Forum > Hot Streaks and SABER and the BOOK
All pythag records do is give us a clearer indicator of a team's true performance level than do actual records. This has been shown through regression analyses that show pythag records more closely correlating with final W-L records than actual W-L records at most points in the season.
As far as the latest winning streak go, we improved our projected record mainly by winning games. But we also did it while putting up an impressive improvement of 53 runs in our season run differential. Had we won 12 straight games by just 1 run, our pythag record would be several games worse than our actual record, saying we had been very fortunate to turn such a small overall run differential into 12 straight wins. But we've won these 12 games by a total of 77-24... which equates to an 11-1 pythag record for that stretch. So we haven't been lucky: just truly much better than our opponents during this run.
Rangers100
Thanks @rangers100 I thought with the pitching and hitting and defense combined it would make since for the projected record to climb.
mhilgtx
But we also did it while putting up an impressive improvement of 53 runs in our season run differential.
One thing I'd absolutely love to see (but have no idea how to run down, beyond manual calculations) is the biggest run differential between the Rangers and their opponents over an X-game stretch. For example, was the Rangers' +53 margin over those 12 games their single-best mark over a 12-game stretch in franchise history? Perhaps, but perhaps not.
Joey Matschulat
Interesting point Joey, and you would have think that would have to be close to correct. Have the Rangers ever had pitching this good top to bottom paired with hitting. I know they have had pitching back in the old days, I know they have had hitting in the last few decades but did they have it all like now?
mhilgtx


So I am grinding my way through "The Book: Playing the percentages in baseball" and as many of you know there is a chapter about streaks both good and bad. Now you can quarrel with the way the author goes about slicing and dicing the numbers. However I suspect the general findings that players tend to play better after periods of playing well and playing poorly to make sense. Basically everyone regresses to the mean of their natural abilities.
Now all of you guys that are old hands at this game and have a deep rich knowledge of how to look at the game from a statistical side tell me what this stretch tells us? I am interest in the Pythagorean analysis part as well. Did this stretch realistically improve the projected record?