Power Ratings, Week 13

  1. You have chosen to ignore posts from Paul_K. Show Paul_K's posts

    Re: Power Ratings, Week 13

    In Response to Re: Power Ratings, Week 13:
    Why do people even argue about power rankings? They're a statistic created by sports media to give themselves something to write and talk about. No different from the Thursday injury report which was created by the NFL to give sports books help in setting lines. All of it is meaningless.
    Posted by Evil2009

    I want to know if the Patriots have any more of a chance (or less) than their on-field play shows, and what their specific strengths and problems might be. 

    ----

    My proposed power rating revision:

    I break a football game into the team’s basic game, a clock-watching asymmetrical game late in the fourth quarter, and a few empty points and meaningless gestures at the end of some games.  I’m moving to two power rating statistics:  a more inclusive power index plus a game winning drive index.

    My standard power index will now cover much of the fourth quarter, as long as the game isn’t getting out of hand. 

    My revised dividing line between the basic and time-dependent game won’t be the end of the third quarter.  It will be based on the number of 8 point scores that the losing team needs to tie the game.  There isn’t a firm dividing line but I need one, so here’s a reasonable line:  A team that needs three scores will be said to abandon its basic game 15 minutes before the end of the game.  A team that needs two scores will be said to abandon its basic game 10 minutes before the end of the game.  A team that needs one score will be said to abandon its basic game 5 minutes before the end of the game.  In a tie game, one team or the other will be said to start hoarding the seconds 2 minutes before the end of the game.  By rule, once the standard game is officially declared over by hitting a time limit or by one team getting further ahead, no score changes can put the standard game back into play again.  There will be only one standard game differential.

    My rule will be, the first time that this line is crossed, that’s the stopping point for the standard power index. 

    In addition to the standard power index, I want to record a game-winning drive index.  May John Elway smile down on me from the broadcast booth or from wherever he hangs out.  I specifically want to differentiate the game-winning drives, and the teams which allow game winning drives, from several types of meaningless and empty points that teams sometimes create at the end of the game.

    Piling-on points beyond the end of the standard game don’t count.  For the Patriots-Tennessee game, this might go all the way back to halftime.  Neither will single touchdowns count when the losing team needed two scores to win, but didn’t bother to play for two scores.  I remember a New England - Pittsburgh game around 2003 where the Steelers called a timeout with a handful of seconds left because they wanted their touchdown, except the Steelers were at least three scores back.  Only real wins and tie games at the end of regulation count for this second index.

    If your team is three scores back and it comes back to win, it gets credit for three game-winning drives.  At the same time, the other team gets a -3 in this rating category for letting your team run wild.  If your team earns a tie game, the game-tying drive only counts for half of one game winning drive.

    Coin flips in overtime are statistical noise and kind of bad football, because whoever wins the coin flip tends to win the game.  A coin flip shouldn’t count in a power index unless the zebras are skillfully bought off to flip two-headed coins.  Personally I’d have the kickoff receiving team always start from the 10 yard line, not from the 20, if they choose not to bring the ball out.  A team that elects to receive and succeeds at a (short) game winning drive earns 0.3 game winning drives if they win.  The other team can earn 0.7 game winning drives if they eventually win.

    Right off, I expect Indianapolis to be the king of game-winning drives, and I expect Miami to be the goat.

    At some point I need to combine these two indices to get a standard power index.  First, let’s see how each index looks.  I suspect that the base index will be pretty tried-and-true, while the winning drive index will operate on scanty data.  We’ll see.

     
  2. You have chosen to ignore posts from zbellino. Show zbellino's posts

    Re: Power Ratings, Week 13

    http://www.usatoday.com/sports/sagarin/nfl09.htm

    They look very similar to the Pure Points scale on the Sagarin index NFL rating.

    That has the Pats as the fourth best team in the NFL.

    I think the reality is closest to Sagarin's.

    Yours is absolutely loony. NE is not the best team. I don't care about noise in the 4th. The 4th is still a quarter, and your reason for removing is not valid, and obviously skews the results.

    It simply is NOT an accurate depiction of who has the most strength, and I wouldn't bet on that scale if my life depended  on it.

    OTOH, our friend here, showing NE that low is just not serious for me.

    NE is in an odd place right now. They are a cut above teams like Baltimore, Pitt, Dallas, etc. But they appear to be a notch below the teams like Minny, NO, and Indy.


