- The LoHud Mets Blog - http://mets.lohudblogs.com -
Some Follow Up on Predictive Value
Posted By Howard Megdal On May 23, 2012 @ 8:00 am In Today's Mets headlines | Comments Disabled
My post from this morning led to some interesting discussion here and elsewhere . One key point that was brought up had to do with the predictive value of 42 games of Pythagorean record. So I thought I’d put that recent history to the test.
Year WP (first 42 games) Pythag (first 42 games) WP (Last 120 games) Pythag (Last 120 games)
2011 .476 .495 .475 .482
2010 .476 .518 .492 .497
2009 .548 .557 .392 .405
2008 .512 .525 .563 .560
2007 .667 .641 .500 .495
So what do we learn from five years of data? Well, for one thing, how a team performs in its first 42 games often deviates from how it performs in its last 120 games. I mean, I think we all knew that, as I remember 2007 all too well, for instance. But it is worth keeping in mind that there’s no guarantee the Mets will continue scoring and allowing runs at such a poor rate. And lest we ignore the reality, they could also get worse.
More to the point: the actual results on the field, whether over 42 games or 120 games, hew pretty closely to the expected W-L record. Over the past five years for the Mets, the gap has been smaller in 120 game samples, but is still awfully close in 42 game samples.
Ultimately, how predictive you believe the first 42 games to be has a lot more to do with whether you expect the Mets to continue playing as they have. But if they do continue playing as they have, they’re much likelier to win games at a .417 pace, as per their Pythagorean record, than at their .524 winning percentage so far.
Put another way: the Mets are on pace to win just under 85 games. Here’s their Pythagorean win total for every year they’ve won between 80 and 90 games:
Year Wins Pythag Wins
2012 85 68
2008 89 89
2007 88 86
2005 83 89
2001 82 73
1998 88 88
1997 88 88
1989 87 91
1984 90 78
1976 86 91
1975 82 83
1973 82 83
1972 83 72
1971 83 86
1970 83 88
So, to recap, in 14 seasons above, Pythagorean record had the win total exactly right three times, within three wins seven times, and missed by double digits only twice in 14 times, with a max spread of 12 wins.
At present, the Mets are winning at a pace 17 games ahead of their Pythag record.
It all comes back to the same conclusion: playing as the Mets have to date, they aren’t likely to win nearly as many games as they have so far. The underlying performance needs to change for them to contend, or they need to buck more than a century of major league history.
Article printed from The LoHud Mets Blog: http://mets.lohudblogs.com
URL to article: http://mets.lohudblogs.com/2012/05/23/some-follow-up-on-predictive-value/
URLs in this post:
 here and elsewhere: http://metspolice.com/2012/05/22/this-says-the-mets-will-win-73-games-i-say-nay/
 Image: http://www.blinklist.com/index.php?Action=Blink/addblink.php&Url=http%3A%2F%2Fmets.lohudblogs.com%2F2012%2F05%2F23%2Fsome-follow-up-on-predictive-value%2F&Title=Some%20Follow%20Up%20on%20Predictive%20Value
 Tweet: https://twitter.com/share