Power Rankings Week 14 Using the Pythagorean Wins Formula...

[Disnardo]

As some of you know this formula is based on team Points Scored and Points Scored Against not on team wins or opponents records. It presents a statistical analysis win %...

1. Colts (12-0) = .693...
2. Seahawks (10-2) = .619...
3. Bears (9-3) = .612...
4. Chargers (8-4) = .609...
5. Panthers (9-3) = .599...
6. Giants (8-4) = .594...
7. Broncos (9-3) = .583...
8. Bengals (9-3) = .577...
9. Jaguars (9-3) = .559...
10. Cowboys (7-5) = .552...
11. Steelers (7-5) = .549...
12. Chiefs (8-4) = .539...
13. Falcons (7-5) = .538...
14. Bucs (8-4) = .531...
15. Redskins (6-6) = .528...
16. Packers (2-10) = .496...
17. Patriots (7-5) = .478...
18. Dolphins (5-7) = .477...
19. Browns (4-8) = .460...
20. Raiders (4-8) = .456...
21. Rams (5-7) = .455...
22. Vikings (7-5) = .445...
23. Eagles (5-7) = .442...
24. Cardinals (4-8) = .441...
25. Lions (4-8) = .440...
26. Titans (3-9) = .428...
27. Bills (4-8) = .426...
28. Ravens (4-8) = .400...
29. Saints (3-9) = .382...
30. Jets (2-10) = .351...
31. Forty Niners (2-10) = .350...
32. Texans (1-11) = .349...

I highlighted those teams still in our schedule and seems like our hardest game is against SD, though "on any given Sunday."...

 

[dolphan117]

Well I know nothing about the forula but I have thought all year the Chargers are better than their record. Hard to beleive they could miss the playoffs.

 

[BennyVW]

Unluckily the formula is flawed. It shows green bay ahead of us.. Most would ask why a 2-10 team is ranked higher. Well the reason is on their first win they madly racked up the score against the saints. final score 52-3. This wildly changes the outcome of the rankings. In statistics you would take aberations like these and exclude them because they give you trends that are unjustified. Basically because they fall out of the normal expectations+ reasonable variance. Kinda of like the lottery. If you choose 100 people at random and give them a lotto ticket there is a 99.9999% chance no one will come out a jackpot winner. But what if one person just happens to win. That changes the odds to 1 in 100. Obviously that is wrong. Your only solutions are to exclude the games when the score goes beyond normal expectation, or you need to increase your field of samples so that it will even out the variance. Thus you would buy 1,000 or 10,000 tickets instead. In football the only way to do that is play more games, or count games from the last few years. And there lies another problem, as teams wildly change because of the Cap and FA's. SO...

You need to come up with a criteria to judge the games, and then examine all results and exclude all variances. Ex.
Packers -Saints 52-3
Eagles- Seahawks- 0-42
ect....


Wow i havent tacken statistics since high school....

 

[Alex44]

Unluckily the formula is flawed. It shows green bay ahead of us.. Most would ask why a 2-10 team is ranked higher. Well the reason is on their first win they madly racked up the score against the saints. final score 52-3. This wildly changes the outcome of the rankings. In statistics you would take aberations like these and exclude them because they give you trends that are unjustified. Basically because they fall out of the normal expectations+ reasonable variance. Kinda of like the lottery. If you choose 100 people at random and give them a lotto ticket there is a 99.9999% chance no one will come out a jackpot winner. But what if one person just happens to win. That changes the odds to 1 in 100. Obviously that is wrong. Your only solutions are to exclude the games when the score goes beyond normal expectation, or you need to increase your field of samples so that it will even out the variance. Thus you would buy 1,000 or 10,000 tickets instead. In football the only way to do that is play more games, or count games from the last few years. And there lies another problem, as teams wildly change because of the Cap and FA's. SO...

You need to come up with a criteria to judge the games, and then examine all results and exclude all variances. Ex.
Packers -Saints 52-3
Eagles- Seahawks- 0-42
ect....


Wow i havent tacken statistics since high school....

