First 2005 Post Mortem

[Guest Rotblatt]

Well, JD, unbeknownst to me, SoSH had a similar thread goin' on and after 3 days of rooting through data in a more sophisticated way than us, here's the inital result for the AL:

]My thought was that "consistency" is not an attribute that implies either more or less runs, but that it moved the run profile to a more valuable shape. If that were true, then you'd expect "consistent" teams (as indicated by the %stdev metric here) to tend to outperform their pythagorans. Sadly, I can't really see that's true. correl bp 0.327504361 correl dif1 -0.64534997 correl dif2 -0.294767694 correl dif3 -0.258967357 http://www.baseballprospectus.com/statistics/standings.php %st = the %stdev metric employed earlier, or stdev/average runs BP Diff = the difference between actual runs and BP's expected runs Dif1, Dif2, and Dif3 are the deltas between actual wins and W1, W2, and W3 (which are various and increasingly complicated ways of estimating expected wins). At bottom are the correlations between each of the last 4 columns and the %stdev column. I don't see anything of value here. Does anyone else? Maybe the %stdev metric doesn't measure consistency meaningfully, or maybe consistency really just doesn't have much value? Or is so evenly distributed among teams that none of it really matters. I'm not very good with stats, so am wide open to suggestions here. If anything it's a negative correlation -- but that might be messed up by the fact that there's interleague play, and the AL teams have underperformed relative to the NL teams (even while beating them more often overall). At least, in W3, which adjusts for strength of schedule, AL teams are down -24. Probably NL teams should be mixed in here, but I didn't want to combine them originally. Should go back and do other years, but this is pretty time consuming.

I'm not quite sure how he got there, otherwise I'd try & replicate it for the NL, but maybe he'll go ahead and examine the NL data and throw that in there.

Cribbing is MUCH easier than doing it yourself.

[seawolf17]

I figured this was the best place to put this, because it's just too sad to make its own thread.

Magic Number Watch
Mets Division Elimination: 7
Mets WC Elimination: 12

[Guest sharpie]

Do we really need it?

[Guest Yancy Street Gang]

I don't think we need it.

Hopefully, if the Yankees get close to elimination, that might be a fun number to track. But we know the Mets are out of it. There's no need to track the math.

[Guest Rockin' Doc]

We don't really need a number to track when the Mets are "mathematically" eliminated from the playoffs. For all intents and purposes they were eliminated by the time they left Atlanta.