     
  3. You have chosen to ignore posts from BBcheats. Show BBcheats's posts

    Re: Power Ratings, Week 13

    In Response to Power Ratings, Week 13:
    OK, I’ll take criticism.   New Orleans spanked the Pats.   OK, that’s one trap game between two division rival games for the Pats.   Maybe New Orleans really wanted the pride and fame. New Orleans was ahead at the end of the third quarter like a good team should be, and so the score brings New England down a notch and New Orleans up.   However, after we add the other 100 NFL games New England is still on top, outweighing this one game.   So let’s look at the numbers: I believe that my “three quarters” statistical power rating system has a fundamental soundness to it across the NFL, a soundness that a simple win-loss record doesn’t reflect.   It’s sensitive enough that we can look at all the outliers and ask what happened in each case. The big positive outliers are Indianapolis and New Orleans.   They both have pass-happy attacks, they both often get behind like bad teams do and they both consistently pull wins out of dangerous situations in the fourth quarter. One big negative outlier is Miami.   Miami has a rookie QB and they run lots and lots of gimmick plays.   My guess is that their opponents figure out the gimmick plays, or Miami can’t either remember or execute all those intricate gimmicks when they’re gassed in the fourth quarter. Another negative outlier is New England.   Best guess, they start talented rookies, and the rookies are always gassed in the fourth quarter.   How’s that? Also they just played two high-stakes gambles against two highly offense-minded teams, and both gambles failed big.   That’s exactly why I designed my rating system, to screen out the big gambles.   Finally, when they’re grossly ahead they play the most consistent prevent in the league.   Or they used to.   Brady is something ridiculous like 75-4 when he’s ahead at the start of the fourth quarter.   Also, the Patriots have had an awesome offense and a suspect defense for years.   BB plays ball control like crazy.   He’d rather have a 20 play drive for a field goal than a 1 play drive for a touchdown.   He’d rather have Sammy Morris dive into the pile three times than throw an interception in the fourth quarter.   He’d rather go for first down on a 4th and 1 late in the game than try a long, meaningless field goal.   He’d rather hike the ball clean out of the endzone than punt.   My ratings system counts ball control. Any other suggestions on why the Pats collapse in the fourth quarter?   1           ne             14.1             (last wk. 1) 2           no             12.7             (last wk. 5) 3           min             12.3             (last wk. 4) 4           mia             12.3             (last wk. 3) 5           ind             11.6             (last wk. 2) 6           sd             11.3             (last wk. 7) 7           gb             11.1             (last wk. 10) 8           cin             10.6             (last wk. 8) 9           pit             10.3             (last wk. 9) 10         az             10.3             (last wk. 6) 11         nyj             10.2             (last wk. 11) 12         bal             9.7             (last wk. 13) 13         hou             9.7             (last wk. 17) 14         dal             8.6             (last wk. 14) 15         atl             8.4             (last wk. 12) 16         den             8.1             (last wk. 20) 17         phi             8.0             (last wk. 16) 18         nyg             8.0             (last wk. 15) 19         buf             7.6             (last wk. 19) 20         car             7.6             (last wk. 18) 21         sf             7.0             (last wk. 21) 22         was             6.4             (last wk. 22) 23         ten             5.8             (last wk. 25) 24         sea             5.0             (last wk. 26) 25         chi             4.9             (last wk. 24) 26         jac             4.8             (last wk. 23) 27         tb             4.4             (last wk. 30) 28         cle             3.4             (last wk. 31) 29         kc             3.2             (last wk. 27) 30         stl             3.1             (last wk. 28) 31         det             3.1             (last wk. 29) 32         oak             2.3             (last wk. 32)
    Posted by Paul_K



    Biggest crock of crap ever......
     
  4. You have chosen to ignore posts from BBcheats. Show BBcheats's posts

    Re: Power Ratings, Week 13

    In Response to Re: Power Ratings, Week 13:
    In Response to Power Ratings, Week 13 : Paul, You are giving all of us in NE a bad reputation by posting this nonsense. Please stop now.
    Posted by nyjoseph


    CORRECT!!!
     
  5. You have chosen to ignore posts from Paul_K. Show Paul_K's posts

    Re: Power Ratings, Week 13

    Revised basic power rankings, out of 16 victories if the season re-started tomorrow.  These rankings now include all fourth quarter scores when the game is not so out of reach that seconds are precious to one team, as defined above.  They don’t yet include a factor for game-winning drives and game tying drives.

