What's new
Fantasy Football - Footballguys Forums

Welcome to Our Forums. Once you've registered and logged in, you're primed to talk football, among other topics, with the sharpest and most experienced fantasy players on the internet.

Pythagorean Win Theorem for 2005/2006 season (1 Viewer)

KnowledgeReignsSupreme

KnowledgeReignsSupreme

Footballguy
Posted this a couple days late (since I wanted to get it in before the playoffs), but here's the results:1: Indianapolis 12.739983092: Seattle 12.331587783: Denver 11.726575624: Carolina 11.621500625: Pittsburgh 11.612002276: New York (N) 10.693130267: Jacksonville 10.680974848: San Diego 10.666755559: Chicago 10.3239868310: Kansas City 9.99620259811: Washington 9.88949145312: Cincinnati 9.72349648313: New England 9.07874656114: Tampa Bay 8.85609040615: Dallas 8.50862678416: Atlanta 8.27389057917: Miami 8.02984821818: Minnesota 6.8973505619: Baltimore 6.86335577520: Green Bay 6.6521307321: St. Louis 6.43668487722: Arizona 5.97248701123: Philadelphia 5.92113286524: Cleveland 5.60708277325: Oakland 5.45460537226: Buffalo 5.24290199627: Detroit 5.21811087828: Tennessee 4.92281277529: New York (A) 4.53391810430: Houston 3.70960878131: New Orleans 3.56682829832: San Francisco 3.213809086Of the top 12 teams, 7 AFC/5 NFC. Last year it was much more lopsided for the AFC (6 of the top 7 were AFC).Biggest overachievers:Tampa Bay +2.14 games (8.85 expected wins, 11 actual wins)Minnesota +2.10 games (6.89 expected wins, 9 actual wins)Biggest underachievers:Green Bay -2.65 games (6.65 expected wins, 4 actual wins)PWT says teams that over/underachieve should return to their expected range the following year.Let's check how PWT did for 2005:Biggest overachievers:#1) Pittsburgh 11.48 expected wins, 15 actual 3.518 difference#2) Atlanta 8.083 expected wins, 11 actual 2.916 difference#3) St. Louis 6.084 expected wins, 8 actual 1.915 differencePWT was money as all three teams hit their expected win total in year N+1And just for jwvdcw who swore last year that Atlanta didn't overachieve last year, but was just good at winning close games... well here's some proof that they just got lucky (I mentioned the reasons why in another thread which boiled down to INT returns for TDs and an abnormal fumble recovery rate).Biggest underachievers:#1) Tampa Bay 9.233 expected wins, 7 actual -2.233 difference#2) Kansas City 8.987 expected wins, 7 actual -1.987 difference#3) Buffalo 10.977 expected wins, 9 actual -1.977 differenceMoney on KC, pretty good on Tampa (who underperformed for two consecutive years and have now overperformed this year) and completely missed the boat on Buffalo.

 
Just wanted to add that Indy's PWT score would rank them #6 best SB winning team since 1985 (assuming they win).1985 CHI 14.0541999 STL 13.8071996 GNB 13.8031991 WAS 13.7902000 BAL 13.4532002 TAM 12.6981989 SFO 12.6331994 SFO 12.4811992 DAL 12.3922004 NWE 12.3821997 DEN 12.2361993 DAL 12.1701998 DEN 12.1381990 NYG 11.9911986 NYG 11.9201995 DAL 11.5472003 NWE 11.3772001 NWE 10.8171988 SFO 10.1031987 WAS 9.9412005 IND 12.7382005 SEA 12.331

 
Posted this a couple days late (since I wanted to get it in before the playoffs), but here's the results:

1: Indianapolis 12.73998309

2: Seattle 12.33158778

3: Denver 11.72657562

4: Carolina 11.62150062

5: Pittsburgh 11.61200227

6: New York (N) 10.69313026

7: Jacksonville 10.68097484

8: San Diego 10.66675555

9: Chicago 10.32398683

10: Kansas City 9.996202598

11: Washington 9.889491453

12: Cincinnati 9.723496483

13: New England 9.078746561

14: Tampa Bay 8.856090406

15: Dallas 8.508626784

16: Atlanta 8.273890579

17: Miami 8.029848218

18: Minnesota 6.89735056

19: Baltimore 6.863355775

20: Green Bay 6.65213073

21: St. Louis 6.436684877

22: Arizona 5.972487011

23: Philadelphia 5.921132865

24: Cleveland 5.607082773

25: Oakland 5.454605372

26: Buffalo 5.242901996

27: Detroit 5.218110878

28: Tennessee 4.922812775

29: New York (A) 4.533918104

30: Houston 3.709608781

31: New Orleans 3.566828298

32: San Francisco 3.213809086

Of the top 12 teams, 7 AFC/5 NFC.  Last year it was much more lopsided for the AFC (6 of the top 7 were AFC).

