05-24-2015, 07:14 PM
http://grantland.com...-us-about-2015/
Season-long statistics have a weaker predictive value in football than they do for just about every other professional sport, specifically because the season is much shorter. You can fit 10 NFL seasons into one MLB campaign with a couple of games to spare. The 16-game schedule means we learn less about the true talent level of each team in football than we do in other sports.
The good news is that we can turn that lack of information into a positive. Because 16 games just aren’t enough to learn much about a team, we can usually safely say that teams that exhibit some extreme characteristic or have some event occur a freakishly high (or low) amount of the time will not have that same experience over the next 16 games.
In every sport I’ve ever examined, the difference between the number of points a team scores and the points it allows is a better indicator of its future win-loss record than its actual win-loss record over that same time span. Pro football is no exception. whereresearch by Daryl Morey (yes, the same Daryl Morey who is very sad about last night’s Rockets game) found that a variant of Bill James’s Pythagorean Expectation formula can estimate what each team’s win-loss record “should” have been, given their points scored and allowed.
(can't cut and paste the chart, but it shows that in 2014 the Bengals won 1.9 games more than they should have based on point differential. That was second highest in the league)
Point differential isn’t exactly complex, but it’s even easier to count up a team’s wins and losses in games decided by seven points or fewer and see whether they were able to produce a similar record the following year. The vast, vast majority of teams win neither a particularly high nor a particularly low percentage of their close games on an annual basis. Most of the differences between teams comes in what they do in the other games, the ones that are decided by two scores or more.
(again I can't cut and paste the chart, but the Bengals were also second best in the league last year in games decided by 7 points or fewer going 3-0-1)
Basically what this article is saying is that the Bengals were not as good as their record last year and are likely to regress toward the mean in 2015. It is an interesting read. I just posted a very small portion of it.
Season-long statistics have a weaker predictive value in football than they do for just about every other professional sport, specifically because the season is much shorter. You can fit 10 NFL seasons into one MLB campaign with a couple of games to spare. The 16-game schedule means we learn less about the true talent level of each team in football than we do in other sports.
The good news is that we can turn that lack of information into a positive. Because 16 games just aren’t enough to learn much about a team, we can usually safely say that teams that exhibit some extreme characteristic or have some event occur a freakishly high (or low) amount of the time will not have that same experience over the next 16 games.
In every sport I’ve ever examined, the difference between the number of points a team scores and the points it allows is a better indicator of its future win-loss record than its actual win-loss record over that same time span. Pro football is no exception. whereresearch by Daryl Morey (yes, the same Daryl Morey who is very sad about last night’s Rockets game) found that a variant of Bill James’s Pythagorean Expectation formula can estimate what each team’s win-loss record “should” have been, given their points scored and allowed.
(can't cut and paste the chart, but it shows that in 2014 the Bengals won 1.9 games more than they should have based on point differential. That was second highest in the league)
Point differential isn’t exactly complex, but it’s even easier to count up a team’s wins and losses in games decided by seven points or fewer and see whether they were able to produce a similar record the following year. The vast, vast majority of teams win neither a particularly high nor a particularly low percentage of their close games on an annual basis. Most of the differences between teams comes in what they do in the other games, the ones that are decided by two scores or more.
(again I can't cut and paste the chart, but the Bengals were also second best in the league last year in games decided by 7 points or fewer going 3-0-1)
Basically what this article is saying is that the Bengals were not as good as their record last year and are likely to regress toward the mean in 2015. It is an interesting read. I just posted a very small portion of it.
