 31-33 percent except in periods 3, 8 and 9, which have rates of 37 percent, 42 percent, and 38 percent, respectively.
Given the large data pool, those variations seem meaningful—good teams have a weak-to-moderate tendency to put their slumps early and their streaks late. My guess is that the prevalence of early slumps results from tinkering that pays off later, but I don't know that. For severe slumps—periods of 6-12 or worse—the only noticeable distribution oddity is that there's just one of them in period 9: the 1964 Phillies. There are a lot of severe slumps in period 8 (see the 2006 Tigers), but apparently those teams recover from their stumbling.
Of course, if your team is in a severe slump, closer to 6-12 than 9-9, none of this is comforting. Very few teams have had two periods of 6-12 or worse and recovered to win 90 games or make the playoffs. Here's the complete list:
1965 Pirates90-72, did not make the playoffs
An incredibly bipolar team, starting off at 12-24 before ripping off a 16-2 section that was negated not too much later. This is a very freaky distribution, and although it is comforting to know that a team can survive that sort of opening horror, I'm sure that's not how you want your favorite team to get to 90 wins.
1984 Royals84-78, won the AL West
This team stood no chance against the Tiger juggernaut that year, but it actually wasn't bad towards the end, going 44-28 at the close of the season. It just took a while to find the groove. It certainly feels better going into the playoffs with that sort of finish than having all the highlights in April and May.
1995 Yankees79-65, won the AL Wild Card
Bonus points for pulling it off in a shortened season. The year 1995 was the first non-strike year that I followed baseball, but I don't remember the Yankees being this erratic (aside from an April or May issue of Baseball Weekly labeling their division as the "AL Least"). Had the Yanks been unable to pull things together by the end, the Angels' historic collapse would have been just a demotion to wild card, as New York needed every one of those 14 wins at the end to reach the playoffs at all.
Of the 268 teams in my list, 48 of them (17 percent) had a period of 6-12 or worse and survived. The percentage is about the same if you consider only the 183 playoff teams from those 268. However, of the 30 playoff teams that sneaked in with fewer than 90 wins, 21 of them (70 percent) have endured such a patch. The Tigers can always hope for a weak AL Central...
So if your team is struggling severely in the early going, be it Detroit this year or any team down the road, a bad record early on is not necessarily panic time; there's every reason to believe that the team will bounce back if the talent's there. Even a severe slump isn't enough by itself to derail a team's hopes for the playoffs. Although a bad start to a season doesn't tell you whether the team will make the playoffs or sink like a stone, more good teams slump early than slump late, and there appears to be no evidence whatsoever that a bad start by a good team is anything but random distribution taking its toll in April instead of July. Unless the Tigers have forgotten how to hit or are all injured without knowing it, it's more likely that chance is playing a big role in the record than that they flat-out stink.
If it's an absolute mystery why your favorite team can't catch a break when you had it pegged for domination, just wait a few weeks and see what happens. It might feel like your team is trying to dig out of a hole, but it's closer to the truth to say that it's holding serve until the team hits the better side of the bell curve. That perspective can be hard for a fan to adopt, but maybe it's enough to keep you sane during periods of frustration.
References and Resources
Baseball-Reference again. Surprise! Since it's hard to think of something to say every week for a site I reference in every article (and it should be assumed even if I don't list it that it's in the article somewhere), I'll just mention that Sean Forman, owner/operator of Baseball-Reference, is a very nice guy and found time at the last Winter Meetings to talk baseball with me for quite a while. I couldn't imagine getting a ridiculously wonderful baseball search engine and database from a nicer guy. It's not every day that sentence makes sense.
Also, thanks to Chicane and Peter Gabriel for writing completely different songs with the same name.
Brandon Isleib is a lawyer and writes about stuff sometimes. He can be reached via the electronic mails.