Biggest overachievers:

Tampa Bay +2.14 games (8.85 expected wins, 11 actual wins)

Minnesota +2.10 games (6.89 expected wins, 9 actual wins)

Biggest underachievers:

Green Bay -2.65 games (6.65 expected wins, 4 actual wins)

PWT says teams that over/underachieve should return to their expected range the following year.

Let's check how PWT did for 2005:

Biggest overachievers:

#1) Pittsburgh 11.48 expected wins, 15 actual 3.518 difference

#2) Atlanta 8.083 expected wins, 11 actual 2.916 difference

#3) St. Louis 6.084 expected wins, 8 actual 1.915 difference

PWT was money as all three teams hit their expected win total in year N+1

And just for jwvdcw who swore last year that Atlanta didn't overachieve last year, but was just good at winning close games... well here's some proof that they just got lucky (I mentioned the reasons why in another thread which boiled down to INT returns for TDs and an abnormal fumble recovery rate).

Biggest underachievers:

#1) Tampa Bay 9.233 expected wins, 7 actual -2.233 difference

#2) Kansas City 8.987 expected wins, 7 actual -1.987 difference

#3) Buffalo 10.977 expected wins, 9 actual -1.977 difference

Money on KC, pretty good on Tampa (who underperformed for two consecutive years and have now overperformed this year) and completely missed the boat on Buffalo.
Click to expand...
good stuff KRS...i know this has been used a lot in baseball & basketball, but haven't seen it transposed to football domain... intuitively it seems like it would be useful in most sports... some teams might win or lose more... than they "should" have... based on flukiness... & you can make inferences about this based on scoring average... margin of victory & loss... etc...

in case of BUF... what seems to be a sharp tool could become a blunt instrument to the degree there have been some drastic, material changes in composition of the roster...

on offense... as erratic as bledsoe was in 2003-2004 seasons... in retrospect, losman was a big dropoff... maybe they thought they had gone as far as they could with bledsoe, & if losman developes, the experience he got this season might later be viewed as worth a down season... nonetheless... it would tend to throw above won-loss "prediction" (2004 data pointing towards 2005) out of whack if not accounted for...

on defense... losing DT pat williams (free agency) & WLB takeo spikes (IR) were huge losses.

this is a case where as good as the numbers are... it is as much (if not more) art than science... in figuring out which teams will be most effected by differences in roster composition from season to season... in the free agent era, nearly all teams will be impacted to a degree... plus injury-related attrition can take a heavy toll.

 
Last edited by a moderator:
Care to share the formula?ETA: Never mind.I thought this might be what the Theorem was.Win % = (points scored)^X / (points scored^X + points against ^X).What I don't know is why you are using an exponent ("X") of 2.37......

 
Last edited by a moderator:
DVOA rankings from footballoutsiders.com:

# TEAM DVOA

1 IND 36.60%

2 DEN 36.50%

3 KC 31.10%

4 SEA 28.20%

5 PIT 26.10%

6 CIN 25.10%

7 NYG 24.60%

8 SD 23.90%

9 WAS 22.60%

10 JAC 21.70%

11 CAR 13.00%

12 CHI 7.00%

13 TB 4.60%

14 MIA 4.50%

15 NE 4.40%

16 DAL 1.70%

17 MIN -0.10%

18 PHI -3.00%

19 OAK -4.30%

20 BAL -6.10%

21 ATL -6.90%

22 ARI -14.40%

23 CLE -15.40%

24 TEN -18.30%

25 DET -19.10%

26 BUF -19.20%

27 NYJ -21.60%

28 GB -22.30%

29 STL -23.00%

30 NO -28.20%

31 HOU -33.30%

32 SF -65.90%

 
Last edited by a moderator:
Care to share the formula?

ETA: Never mind.

I thought this might be what the Theorem was.

Win % = (points scored)^X / (points scored^X + points against ^X).

What I don't know is why you are using an exponent ("X") of 2.37......
Click to expand...
2.37 is the football exponent. The NBA and MLB use different exponents.
 
Care to share the formula?

ETA: Never mind.

I thought this might be what the Theorem was.

Win % = (points scored)^X / (points scored^X + points against ^X).

What I don't know is why you are using an exponent ("X") of 2.37......
Click to expand...
2.37 is the football exponent. The NBA and MLB use different exponents.
Click to expand...
Is it too lengthy of a response, or can you give a reason for that number?Best fit to the model?

