WEEK FIVE NFL Picks and "Game of the Week"

  1. This post has been removed.

  2. You have chosen to ignore posts from p-mike. Show p-mike's posts

    Re: WEEK FIVE NFL Picks and "Game of the Week"

    Is it just me, or is this week's schedule littered with dog matchups?

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

    Re: WEEK FIVE NFL Picks and "Game of the Week"

    Week 5 power ratings and game of the week -- the rise of the AFC

    At this point the AFC dominates the five top slots, and the AFC East dominates the top two slots.  New England played a mercy killing game this week against decent opposition and bounced upwards.  Both the Jets and Sandy Eggo did all that and a bag of chips against rumored NFL teams.  The Jets, a visiting team playing against Buffalo that had at least some respectability prior to switching quarterbacks, gained the most in ratings.

    In other news, both New Orleans and Green Bay looked like putzes against bad teams but won.  Their ratings dropped.  Carolina isn’t terrible, but the Saints almost lost at home.  Detroit almost beat someone respectable, and St. Louis is a nearly respectable team now, which gives Oakland a slightly better chance of getting the #1 pick in the entire 2011 draft, if they hadn’t already traded that pick away.

    My stats can’t care whether Michael Vick climbs back up out of his sick bed in a week or not -- last week Philly played bad for once, and that’s what counts here.  Baltimore got a leg up on Pitt -- again, note that the expected return of Mr. Men’s Room to under center isn’t yet reflected in the current stats.

    1          nyj            13.2            (last wk. 8)

    2          ne            12.9            (last wk. 4)

    3          bal            12.3            (last wk. 7)

    4          pit            11.7            (last wk. 2)

    5          sd            11.4            (last wk. 10)

    6          gb            10.9            (last wk. 1)

    7          ind            10.8            (last wk. 6)

    8          phi            10.8            (last wk. 3)

    9          no            10.6            (last wk. 5)

    10        atl            9.8            (last wk. 9)

    11        kc            9.6            (last wk. 13)

    12        dal            9.4            (last wk. 12)

    13        min            9.3            (last wk. 11)

    14        cin            9.0            (last wk. 15)

    15        den            8.6            (last wk. 19)

    16        ten            8.4            (last wk. 14)

    17        hou            8.2            (last wk. 18)

    18        cle            7.7            (last wk. 20)

    19        was            7.3            (last wk. 23)

    20        mia            7.0            (last wk. 17)

    21        chi            6.7            (last wk. 16)

    22        stl            5.9            (last wk. 27)

    23        tb            5.8            (last wk. 22)

    24        sf            5.7            (last wk. 25)

    25        car            5.4            (last wk. 28)

    26        nyg            5.4            (last wk. 29)

    27        buf            4.3            (last wk. 21)

    28        sea            4.3            (last wk. 24)

    29        jac            3.8            (last wk. 31)

    30        az            3.4            (last wk. 26)

    31        oak            3.3            (last wk. 30)

    32        det            3.2            (last wk. 32)

    Now that New England, the Jets and San Diego all benefited from blowouts, the New York trolls and Diego fans at least may suddenly pronounce the ratings to be pretty darn accurate after a year of their complaining. 

    Perhaps I should let them have their day in the sun right now.  However, this is a good time to bring up a possible (not implemented quite just yet) adjustment in my ratings system that lowers the weighting of blowouts.  I’m finding that heavily weighting blowouts doesn’t accurately reflect a team’s future performance.  Right now I use a chi-square method of finding an optimal fit.  Instead of optimally fitting the squares (the second power) of all the variances for all games, I want to try optimally fitting the variances raised only to, say, x**1.7 versus x**2.  If you don't really dig the math, sorry about that.

    All last year the Pats were blowing out someone one week and losing by 1 point the next week.  In the end the Pats were 10-5.  The 10-5 result did not reflect the Patriots’ high power rating, and so I’m suspicious of the extra weight given for the Patriots’ blowouts.  I have read someone else’s system, and he already compensates for blowouts. 

    One possibility is to run the original and the new power ratings in parallel in future weeks, to see how people think. 