     

    team  week 4  week 5  week 6  week 7  week 8  week 9  week 10  week 11  week 12

    buf   6.2   5.8   7.4   7.6   7.2   7.6   6.4   6.9   7.6

    mia   8.1   8.9   9.0   9.6   10.4   10.7   9.2   10.2   9.4

    ne   8.1   10.2   13.2   13.3   13.9   14.0   14.3   14.8   14.0

    nyj   10.5   10.6   10.0   11.6   11.2   11.6   9.7   10.0   9.7

    cin   7.3   8.0   7.1   10.1   10.2   10.2   11.0   10.0   9.7

    pit   9.8   10.6   9.6   10.3   10.3   11.3   11.0   9.6   10.2

    bal   8.3   9.4   9.3   9.9   11.6   10.3   10.6   10.5   10.5

    cle   4.2   5.2   5.6   5.1   4.1   3.2   2.9   3.3   4.0

    ind   14.8   14.2   13.4   13.5   12.0   12.1   11.9   11.9   12.6

    jac   11.1   6.9   5.1   5.3   2.6   4.0   5.6   5.3   4.3

    hou   8.0   6.5   7.6   9.1   9.5   10.4   10.3   9.4   8.6

    ten   7.7   6.2   2.8   3.0   3.5   6.0   6.8   7.2   6.9

    den   10.3   10.4   11.8   12.7   11.7   10.2   8.7   7.0   8.4

    kc   3.0   3.6   3.5   2.0   2.5   1.9   2.4   4.0   3.1

    oak   3.0   1.9   3.4   2.2   2.6   2.3   1.0   2.1   1.9

    sd   6.4   6.7   6.6   8.3   7.8   7.4   7.6   9.7   10.3

    dal   8.5   8.8   8.9   10.5   12.1   12.5   10.8   9.7   9.7

    nyg   10.1   11.4   11.0   10.3   8.9   9.3   9.1   9.4   8.1

    phi   9.9   10.9   8.9   8.7   11.4   11.5   10.6   10.8   9.9

    was   3.9   3.7   2.6   1.5   2.4   3.0   4.4   5.2   5.9

    chi   9.8   10.2   10.0   6.5   6.9   4.7   5.3   5.0   5.0

    det   6.6   5.9   4.2   3.6   2.7   2.4   4.2   3.4   3.2

    gb   6.6   6.3   8.3   9.7   8.9   7.5   9.7   10.2   11.2

    min   10.9   10.8   12.4   12.2   12.9   12.3   12.1   12.9   13.5

    atl   6.8   12.2   12.8   9.9   11.1   10.2   9.2   9.5   8.2

    car   3.3   4.6   3.6   3.0   6.4   7.8   8.1   7.6   7.2

    no   13.1   13.5   14.6   14.6   14.8   14.3   13.0   14.4   15.1

    tb   3.0   2.9   2.1   2.0   2.9   4.0   4.8   3.9   4.9

    az   12.0   8.8   11.1   11.9   8.2   10.4   10.5   9.6   9.6

    stl   2.8   1.0   1.8   1.6   1.0   0.5   2.2   2.4   1.5

    sf   12.7   9.4   10.1   8.6   8.9   7.0   6.9   6.0   6.8

    sea   9.2   10.4   8.3   7.7   5.4   5.1   5.7   4.0   5.1

     

    Week 12 rankings in order:

    1   no   15.1   (last wk. 2)

    2   ne   14.0   (last wk. 1)

    3   min   13.5   (last wk. 3)

    4   ind   12.6   (last wk. 4)

    5   gb   11.2   (last wk. 8)

    6   bal   10.5   (last wk. 6)

    7   sd   10.3   (last wk. 12)

    8   pit   10.2   (last wk. 13)

    9   phi   9.9   (last wk. 5)

    10   nyj   9.7   (last wk. 9)

    11   cin   9.7   (last wk. 10)

    12   dal   9.7   (last wk. 11)

    13   az   9.6   (last wk. 14)

    14   mia   9.4   (last wk. 7)

    15   hou   8.6   (last wk. 16)

    16   den   8.4   (last wk. 20)

    17   atl   8.2   (last wk. 15)

    18   nyg   8.1   (last wk. 17)

    19   buf   7.6   (last wk. 21)

    20   car   7.2   (last wk. 18)

    21   ten   6.9   (last wk. 19)

    22   sf   6.8   (last wk. 22)

    23   was   5.9   (last wk. 24)

    24   sea   5.1   (last wk. 26)

