SITE NEWS: We are moving all of our site and company news into a single blog for Sports-Reference.com. We'll tag all PFR content, so you can quickly and easily find the content you want.

Also, our existing PFR blog rss feed will be redirected to the new site's feed.

Pro-Football-Reference.com » Sports Reference

For more from Chase and Jason, check out their work at Football Perspective and The Big Lead.

Records in close games

Posted by Doug on July 28, 2006

This is just a quick one: team's records in close (decided by seven or fewer points) and non-close games:


TM YR Close NonClose
=========================
pit 2005 2- 4 9- 1
sdg 2005 2- 5 7- 2
ind 2005 3- 0 11- 2
nyg 2005 3- 3 8- 2
den 2005 4- 2 9- 1
car 2005 4- 3 7- 2
was 2005 5- 5 5- 1
gnb 2005 2- 8 2- 4
sea 2005 6- 2 7- 1
atl 2005 3- 4 5- 4
ari 2005 2- 5 3- 6
buf 2005 1- 4 4- 7
det 2005 2- 5 3- 6
tam 2005 6- 3 5- 2
dal 2005 6- 5 3- 2
cin 2005 4- 1 7- 4
kan 2005 4- 2 6- 4
cle 2005 4- 5 2- 5
chi 2005 5- 1 6- 4
hou 2005 1- 6 1- 8
mia 2005 5- 3 4- 4
ten 2005 1- 4 3- 8
bal 2005 3- 4 3- 6
phi 2005 5- 5 1- 5
nor 2005 2- 5 1- 8
nyj 2005 2- 4 2- 8
oak 2005 2- 4 2- 8
min 2005 4- 1 5- 6
jax 2005 8- 2 4- 2
stl 2005 4- 4 2- 6
nwe 2005 5- 1 5- 5
sfo 2005 4- 4 0- 8

I've attempted to sort the list so that teams that did relatively better in non-close games are at the top. Though it's not important, the actual formula I used was this:


(NonCloseWins - NonCloseLosses) - (CloseWins - CloseLosses)

Fact 1: the correlation between close game winning percentage in Year N and overall winning percentage in Year N+1 is .08.

Fact 2: the correlation between non-close game winning percentage in Year N and overall winning percentage in Year N+1 is .30.

Fact 3: the correlation between overall winning percentage in Year N and overall winning percentage in Year N+1 is also .30.

Fact 4: if you regress Year N+1 overall winning percentage on the two independent variables: Year N close game winning percentage and Year N non-close game winning percentage, the first one (close games) is highly insignificant.

In other words, from a predictive standpoint, the first column is just noise. If you're thinking about, say, the Patriots and their 2006 prospects, instead of saying, "Well, they were 10-6 last year, they gained Players A, B, and C, and lost players X, Y, and Z," you should probably be saying, "Well, they were 5-5 last year, they gained Players A, B, and C, and lost players X, Y, and Z."

This entry was posted on Friday, July 28th, 2006 at 5:24 am and is filed under General. You can follow any responses to this entry through the RSS 2.0 feed. Both comments and pings are currently closed.