NFL 2015: Which Team is the "Best"?: Final

Introduction

In two previous posts (here and here), I claimed twice that the Patriots were the "best" team in the sense that, according to my model, they would have the most wins in a tournament in which every team in the NFL played every other team in the NFL.  Now that the season is over, the Broncos have won the Superbowl, and there are no more games to be played, does this assertion still make sense?

Method 1:

Using the most recent data, I re-ran the simulation, pitting every team against every other team once and assigning a win to the team most likely to win in the matchup.  There are 32 games total for each team (and playing yourself is a win).

Which team is the "best"? The tops 5 are:
     1. Bengals
     2. Panthers
     3. Patriots
     4. Seahawks
     5. Steelers

Which team is the "worst"?  The worst 5 are:        
     32. Redskins
     31. Rams
     30. Giants
     29. Jaguars
     28. Titans

Bengals move to the top, while the Patriots slip to third.  The Broncos, despite winning the Superbowl, are at 9.  Here is the full list:

Rank	Team	Wins	Losses
1	Bengals	32	0
2	Panthers	31	1
3	Patriots	30	2
4	Seahawks	29	3
5	Steelers	28	4
6	Cardinals	27	5
7	Packers	26	6
8	Falcons	25	7
9	Broncos	24	8
10	Vikings	23	9
11	Bills	22	10
12	Ravens	21	11
13	Chiefs	20	12
14	Lions	19	13
15	Colts	18	14
16	Texans	17	15
17	Jets	16	16
18	Dolphins	15	17
19	Cowboys	14	18
20	Eagles	13	19
21	Raiders	12	20
22	Saints	11	21
23	Chargers	10	22
24	Browns	9	23
25	49ers	8	24
26	Bears	7	25
27	Buccaneers	6	26
28	Titans	5	27
29	Jaguars	4	28
30	Giants	3	29
31	Rams	2	30
32	Redskins	1	31

Method 2: 

In the foregoing method, all "wins" are equal.  For example, a predicted 80% win counts the same as a 50.1% win.  This allows for the possibility that a team may "lose" every game by a narrow margin, making it look like the worst team, while another team may barely "win" a few and "lose" all the rest by a great margin.  The second team, in the long run, may actually win fewer games than the former team  if we allow chance to have a role.  And as we know from watching the post season, just because a team is predicted to win doesn't mean it actually will win.  Each team has a chance to win in each game.

To account for this, instead of giving the full win value to the team which has a higher probability of winning (even if barely), this second approach splits the "win" according to the probabilities assigned.  To use the examples above, if a team has an 80% likelihood of winning, it gets 0.8 of the "win" and the opposing team gets 0.2 of the "win". 

These "wins" are once again summed up, where the highest number of "wins" possible is 31.5 (100% likelihood of winning each game except against yourself, which is 50%) and the lowest number of "wins" possible is 0.5 (0% likelihood of winning each game except against yourself, which is 50%).

Here are the results:

Which team is the "best"? The tops 5 are:
     1. Patriots
     2. Bengals
     3. Panthers
     4. Seahawks
     5. Steelers

Which team is the "worst"?  The worst 5 are:        
     32. Redskins
     31. Rams
     30. Giants
     29. Jaguars
     28. Titans

Patriots are back on top, while the Bengals are in second.  The Panthers move back to third.  The Broncos are still at 9.  The Texans and Jets switch spots while the 49ers move up to 23 from 25.  Everything else stays the same.

Here is the full list:

Rank	Team	Wins
1	Patriots	21.476
2	Bengals	21.442
3	Panthers	21.03
4	Seahawks	20.262
5	Steelers	20.164
6	Cardinals	19.271
7	Packers	18.963
8	Falcons	17.984
9	Broncos	17.841
10	Vikings	17.757
11	Bills	17.257
12	Ravens	17.183
13	Chiefs	16.755
14	Lions	16.656
15	Colts	16.591
16	Jets	16.391
17	Texans	16.343
18	Dolphins	16.093
19	Cowboys	15.914
20	Eagles	15.033
21	Raiders	14.872
22	Saints	14.091
23	49ers	13.968
24	Chargers	13.965
25	Browns	13.703
26	Bears	13.461
27	Buccaneers	13.163
28	Titans	13.079
29	Jaguars	11.803
30	Giants	11.164
31	Rams	9.762
32	Redskins	8.563

Conclusion

Upon reflection, I like the second method better.  While the actual ranking does not change much, it does a better job showing how much better or worse one team is compared to another.  For example, the Patriots and Bengals are about dead even.  It is also more realistic.  The Patriots or Bengals would not win all 32 games, whereas 21 wins each is much more reasonable.  At the other end, the Rams and Redskins get many more victories than they would otherwise. Thus, this approach shows that the NFL is much more competitive than the previous approach would suggest.

However, in the end, the answer to our question remains the same.  My Superbowl predictions before and during the early playoffs claimed that the Patriots or the Bengals would meet the Panthers in the Superbowl.  While the Broncos beat out the Patriots and Bengals for the AFC spot and beat the Panthers in the Superbowl, we should simply say that the Broncos had the best playoff performance and were the best team in each game they played.  Don't get me wrong.  I am very pleased that the Broncos won the Superbowl.  But that does not make them the "best" team overall, at least according to my use of the word "best".

Thus, at the end of the season, in my judgment, the Patriots, Bengals, and Panthers are the best three teams of the 2015 NFL season.