Great win! Holy Turnovers. Regression?
Another dominating performance against the Bills. Nasty. Jacks our TO differential to +27 for the season. Now I know we are killing it in several other more consistent facets of the game (e.g. Pass Yds/Attempt), but I was just wondering what sort of regression we should expect in the TO/game category? Obviously because there is so much variability in this stat - tips, bad throws, deflections, etc.
If a TO equals about 3 or 4 points in scoring differential, which I have seen bandied about in numerous sabermetric articles, we have about 90 to 105 points scores from such events this year. Our overall point differential after week 15 is +174 pts or +11.6 a game. Not including strength of schedule or other refinements. Assuming, and this is the question posed here, a regression of 25%, 50% or 75% we can expect approximately from 25 to 75 less points, or roughly 2 to 5 points less per game (differential) when luck evens out in games going forward. In addition to all other regressions of course.
I did find one article on the matter. And it seems to point toward a heavy regression. But this particular study was year to year when we would be better served looking at intra-year with the same personnel (TFB), coaching, game planning, etc. So I would expect somewhat a bit more stability in our case. Check if inclined:
linky
Anyway just some food for thought. Not too much a big deal, even a true +3.5 point strength advantage equals about a 60% win probability. We're looking pretty good at +11.6 (w/home field advantage to boot) even after it gets whittled down a bit. But I think it's something we shouldn't expect as much as every game going forward. So I expect we will get more typical NFL nail biter games in the playoffs. Not that that is any great revelation.
Anyway Merry Xmas and Season Greetings and such PatsFans. Pumped watching this team this year. Another good W!