    25   chi   5.0   (last wk. 25)

    26   tb   4.9   (last wk. 28)

    27   jac   4.3   (last wk. 23)

    28   cle   4.0   (last wk. 30)

    29   det   3.2   (last wk. 29)

    30   kc   3.1   (last wk. 27)

    31   oak   1.9   (last wk. 32)

    32   stl   1.5   (last wk. 31)

     
  6. You have chosen to ignore posts from Paul_K. Show Paul_K's posts

    Re: Power Ratings, Week 13

    In Response to Re: Power Ratings, Week 13:
    http://www.usatoday.com/sports/sagarin/nfl09.htm They look very similar to the Pure Points scale on the Sagarin index NFL rating. That has the Pats as the fourth best team in the NFL. I think the reality is closest to Sagarin's. Yours is absolutely loony. NE is not the best team. I don't care about noise in the 4th. The 4th is still a quarter, and your reason for removing is not valid, and obviously skews the results.
    Posted by zbellino


    I put much of the 4th quarter back in the revised system, although I still care about removing noise late in the 4th.  In that respect it's not a Sagarin pure points system. 

    It also uses a more subtle weighting system than Sagarin's in two respects.  First, it reduces the weight of the September games.  Second, instead of Sagarin's system of looking at the opponents' win-loss record, my system seeks to miniimize the total deviations NFL-wide in order to come up with a best-fit set of power ratings.  Because teams from the same division play each other twice and one team has to win and one team has to lose, and the total win-loss record of a division is diluted.  The power rating of a strong division is then diluted with Sagarin's pure points system.  My system is not afraid to give due credit or blame to weak and strong divisions.

    With the revision, N.O. benefited considerably.  In this power rating, NE is second this week behind New Orleans. 

    Apologies to Indianapolis, as I haven't factored in the game winning drive factor yet.  Good night for now.
     
  7. You have chosen to ignore posts from carawaydj. Show carawaydj's posts

    Re: Power Ratings, Week 13

    Paul, since I'm just a typical nughead, I take a very nughead view of things.  In my view, the only substantive value power rankings can have is if they do a good job predicting if team A can beat team B.  Have you kept track of this so far?

    That just seems to be the ultimate test of a ranking system.  Nothing will be 100% accurate, but the bottom line assessment is simply this, does it predict wins against the spread more often than not?  If it does, I could care less if you factor in the color hair of the cheerleaders.  Other than that, the power rankings are just a way to satisfy us fans requirements to argue who is best before the teams play it out in the playoffs and prove it.  I'm all for people thinking out of the box and taking the time to innovate.  Where does it stand against the ultimate test?
     
  8. You have chosen to ignore posts from BBcheats. Show BBcheats's posts

    Re: Power Ratings, Week 13

    In Response to Re: Power Ratings, Week 13:
    In Response to Re: Power Ratings, Week 13 : I put much of the 4th quarter back in the revised system, although I still care about removing noise late in the 4th.  In that respect it's not a Sagarin pure points system.  It also uses a more subtle weighting system than Sagarin's in two respects.  First, it reduces the weight of the September games.  Second, instead of Sagarin's system of looking at the opponents' win-loss record, my system seeks to miniimize the total deviations NFL-wide in order to come up with a best-fit set of power ratings.  Because teams from the same division play each other twice and one team has to win and one team has to lose, and the total win-loss record of a division is diluted.  The power rating of a strong division is then diluted with Sagarin's pure points system.  My system is not afraid to give due credit or blame to weak and strong divisions. With the revision, N.O. benefited considerably.  In this power rating, NE is second this week behind New Orleans.  Apologies to Indianapolis, as I haven't factored in the game winning drive factor yet.  Good night for now.
    Posted by Paul_K



    No apologies necessary....its an idiotic analysis....garbage in = garbage out
     
  9. You have chosen to ignore posts from Paul_K. Show Paul_K's posts

    Re: Power Ratings, Week 13

    In Response to Re: Power Ratings, Week 13:
    Paul, since I'm just a typical nughead, I take a very nughead view of things.  In my view, the only substantive value power rankings can have is if they do a good job predicting if team A can beat team B.  Have you kept track of this so far? That just seems to be the ultimate test of a ranking system.  Nothing will be 100% accurate, but the bottom line assessment is simply this, does it predict wins against the spread more often than not?  If it does, I could care less if you factor in the color hair of the cheerleaders.  Other than that, the power rankings are just a way to satisfy us fans requirements to argue who is best before the teams play it out in the playoffs and prove it.  I'm all for people thinking out of the box and taking the time to innovate.  Where does it stand against the ultimate test?
    Posted by carawaydj

    I just reworked the stats at 1:30 a.m.  Analysis could take time.  Real analysis takes place in real time. 

