Post by Richie3Jack on Sept 28, 2010 8:32:42 GMT -5
Recently I’ve been working on a way to better predict team wins. Usually I use the QBRD (QB Rating Differential formula: QB Rating for the offense – QB Rating Allowed for the defense). This is still a very strong correlation to winning. Over the last 5 seasons it has had a correlation coefficient of +0.8, which is very strong…meaning that the higher the QBRD, the more likely a team is going to win more games and vice versa.
Still though, there were some problems as usually 2-3 teams a year did much worse or much better than their QBRD suggests.
A few of the teams that were shining examples of this were in the ’08 season with Denver and Carolina doing far better than their QBRD suggested and the Packers doing far worse than their QBRD suggested.
So essentially I needed to find other statistics that correlate to winning and losing in football and try to make the correlation coefficient higher (a statistic that tells you how strong the correlation of one thing is to another).
I finally discovered some other stats that helped me raise that correlation coefficient. What I did was combine the following:
QBRD
Adjust Sacks Allowed Ratio
Turnover Margin
Pts Scored Per Yards Gained
I already went over QBRD.
Adjusted Sacks Allowed Ratio is something created by FootballOutsiders.com. It’s essentially a statistic that takes into account how often a team’s QB is sacked per pass attempt thrown. But it also figures in things like down and distance, the opposing defense’s pass rush, etc. The correlation between ASAR and winning or losing in the NFL is so-so (about 0.54 over the year), but it pretty much explained away all of those teams that did far better or worse than their QBRD suggested. Essentially, if you did far better than your QBRD suggests, your ASAR was very good and vice versa. I think most would agree that the O-Line is a part of importance to winning in the NFL. If your QB gets sacked a lot and you have a good team, you are risking injury, rattling your QB and losing yardage. That’s the big flaw that I saw in the ’08 Packers. They couldn’t stop the run either that year, but I ran a bunch of stats on the run and just couldn’t come up with anything that was of any decent correlation to winning or losing in the NFL.
Turnover Margin is the statistic that most NFL coaches that are more statistically driven point to as the ‘#1 stat to winning or losing in the NFL.’ Unfortunately, that’s not true. Over the past 5 seasons, the correlation is at 0.61, which is solid, but certainly nowhere near in the same league as QBRD. Still though, it’s a helpful predictor of success for me because a good team that is strong everywhere can be derailed by a poor turnover margin.
Lastly I have points scored per yards gained. Simply put, here’s the formula:
(points scored / yards gained) * 1000
This has a correlation of 0.75, which is very good, but still not as strong as QBRD. I think this stat is important though because it shows the importance of a team not only being able to move the ball, but to turn it into points. The gripe is that a team could only average 30 yards a drive and score field goals each time. But that would require amazing field position and a great kicker. Truth is that the teams that rank well here are just great offenses because they have to move the ball in order to be in a position to score and then actually score points.
Anyway, here’s the current rankings and the team’s projected win total.
Obviously, these projected wins will fluctuate throughout the season and it doesn’t account for strength of schedule.
1……Indianapolis Colts……11.06
2……New York Jets……10.32
3……Philadelphia Eagles……10.22
4……Atlanta Falcons……9.76
5……Pittsburgh Steelers……9.72
6……New England Patriots……9.47
7……New Orleans Saints……9.29
8……Green Bay Packers……9.18
9……Chicago Bears……9.16
10……Cincinnati Bengals……8.82
11……Tennessee Titans……8.71
12……Seattle Seahawks……8.67
13……Tampa Bay Buccaneers……8.64
14……San Diego Chargers……8.33
15……Kansas City Chiefs……7.77
16……Dallas Cowboys……6.52
17……Miami Dolphins……6.47
18……St. Louis Rams……6.28
19……Washington Redskins……6.18
20……Denver Broncos……5.64
21……Cleveland Browns……5.51
22……Detroit Lions……4.75
23……Baltimore Ravens……4.60
24……New York Giants……4.08
25……Arizona Cardinals……4.02
26……Houston Texans……3.48
27……Oakland Raiders……3.43
28……Jacksonville Jaguars……2.57
29……Minnesota Vikings……2.46
30……Buffalo Bills……2.35
31……San Francisco 49ers……1.24
32……Carolina Panthers……0.63
3JACK
Still though, there were some problems as usually 2-3 teams a year did much worse or much better than their QBRD suggests.