    - - - -

    In general I really like the ratings but I hate my point-spread stats.  My point spreads so far seem to reflect what the most simple-minded bettors are thinking, and so I regret to say that they are becoming effective counter indicators.  The phenomenon is worth studying.  Here goes: 

    Visitor  home          PaulK's stats   Betting public   Difference

    den       bal                                8.6            7            1.6

    nyg       hou                               7.2            3            4.2

    stl         det                               -1.0            3            -4.0

    gb        was                              -2.4            -2.5            0.1

    tb         cin                                7.8            7            0.8

    jac        buf                               3.8            0            3.8

    chi        car                               1.1            -2            3.1

    kc        ind                                4.9            8.5            -3.6

    atl         cle                                -0.1            -3            2.9

    no        az                                 -7.7            -7            -0.7

    ten        dal                                4.6            6.5            -1.9

    sd         oak                              -9.1            -6.5            -2.6

    phi        sf                                  -4.7            -3            -1.7

    min       nyj                                8.8            4            4.8

    Game of the week:

    The good news is that my games of the week usually have worked out. 

    min       nyj                                8.8            4            4.8

    I’m with Tex.  This is the Jets playing at their own Giants-Jets new home (naturally the Giants still get top billing), a place designed like an echo chamber for so much noise the fans can barely stand it, and on that slippery awful new fake turf.

    Minor calls, see below

    nyg       hou                               7.2            3            4.2

    stl         det                               -1.0            3            -4.0

    Last week: 

    Chi over giants, loser

    Atlanta over sf, loser (with the points)

    Cleveland over Cincinnati, winner

    Philadelphia over Washington, loser

    I seem to have found the holy grail of counterindicators. 

  4. You have chosen to ignore posts from jbolted. Show jbolted's posts

    Re: WEEK FIVE NFL Picks and "Game of the Week"

    1.) Minnesota Vikings (+4) @ NY Jets (Monday night): The Jets will run for less then 90 vs MIN and lose.
    PICK: Vikes;

    2.) Denver Broncos (+7) @ Baltimore Ravens: NO trap game here, the Donks will lose to a good defense on the road. PICK: Ravens;

    3.) Jacksonville Jaguars @ Buffalo (+1): Buffalo are the  worst team in the AFC:
    PICK: Jaguars;

    4.) Kansas City Chiefs (+8.5) @ Indianapolis Colts: The Chiefs' losing streak begins: 
    PICK: Colts;

    5.) St. Louis Rams (+3) @ Detroit Lions: High Scoring affair with DETROIT getting their first. If they can’t get it here, when will they?:
    PICK: Lions;

    6.) Atlanta Falcons (+3) @ Cleveland Browns: Falcons could be the NFC rep in the bowl, maybe. We’ll see.: PICK: Falcons;

    7.) Tampa Bay Buccaneers (+6.5) @ Cincinnati Bengals: I don’t think Tampa can win on the road, outside of the familiar teams:
    PICK: Bengals;

    8.) Chicago Bears @ Carolina Panthers (+2.5): Take da Bears, 20-6:
    PICK: Bears;

    9.) Green Bay Packers @ Washington Redskins (+2.5):The Pack has to be able to beat these guys by 3 right? :PICK: Packers;

    10.) NY Giants (+3) @ Houston Texans: Tex lives near Houston, I don’t
    PICK: Giants;

    11.) New Orleans Saints @ Arizona Cardinals (+7): The Cards are awful. I think the Saints will light it up in this one. After giving 40 to SD, I say at least 30 by NO. How about 30-17 :
    PICK: Saints;

    12.) San Diego Chargers @ Oakland Raiders (+6): The Chargers' without turnovers/special teams miscues are the best team in the NFL. The Bolts are way better then what the record says. They’ll be 5-2 on 10/25. 
    PICK: Chargers;

    13.) Tennessee Titans (+6.5) @ Dallas Cowboys: Did Cecil collar that bird? Titans cover in a loss:
    PICK:  Titans;

    14.) Philadelphia Eagles (+3) @ San Francisco 49ers: 9ers get their first win:
    PICK: 9ers.


  5. This post has been removed.

  6. This post has been removed.