 
Pythagoras: Anaximander! I have toiled long and hard, and have arrived at an incredible mathematical truth which states that in any right triangle, the area of the square whose side is the hypotenuse is equal to the sum of the areas of the squares of the other two sides. I shall call it the Pythagorean Theorem, and its applications throughout the history of man shall be countless!Anaximander: Excellent work, Pythagoras! Will it even be able to estimate wins of NFL teams 2500 years from now?Pythagoras: Yes! Well, except in Tampa Bay. For some reason I just can't quite seem to get that one...Anaximander: Forget it then. The theorem is fla\/\/ed. Throw it out. :thumbdown:

 
Last edited by a moderator:
1: Indianapolis 12.73998309

2: Seattle 12.33158778

3: Denver 11.72657562

4: Carolina 11.62150062

5: Pittsburgh 11.61200227

6: New York (N) 10.69313026

7: Jacksonville 10.68097484

8: San Diego 10.66675555

9: Chicago 10.32398683

10: Kansas City 9.996202598

11: Washington 9.889491453

12: Cincinnati 9.723496483

13: New England 9.078746561

14: Tampa Bay 8.856090406

15: Dallas 8.508626784

16: Atlanta 8.273890579

17: Miami 8.029848218

18: Minnesota 6.89735056

19: Baltimore 6.863355775

20: Green Bay 6.65213073

21: St. Louis 6.436684877

22: Arizona 5.972487011

23: Philadelphia 5.921132865

24: Cleveland 5.607082773

25: Oakland 5.454605372

26: Buffalo 5.242901996

27: Detroit 5.218110878

28: Tennessee 4.922812775

29: New York (A) 4.533918104

30: Houston 3.709608781

31: New Orleans 3.566828298

32: San Francisco 3.213809086
Click to expand...
Teams 2-5 have made the big conference game. Score another one for PWT, although a minus for not predicting chokes. :kicksrock:
 
And this is probably as good a place as any to post the final 2005 rankings based on the touchdown-firstdown per play ratio.

Code: 
OFFENSE   DEFENSE  TEAM         RANK   RANK   TOTALSeattle         2     9      11Indianapolis    1    15      16Pittsburgh      9     7      16Washington     11     5      16Carolina       17     2      19Jacksonville   14     6      20Denver          6    17      23New York (N)   12    11      23Dallas         18     8      26Tampa Bay      23     3      26Oakland        13    14      27Green Bay      15    13      28San Diego       4    24      28Kansas City     5    25      30New England     7    23      30Baltimore      27     4      31Chicago        31     1      32Cincinnati      3    29      32Miami          24    10      34Philadelphia   22    12      34Atlanta        10    26      36Minnesota      16    20      36Arizona        21    16      37New Orleans    19    18      37St. Louis       8    30      38Tennessee      20    27      47Cleveland      29    19      48New York (A)   28    22      50Detroit        30    21      51Buffalo        25    31      56Houston        26    32      58San Francisco  32    28      60
PWT says Seattle is the best of what's left, the ratio says Seattle was the best all along anyway.
 
Last edited by a moderator:
PWT says teams that over/underachieve should return to their expected range the following year.
Click to expand...
No, it doesn't.
Click to expand...
Okay, it doesn't then according to you.
Click to expand...
Dude, It Doesn'tI actually appreciate the time you've spent on this thread. Interesting stuff. My only quibble is with attributing things to James and his work which he never actually said. He discusses these issues in the 1981 Abstract. More recently, the Red Sox had him work on a question directly related to this point.

 
1: Indianapolis 12.73998309

2: Seattle 12.33158778

3: Denver 11.72657562

4: Carolina 11.62150062

5: Pittsburgh 11.61200227
Click to expand...
Teams 2-5 have made the big conference game. Score another one for PWT, although a minus for not predicting chokes. :kicksrock:
Click to expand...
One Rothlisberger tackle away from 1-4. :mellow:
 
Ok, so this is a nice way of looking back in time..... but what use is this to predict anything?Asides from (possible) playoff predictions, what value is this exercise?

 
Ok, so this is a nice way of looking back in time..... but what use is this to predict anything?

Asides from (possible) playoff predictions, what value is this exercise?
Click to expand...
Best PWT is a good indicator of the SB winner. It's not perfect, but it's pretty good.
 
Ok, so this is a nice way of looking back in time..... but what use is this to predict anything?

Asides from (possible) playoff predictions, what value is this exercise?
Click to expand...
Best PWT is a good indicator of the SB winner. It's not perfect, but it's pretty good.
Click to expand...
Fair enough. Just hoping that this wouldn't be suggested as a next year indicator.
 
So it's saying Seattle will beat Denver in the Super Bowl,unless one of them pulls a "Colt"? :eek:
Click to expand...
It's saying all of them are about as good as their record. Seattle and Indy have historically very good teams.Denver, Pitt and Carolina are not historically "very good", but are more than good enough to win it all.

I just noticed something else.

These are the playoff teams with final wins +1 game over their expected wins (which isn't generally a cause for concern).

Jacksonville +1.31

Cincinnati +1.27

Denver +1.27

Indianapolis +1.26

3/4 lost their first game.

 
Last edited by a moderator:
You must log in or register to reply here.

Similar threads

QuizGuy66
*** 2024 QuizGuy66 Somewhat Predictable Season Prediction Thread - feel free to add your own
Replies
0
Views
40
QuizGuy66
QuizGuy66
Ministry of Pain
Projected Win totals for all 32 teams
Replies
24
Views
522
Alex P Keaton
A
T
Travdogg's Predictions for the 2024 NFL season
Replies
25
Views
2K
Moonlight
Moonlight

Users who are viewing this thread

Total: 2 (members: 0, guests: 2)
Top