A few of the teams that were shining examples of this were in the ’08 season with Denver and Carolina doing far better than their QBRD suggested and the Packers doing far worse than their QBRD suggested.
So essentially I needed to find other statistics that correlate to winning and losing in football and try to make the correlation coefficient higher (a statistic that tells you how strong the correlation of one thing is to another).
I finally discovered some other stats that helped me raise that correlation coefficient. What I did was combine the following:
QBRD
Adjust Sacks Allowed Ratio
Turnover Margin
Pts Scored Per Yards Gained
I already went over QBRD.
Adjusted Sacks Allowed Ratio is something created by FootballOutsiders.com. It’s essentially a statistic that takes into account how often a team’s QB is sacked per pass attempt thrown. But it also figures in things like down and distance, the opposing defense’s pass rush, etc. The correlation between ASAR and winning or losing in the NFL is so-so (about 0.54 over the year), but it pretty much explained away all of those teams that did far better or worse than their QBRD suggested. Essentially, if you did far better than your QBRD suggests, your ASAR was very good and vice versa. I think most would agree that the O-Line is a part of importance to winning in the NFL. If your QB gets sacked a lot and you have a good team, you are risking injury, rattling your QB and losing yardage. That’s the big flaw that I saw in the ’08 Packers. They couldn’t stop the run either that year, but I ran a bunch of stats on the run and just couldn’t come up with anything that was of any decent correlation to winning or losing in the NFL.
Turnover Margin is the statistic that most NFL coaches that are more statistically driven point to as the ‘#1 stat to winning or losing in the NFL.’ Unfortunately, that’s not true. Over the past 5 seasons, the correlation is at 0.61, which is solid, but certainly nowhere near in the same league as QBRD. Still though, it’s a helpful predictor of success for me because a good team that is strong everywhere can be derailed by a poor turnover margin.
Lastly I have points scored per yards gained. Simply put, here’s the formula:
(points scored / yards gained) * 1000
This has a correlation of 0.75, which is very good, but still not as strong as QBRD. I think this stat is important though because it shows the importance of a team not only being able to move the ball, but to turn it into points. The gripe is that a team could only average 30 yards a drive and score field goals each time. But that would require amazing field position and a great kicker. Truth is that the teams that rank well here are just great offenses because they have to move the ball in order to be in a position to score and then actually score points.
Anyway, here’s the current rankings and the team’s projected win total.
Obviously, these projected wins will fluctuate throughout the season and it doesn’t account for strength of schedule.
1……Indianapolis Colts……11.06
2……New York Jets……10.32
3……Philadelphia Eagles……10.22
4……Atlanta Falcons……9.76
5……Pittsburgh Steelers……9.72
6……New England Patriots……9.47
7……New Orleans Saints……9.29
8……Green Bay Packers……9.18
9……Chicago Bears……9.16
10……Cincinnati Bengals……8.82
11……Tennessee Titans……8.71
12……Seattle Seahawks……8.67
13……Tampa Bay Buccaneers……8.64
14……San Diego Chargers……8.33
15……Kansas City Chiefs……7.77
16……Dallas Cowboys……6.52
17……Miami Dolphins……6.47
18……St. Louis Rams……6.28
19……Washington Redskins……6.18
20……Denver Broncos……5.64
21……Cleveland Browns……5.51
22……Detroit Lions……4.75
23……Baltimore Ravens……4.60
24……New York Giants……4.08
25……Arizona Cardinals……4.02
26……Houston Texans……3.48
27……Oakland Raiders……3.43
28……Jacksonville Jaguars……2.57
29……Minnesota Vikings……2.46
30……Buffalo Bills……2.35
31……San Francisco 49ers……1.24
32……Carolina Panthers……0.63
3JACK