Rubbing Mud: The Cole ... (07/02)
the archives are now free.
3 comments have been left for this article. BP Comment Quick Links
| |
|
|
July 3, 2015 Rubbing MudThe Variation of All Things
About a week and a half ago, a gem of an edition of the Joe Sheehan Newsletter landed in my inbox. Joe’s topic du jour was the Mets, whose hot start he had dismissed somewhat ruthlessly back in April, and whose fans had pushed back pretty hard. (The Mets had fallen back to .500 at the time of the piece, and they’re back there again now.) Joe didn’t confine himself to the Mets alone, though. His thesis, a favorite of his (and—by no coincidence—mine), was that we tend to underestimate the degree of fluctuation in performance teams and players can experience, even over significant chunks of a season. A snip: I've been accused of the Gambler's Fallacy, over and over again, and I don't buy it, because baseball teams are not dice and baseball games are not independent trials. Baseball teams do play better sometimes and play worse sometimes; that those stretches are not predictable at all doesn't make them any less real. The Phillies scored 19 runs in ten games, then lit up the Cardinals for nine, the Yankees for 11 and the Yankees again for 11 more. That's baseball. The Mets went 13-3 and then 23-33. That's baseball. In other words, while the best words in our vocabulary for the swings in performance within a season are words like ‘variance’ and ‘randomness,’ those words fall short of describing the phenomenon. The season is so long, with such a demanding and uneven schedule, that teams can have sustained performances that reach to the extremes, and those stretches can be backed by real things, even though they’re ultimately unexplainable (except in hindsight, which is no way to go about explaining something). More Sheehan: Remember, I didn't cherry-pick those correlations above. The Gregorians did. [He’d just run through some monthly records.] If I were cherry-picking, I'd probably support my case by pointing at the Phillies going 11-23, then 6-0, then 6-24. The Dodgers followed up 22-10 with 17-23. The Pirates went 12-15, then 27-12. These aren't outliers. These are just baseball teams bouncing through 162 games the way teams do every single season. In that light, I thought I would mention this: Our Playoff Odds report through the games of July 1 shows the Cardinals averaging 96.8 wins at season’s end, across the 10,000 simulations we use to build the report. At the dawn of the season, the same simulation regimen—not yet privy to the season-ending injuries that would befall Adam Wainwright and Matt Adams—churned up an average Cardinals win total of 87.2. Halfway into the season, despite losing two key cogs, St. Louis has gotten off to a hot enough start to run away with the NL Central, at least in the computers’ estimation. But is that true? We have the entire two-Wild Card Era’s Playoff Odds Reports archived. You can see them yourself: just add “?maxdate=YYYY-MM-DD” with the date you want to see to the end of the URL of the normal Odds Report. To save you a little time, though, I went through each of the seasons we have, and looked up the projected win totals for every team on Opening Day, then on the Fourth of July, and then I looked up their final win totals. Here’s all of that in tables! 2012 American League
Now, these projected records bake in a little more than just the variance Sheehan described in his column. They also move in response to injuries and trades. Still, there’s some evidence that his theory is correct here, and some to contradict it: banked deviation from an expectation does seem to matter, but maybe that’s only true if the expectation was fundamentally broken in the first place. It appears that very large changes over that first half overwhelmingly indicate a real and lasting change, whereas any shift of fewer than seven games is dicey. There are exceptions, though. Consider the 2014 A’s and Brewers, both of whom saw their projected win total shoot up during the first half, only to end up more or less right where their preseason expectations had them. Obviously, I don’t have a crystal ball here. I can’t offer firm answers on whether Sheehan is right or wrong. I find the notion of real but intangible variance compelling, but the math is on the side of those who would treat St. Louis’s 108-win first-half pace as a virtual knockout punch to the Pirates and Cubs. Then again, maybe the way to parse this is to use history as a guide to the numbers. When you see the Royals go from 76.5 to 87.2 projected wins, as they already have this season, mentally note that there’s a chance (perhaps one in six, perhaps one in 10) that they come all the way back to Earth, and a chance (larger, though no more than 30 percent) that they end up somewhere around .500. That still leaves a 60-percent chance, or thereabouts, that Kansas City turns this hot start into somewhere north of 85 wins, a better than 50/50 shot that they make it to the Postseason. The Playoff Odds Report gives the Royals a 77.4-percent chance of reaching October, so Royals fans probably won’t appreciate having 20 points shaved off that number. If you see the season the way Sheehan does, though, that’s the thing to do.
Matthew Trueblood is an author of Baseball Prospectus. Follow @MATrueblood
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Free Agent Watch: Week 13
It would be interesting to see each team's prorated record on July 4th as well, so we could compare whether the updated projection was a better predictor of the final number of wins than just using the team's W/L record on July 4th.
Agreed. I was expecting to find 2015s number to compare.