View unanswered posts | View active topics It is currently October 2nd, 2014, 2:44 am



Reply to topic  [ 8 posts ] 
 Statistical Analysis of NFL 
Author Message
QB Coach
User avatar

Joined: August 21st, 2005, 3:36 am
Posts: 3130
Location: Philadelphia, PA
Post Statistical Analysis of NFL
So in another thread I mentioned that I was going to try to run some statistical analysis of a few NFL stats. I'm in the process of doing this now. I'm basically taking all the stats from nfl.com for every season back to 2001, putting them in excel, and seeing what I can come up with. So far I'm running a correlation between wins and total offense yds/game, yds/play, and time of possession. Then doing the same with wins against rushing yds/carry & yds/game, and passing yds/att & yds/game. Then I'm doing defensive stats vs. winning (yds/game, yds/play, rushing yds/play, yds/game, and passing yds/play, yds/game. I'll also probably run a few regressions agains the same stats.

Before I do it all, anyone have any other specific stats they'd like me to check? (yes I know I'm a big geek)

_________________
"Facts are stubborn things; and whatever may be our wishes, our inclinations, or the dictates of our passions, they cannot alter the state of facts and evidence." - John Adams

“The good thing about science is that it's true whether or not you believe in it.” - Neil deGrasse Tyson


October 29th, 2011, 12:26 pm
Profile
Heisman Winner
User avatar

Joined: July 14th, 2005, 11:58 am
Posts: 782
Post Re: Statistical Analysis of NFL
Touchdown Jesus wrote:
So in another thread I mentioned that I was going to try to run some statistical analysis of a few NFL stats. I'm in the process of doing this now. I'm basically taking all the stats from nfl.com for every season back to 2001, putting them in excel, and seeing what I can come up with. So far I'm running a correlation between wins and total offense yds/game, yds/play, and time of possession. Then doing the same with wins against rushing yds/carry & yds/game, and passing yds/att & yds/game. Then I'm doing defensive stats vs. winning (yds/game, yds/play, rushing yds/play, yds/game, and passing yds/play, yds/game. I'll also probably run a few regressions agains the same stats.

Before I do it all, anyone have any other specific stats they'd like me to check? (yes I know I'm a big geek)


What about Average drive start position and punt and kickoff return yards? Adding some sort of Special teams analysis would be interesting to see as well.


October 29th, 2011, 1:10 pm
Profile
Player of the Year - Defense

Joined: September 25th, 2007, 3:20 am
Posts: 2783
Post Re: Statistical Analysis of NFL
If you haven't checked them out, also check out profootballfocus.com. More statistically correct then even NFL at times.


October 29th, 2011, 1:38 pm
Profile
QB Coach
User avatar

Joined: August 21st, 2005, 3:36 am
Posts: 3130
Location: Philadelphia, PA
Post Re: Statistical Analysis of NFL
Roy,Roy,Roy...Roy wrote:
Touchdown Jesus wrote:
So in another thread I mentioned that I was going to try to run some statistical analysis of a few NFL stats. I'm in the process of doing this now. I'm basically taking all the stats from nfl.com for every season back to 2001, putting them in excel, and seeing what I can come up with. So far I'm running a correlation between wins and total offense yds/game, yds/play, and time of possession. Then doing the same with wins against rushing yds/carry & yds/game, and passing yds/att & yds/game. Then I'm doing defensive stats vs. winning (yds/game, yds/play, rushing yds/play, yds/game, and passing yds/play, yds/game. I'll also probably run a few regressions agains the same stats.

Before I do it all, anyone have any other specific stats they'd like me to check? (yes I know I'm a big geek)


What about Average drive start position and punt and kickoff return yards? Adding some sort of Special teams analysis would be interesting to see as well.

Not sure if I can find those specific stats. Kickoff return yards I think I can get. I'll see what I can do.

So far what I'm seeing is that (as I kind of expected) there aren't any single stats that really show who will win or lose. But there are some surprising things showing that some stats are really not good predictors of winning at all.

I'll have some details hopefully a bit later today.

_________________
"Facts are stubborn things; and whatever may be our wishes, our inclinations, or the dictates of our passions, they cannot alter the state of facts and evidence." - John Adams

“The good thing about science is that it's true whether or not you believe in it.” - Neil deGrasse Tyson


October 30th, 2011, 11:44 am
Profile
QB Coach
User avatar

Joined: August 21st, 2005, 3:36 am
Posts: 3130
Location: Philadelphia, PA
Post Re: Statistical Analysis of NFL
Ok, this is turning out to be more difficult/confusing than I had hoped, but here's where I'm at.

(Disclaimer: these stats are all through 7 weeks and do not include week 8 )
I took all the stats from nfl.com from 2001-2011 and put them into excel. I mostly focused on things like totals yds/game, avg. yds/play, time of possession, rushing yds/att & yds/game, passing yds/att & yds/game. I also looked at the same stats on defense. I then looked at correlation of the stats with wins, as well as trying some regressions to see how things really shake out. Here's what I found for correlation since 2006 (haven't had time to go through everything else:

Stats Most Highly Correlated With Wins (in order of most correlated to least)
2011

Offense
1. pass yds/att
2. pass yds/game
3.ToP
4. total yds/play
5. total yds/game.
Interestingly, rushing yds/att was the least correlated with winning. In fact, it is negatively correlated (which means that for every increase by 1 in rushing yds/att, wins decrease by .098.
Defense
1. pass yds/att
2. rush yds/game
3. total yds/game
4. total yds/play
5. pass yds/game
Again, rush yds/att is negatively correlated with winning.