    All statistics are blind to what comes over the news wire.  A prognosticator would want to combine stats with injury reports, with weather reports for game time (they issue slightly accurate 10 day forecasts these days), and with known trends such as favoring home underdogs slightly, and favoring huge pointspreads because everybody puts money on the big winner.

    I have some game winning drive data to factor in.  I haven't decided how much weight it gets and how stale the data gets with age. 

    Just a hunch, I expect my system to be measurably superior to Jeff Sagaran's pure point system, especially because Sagarin's system has a long tail.  Data from the Patriots' first two games, where the Pats stunk, is no longer as relevant as the data from the last 4 or 8 games.  Sagarin's chess system is just buttering up the winners, and I don't think that helps anyone.

    P.S.  Gambling is for people who don't understand mathematics.  Or the nature of bookmaking operations.
     
  10. You have chosen to ignore posts from BBcheats. Show BBcheats's posts

    Re: Power Ratings, Week 13

    In Response to Re: Power Ratings, Week 13:
    In Response to Re: Power Ratings, Week 13 : I just reworked the stats at 1:30 a.m.  Analysis could take time.  Real analysis takes place in real time.  All statistics are blind to what comes over the news wire.  A prognosticator would want to combine stats with injury reports, with weather reports for game time (they issue slightly accurate 10 day forecasts these days), and with known trends such as favoring home underdogs slightly, and favoring huge pointspreads because everybody puts money on the big winner. I have some game winning drive data to factor in.  I haven't decided how much weight it gets and how stale the data gets with age.  Just a hunch, I expect my system to be measurably superior to Jeff Sagaran's pure point system, especially because Sagarin's system has a long tail.  Data from the Patriots' first two games, where the Pats stunk, is no longer as relevant as the data from the last 4 or 8 games.  Sagarin's chess system is just buttering up the winners, and I don't think that helps anyone. P.S.  Gambling is for people who don't understand mathematics.  Or the nature of bookmaking operations.
    Posted by Paul_K


    You just keep on massaging your fantasy figures until the patsies come out on top.  I suggest adding something on # of times quitting early.....patsies would lead the league there...
     
  11. You have chosen to ignore posts from carawaydj. Show carawaydj's posts

    Re: Power Ratings, Week 13

    In Response to Re: Power Ratings, Week 13:
    ...P.S.  Gambling is for people who don't understand mathematics.  Or the nature of bookmaking operations.
    Posted by Paul_K


    How correct you are.  I should not have said that the ultimate test is how well it can predict wins against the spread, and simply said how well it can predict a win period.  The spread is just used to even out the bets.  Anyway, the moral to my story is this.  You will never gain widespread agreement as to which factors should be included/excluded, or given more weight or less weight.  However, if you were to say that out of 6 major power ranking systems, your system is the 2nd most accurate at predicting wins and so on...  Well, that is kind of the bottom line.  People can knock how you come up with it all they want, but you can't knock the end result if it is good.  

    That's one of the things missing from the various power rankings; a score to inform the readers of its accuracy in predicting wins.  Hmmm...  there's an idea.  We need a power ranking that ranks power ranking systems lol.
     
  12. You have chosen to ignore posts from DFree93HOF. Show DFree93HOF's posts

    Re: Power Ratings, Week 13

    Paul,  Stop please.  You're power rankings are junk.  You are using big words and skewed math to try and make it seem like you are the smartest guy in the room, but your rankings just show that you are a typical homer.  Your math is B.S., your "System" is B.S. and your rankings are dumb.  This is just your biased opinion wraped up in "fuzzy" math, nothing more. 
     
  13. You have chosen to ignore posts from DFree93HOF. Show DFree93HOF's posts

    Re: Power Ratings, Week 13

    I wonder if the Pats will still be number 1 now.  They were winning for the first 3 qtrs, so in Pauls system they won the game.  Is that how it works Paul?? 
     
  14. You have chosen to ignore posts from Marcus1973. Show Marcus1973's posts

    Re: Power Ratings, Week 13

    Be interesting to see the reason this fool has the Pats number 1 this coming week. I had them too high at number 7.
     
Sections
Shortcuts

Share