thebandit27 Posted April 22, 2013 Posted April 22, 2013 LOLZ Premise: shot In 2012 Teams coming off Thurs games were 16-18 Teams coming off byes were 15-15-2 Soooooooooooooooo... :-) /thread?
MattM Posted April 22, 2013 Posted April 22, 2013 LOLZ Premise: shot In 2012 Teams coming off Thurs games were 16-18 Teams coming off byes were 15-15-2 Soooooooooooooooo... :-) Not really--unless you controlled for other variables like W-L record of those playing. That's why things like the 10 year data set linked by Poo above that showed ten years of results (anomalies more likely to even out) that showed effectively a 4 to 5% advantage are more reliable....
The Big Cat Posted April 22, 2013 Posted April 22, 2013 Not really--unless you controlled for other variables like W-L record of those playing. That's why things like the 10 year data set linked by Poo above that showed ten years of results (anomalies more likely to even out) that showed effectively a 4 to 5% advantage are more reliable.... I understand, but is 4-5% statistically relevant over 6 occurrences, i.e. the 2013 bye/Thurs opponents?
Doc Posted April 22, 2013 Posted April 22, 2013 Not really--unless you controlled for other variables like W-L record of those playing. That's why things like the 10 year data set linked by Poo above that showed ten years of results (anomalies more likely to even out) that showed effectively a 4 to 5% advantage are more reliable.... The overall trend is more important, unless/until the last 2 years' trend proves to be the new normal and not aberrations. Going along with your controlling for other variables, I'd like to know the records of the teams coming-off first-round byes in the playoffs, since those teams are more closely matched (being playoff teams). I wonder what the players think? If they don't care or rather not play teams after the bye, perhaps they should get rid of them? Or if they really want that bye, have teams coming off byes play each other the following week?
Cash Posted April 22, 2013 Posted April 22, 2013 The overall trend is more important, unless/until the last 2 years' trend proves to be the new normal and not aberrations. Going along with your controlling for other variables, I'd like to know the records of the teams coming-off first-round byes in the playoffs, since those teams are more closely matched (being playoff teams). I wonder what the players think? If they don't care or rather not play teams after the bye, perhaps they should get rid of them? Or if they really want that bye, have teams coming off byes play each other the following week? I read a great article a year or two ago (that I sadly can't find anymore) that broke down the bye week advantage into home and road splits. What it found was fascinating: home teams coming off a bye essentially weren't helped by it -- they won about 55% of their games, but that's consistent with home-field advantage in general. The big jump was for road teams, who won over half their post-bye games, compared to about 45% normally. The specific numbers are from memory and not necessarily accurate, but the important point was that home-team winning percentage didn't increase after a bye, but road-team winning percentage did. Personally, I don't see any reason why last year's aberrant result would be anything other than a fluke, and I'll be very surprised if it continues. And honestly, it's probably not really aberrant in the true statistical sense. If 54% is the true winning percentage of post-bye teams, the odds of producing a 50% winning percentage in a 26-game sample are pretty good. (It's a 26-game sample instead of 32 because 6 teams played each other when both were coming off a bye: Falcons/Eagles, Bills/Texans, and 49ers/Rams. Those 6 teams had to have a combined .500 record in those games no matter what.) As for Thursday games, we didn't have a lot of data prior to last year, but it seemed as though having an extra 3 days would be a similar but smaller advantage than the bye-week advantage. Maybe it isn't for some reason, but 1 season of 47% win percentage doesn't tell us much. If the true win percentage of post-Thursday teams is something like 51% or 52%, you should expect to see them go a little below .500 sometimes.
Recommended Posts