2010
Offense
1. ToP
2. pass yds/att
3. total yds/game
4. total yds/play
5. rush yds/game
6. pass yds/game
7. rush yds/carry
No negative correlations this time

Defense
1. rush yds/game
2. pass yds/att
3. total yds/game
4. total yds/play
5. rush yds/att
6. pass yds/game
No negative correlations

2009
Offense
1. total yds/play
2. pass yds/att
3. total yds/game
4. pass yds/game
5. ToP
6. rush yds/game
Rush yds/att was negatively correlated with winning.

Defense
1. rush yds/game
2. pass yds/att
3. total yds/game
4. total yds/play
5. rush yds/att
6. pass yds/game
No negative correlations

2008
Offense
1. ToP
2. total yds/game
3. rush yds/game
4. pass yds/att
5. total yds/play
6. pass yds/game
7. rush yds/carry
no negative correlations

Defense
1. total yds/game
2. rush yds/game
3. pass yds/att
4. total yds/play
5. rush yds/att
6. pass yds/game

2007
Offense
1. pass yds/att
2. total yds/game
3. total yds/play
4. pass yds/game
5. ToP
6. rush yds/game
7. rush yds/att

Defense
1. pass yds/att
2. total yds/play
3. rush yds/game
4. total yds/game
5. rush yds/att
6. pass yds/game

2006
Offense
1. total yds/game
2. ToP
3. total yds/play
4. pas yds/att
5. rush yds/game
6. pass yds/game
7. rush yds/carry

Defense
1. pass yds/att
2. total yds/game
3. total yds/play
4. pass yds/game
5. rush yds/game
6. rush yds/att

What this all works out to, from a correlation standpoint, is that (over the past 5 seasons) the stats most correlated with wins are:

Overall
1. Offense total yds/gm
2. Offense pass yds/att
3. Offense total yds/play
4. Defense pass yds/att
5. ToP
6. Defense rush yds/game
7. Defense yds/game
8. Offense pass yds/game
9. Defense total yds/play
10. Offense rush yds/game
11. Defense pass yds/game
12. Defense rush yds/att
13. Offense rush yds/att

Breaking it down:

Offense
1. total yds/game
2. pass yds/att
3. total yds/play
4. ToP
5. pass yds/game
6. rush yds/game
7. rush yds/carry

Defense
1. pass yds/att
2. rush yds/game
3. total yds/game
4. total yds/play
5. pass yds/game
6. rush yds/att

Whew! I know that's a really long post, but it's a lot of data. Anyway, my analysis is this:

The most interesting thing here is that offensive rushing is not highly correlated with winning. Either in terms of yds per carry or yds per game, being a strong rushing team just doesn't seem to be (statistically speaking) a key to wins. Not surprisingly, the keys seem to be gaining a lot of total yards on offense, as well as having a higher yds/play average. This seems to point to having big play capability as a key to winning games. Combine that with a defense that doesn't give up a lot of big plays (low pass yds/att) and you have a winning combo.

I will add that I know this may be confusing or not really be revealing any earth shattering insights. But, I think it's interesting to see how the numbers actually shake out. I'll try to do some more later when I have more time.

_________________
"Facts are stubborn things; and whatever may be our wishes, our inclinations, or the dictates of our passions, they cannot alter the state of facts and evidence." - John Adams

“The good thing about science is that it's true whether or not you believe in it.” - Neil deGrasse Tyson


November 2nd, 2011, 10:40 am
Profile
Heisman Winner
User avatar

Joined: July 14th, 2005, 11:58 am
Posts: 782
Post Re: Statistical Analysis of NFL
Do you think the rushing yards per attempt having a negative impact on winning might be skewed because of the kneel down, that would count as a rush for negative yards correct?


November 2nd, 2011, 10:58 am
Profile
QB Coach
User avatar

Joined: August 21st, 2005, 3:36 am
Posts: 3130
Location: Philadelphia, PA
Post Re: Statistical Analysis of NFL
aManNamedSuh wrote:
Do you think the rushing yards per attempt having a negative impact on winning might be skewed because of the kneel down, that would count as a rush for negative yards correct?

It could be, but I would guess that there aren't enough kneel down plays to really impact the rushing average. If you think about it, not every game has kneel downs at the end. Even when there are, there are usually only 2 or 3 at most. That's minuscule in the grand scheme of things, especially over a 5 year period. So my guess would be that it doesn't have a major effect.

_________________
"Facts are stubborn things; and whatever may be our wishes, our inclinations, or the dictates of our passions, they cannot alter the state of facts and evidence." - John Adams

“The good thing about science is that it's true whether or not you believe in it.” - Neil deGrasse Tyson


November 2nd, 2011, 7:35 pm
Profile
Pro Bowl Player
User avatar

Joined: April 19th, 2005, 2:10 pm
Posts: 2478
Location: Michigan
Post Re: Statistical Analysis of NFL
and if anyone wants to know what the Lions are building, Check out passing yards per attempt on offense on game's they've won and passing yards per attempt on defense in general.

That's how they are building a winner.

_________________
[b]New York Giants 26 - San Fransisco 3 - CandleStick Park - Who's got it better than us???[/b]


November 4th, 2011, 1:29 am
Profile WWW
Display posts from previous:  Sort by  
Reply to topic   [ 8 posts ] 

Who is online

Users browsing this forum: No registered users and 2 guests


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum

Search for:
Jump to:  
cron
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group.
Designed by STSoftware.