One way to eliminate the big score games would be to discard both the biggest win and worst loss, then you come out with an average that isnt blown up by a huge game

 

[Disnardo]

Unluckily the formula is flawed. It shows green bay ahead of us.. Most would ask why a 2-10 team is ranked higher. Well the reason is on their first win they madly racked up the score against the saints. final score 52-3. This wildly changes the outcome of the rankings. In statistics you would take aberations like these and exclude them because they give you trends that are unjustified. Basically because they fall out of the normal expectations+ reasonable variance. Kinda of like the lottery. If you choose 100 people at random and give them a lotto ticket there is a 99.9999% chance no one will come out a jackpot winner. But what if one person just happens to win. That changes the odds to 1 in 100. Obviously that is wrong. Your only solutions are to exclude the games when the score goes beyond normal expectation, or you need to increase your field of samples so that it will even out the variance. Thus you would buy 1,000 or 10,000 tickets instead. In football the only way to do that is play more games, or count games from the last few years. And there lies another problem, as teams wildly change because of the Cap and FA's. SO...

You need to come up with a criteria to judge the games, and then examine all results and exclude all variances. Ex.
Packers -Saints 52-3
Eagles- Seahawks- 0-42
ect....


Wow i havent tacken statistics since high school....

Thanks for the info...

Of course we know that all formulas and stats are only good till an outlier is found... GB seems to be one, albeit, the problem with them is not just that lopsided win, because Indi, as well as SD, have had them. The main problem is that have lost most games by a vary little margin (6 games by a combined total of 17 points), and the Saints game (+49 points) makes a positive influence in the stat..

 

[BennyVW]

I think alex22 is on to something...Let me try the rankings when the we discard the best win, and worst loss.



Outlier is that freaking word i was trying to remember...

 

[TheAnswer1313]

Hmm this is the first time ive seen this used for football, it works alot better with baseball given the fact they play 162 games instead of just 16. With only 16 games u have a much smaller sample, thus those blowout games have a bigger effect on the formula.

 

[SQuinn17]

that system is crazy. how are the packers better than the patriots???

 

[Disnardo]

that system is crazy. how are the packers better than the patriots???

No system is perfect and the GB outlier proves it...

The reason it shows GB better than the Pats is not based on team wins... it's based on the differences in Points Scored and Points Allowed...

Green Bay has only a difference of -3 points between their D and O, compared to NE which has allowed Opponents to score 23 points (-23) more than their Offense...thus GB shows to be a better team, in this context, statistically speaking...

 

[Dolphin North]

One way to eliminate the big score games would be to discard both the biggest win and worst loss, then you come out with an average that isnt blown up by a huge game

Or do a standard deviation on the win marguns. STDEV command on EXCEL, or get a calculator for it on the net.

 

[Dol-Fan Dupree]

Unluckily the formula is flawed. It shows green bay ahead of us.. Most would ask why a 2-10 team is ranked higher. Well the reason is on their first win they madly racked up the score against the saints. final score 52-3. This wildly changes the outcome of the rankings. In statistics you would take aberations like these and exclude them because they give you trends that are unjustified. Basically because they fall out of the normal expectations+ reasonable variance. Kinda of like the lottery. If you choose 100 people at random and give them a lotto ticket there is a 99.9999% chance no one will come out a jackpot winner. But what if one person just happens to win. That changes the odds to 1 in 100. Obviously that is wrong. Your only solutions are to exclude the games when the score goes beyond normal expectation, or you need to increase your field of samples so that it will even out the variance. Thus you would buy 1,000 or 10,000 tickets instead. In football the only way to do that is play more games, or count games from the last few years. And there lies another problem, as teams wildly change because of the Cap and FA's. SO...

You need to come up with a criteria to judge the games, and then examine all results and exclude all variances. Ex.
Packers -Saints 52-3
Eagles- Seahawks- 0-42
ect....


Wow i havent tacken statistics since high school....

I was reading something about it on some article and they took away taht big game against the Saints. The Packers still had a good Pythagorean win percentage

 

[Awsi Dooger]

I always use the basic ratio of Points Scored Divided By Points Allowed. The league leader in that category has won the Super Bowl 7 or the past 9 years. The two misses were the '98 Vikings who blew the NFC Championship Game and the '01 Rams who lost the Super Bowl on the final play.

The Green Bay situation is very rare. But I don't discount those games. Excluding from an already short sample is unwise, IMO. Winning big in a balanced league like the NFL is very impressive. Many of my friends isolate underrated teams as ones with losing records but that win big on occasion. Those teams generally have more talent than their won-loss records indicate and you get good value the following year on the season over/under wins and also pointspreads on their games early in the season.