2008 NFL Week 10 Picks

Beatpaths was 8-6 last week, for a total record of 75-55. My “personal picks” that conflict with the Beatpaths picks were 1-2, so I’m at 10-9 for the year.

Denver at CLEVELAND: Denver could really help themselves with a win by shedding their beatloss to Jacksonville. Cleveland is playing for a beatwin. Cleveland is favored.

Seattle at MIAMI: Miami’s play yields rewards as they’re favored in a matchup. They actually have a beatpath to Seattle through Buffalo.

GREEN BAY at Minnesota: Green Bay has a beatpath to Minnesota through Indianapolis, as well as a direct beatwin.

New Orleans at ATLANTA: Atlanta has a very long beatpath to New Orleans. If New Orleans wins, then NO->ATL->GB->MIN->NO could be broken apart, which would hurt Atlanta.

TENNESSEE at Chicago: A good test for Tennessee. Both teams are playing for a beatwin here, so a loss would hurt Tennessee. Tennessee is favored.

BALTIMORE at Houston: A fairly even matchup, both teams are playing for a beatwin. Baltimore is favored. I’ll personally pick Houston due to home field advantage.

St. Louis at NY JETS: The NY Jets are favored due to what are probably several beatpaths to St. Louis.

BUFFALO at New England: The closest-matched game of the week, as they’re right next to each other in the rankings. Buffalo is favored. I’ll pick New England due to home field advantage.

JACKSONVILLE at Detroit: Jacksonville has a beatpath to Detroit.

CAROLINA at Oakland: The mismatch of the week, part of the NFC South’s “Demoralize Oakland” project. Carolina is favored due to having a beatpath to Oakland.

Indianapolis at PITTSBURGH: Indianapolis is playing for a beatwin, while Pittsburgh’s beatwin would be beatlooped away due to Jacksonville. Pittsburgh is favored.

KANSAS CITY at San Diego: Kansas City is actually favored due to a brutal beatpath that is pinning down San Diego. This is largely a meaningless game in terms of this week’s results, due to reinforced beatloops. I’ll personally pick San Diego due to home field advantage.

NY GIANTS at Philadelphia: A strange matchup, as neither team has beatlosses, and both teams are fairly low in the rankings. The NY Giants are favored. The Giants are playing for a beatwin, while Philadelphia would continue to have some beatloops to contend with.

San Francisco at ARIZONA: The second most lopsided matchup of the week. Arizona has a beatpath to San Francisco and is favored.

11 Responses to 2008 NFL Week 10 Picks

  1. doktarr says:

    Funky situation there with Indy/Pittsburgh – the favored team is the one that can’t get a beatpath.

    NYG-Philly is the absolute game of the week in iterative: #1 versus #2. The iterative algorithm also agrees with you and picks San Diego over KC – that’s the only conflict pick, I think.

  2. Mike T says:

    I’m a bit confused. Pittsburgh beat Jacksonville, so how would a Pittsburgh win over Indy have any beatloop potential?

  3. ThunderThumbs says:

    You’re right, it’s Indianapolis’ win that would be beatlooped away.

  4. boga says:

    Moose

    Well, my proposed method starts to make gains as early as week 4 this season. The beatloops for that week was

    DEN-NO -TB -ATL-KC-DEN
    IND-MIN-CAR-CHI-IND
    NYJ-ARI-MIA-NE -NYJ
    DEN-SD -OAK-KC -DEN
    NYJ-MIA-NE- NYJ
    KC -DEN-OAK-KC

    The iterative method eliminates 10 beatwins (the two small loops and the ind-min loop). My proposed method eliminates 7 beatwins (kc-den, mia-ne, ne-nyj, and the same ind-min loop).

    Later I’ll catch up with the rest of the season and see what the difference is on week 10.

    I coded up a rough code (I have to manually input the wins/losses and don’t have any graphical output) and not sure its completely debugged. I still need to go back and make sure I’ve kept the knowledge of every beatwin (either through beatloops or true wins) and it hasn’t made any up. I was curious was language y’all used. I used perl.

    Boga

  5. Tom says:

    Rated by “confidence”:

    NY Jets-St. Louis 100.0 (100.0-0.0)
    Carolina-Oakland 97.7 (100.0-2.3)
    Arizona-San Francisco 84.0 (90.0-6.0)
    Atlanta-New Orleans 54.6 (72.0-17.4)
    Baltimore-Houston 38.4 (84.4-46.0)
    Miami-Seattle 33.3 (40.4-7.1)
    Pittsburgh-Indianapolis 32.4 (91.7-59.3)
    Tennessee-Chicago 29.2 (100.0-70.8)
    Jacksonville-Detroit 28.0 (28.0-0.0)
    Cleveland-Denver 14.9 (38.9-24.0)
    Green Bay-Minnesota 14.5 (66.7-52.2)
    NY Giants-Philadelphia 5.6 (77.8-72.2)
    Buffalo-New England 5.4 (36.0-30.6)
    Kansas City-San Diego 0.3 (7.1-6.8)

    According to this type of ranking, Ari-SF is not the most lopsided–in fact two other games we should have greater confidence in: Jets over Rams, and Panthers over Raiders.

    Change since last week is that there are far fewer picks with a “confidence” of over 50. Lots more close games, as perceived by the beatpower rankings at least. We can predict that the beatpaths record at making picks might take a hit this week.

  6. Tom says:

    My mistake, Cards-49ers is not the *second* most lopsided, but nor is Pathers Oakland the most lopsided. Jets-St. Louis is, going by beatpower instead of straight rankings.

  7. The MOOSE says:

    I use C#

    Found a bug while running week 12 of 2004 in the iterative method. Working out the kinks so I can post more historical data.

  8. doktarr says:

    boga, could you briefly explain the difference between your approach and the standard iterative approach, or point me to a past post where you explain it?

  9. The MOOSE says:

    The difference is that Iterative starts with the smallest loops and works upwards. Boga’s suggested method first finds ALL loops of any size and eliminates the most common link from the entire set. It is sort of an extended iterative.

    Turns out the bug I mentioned above wasn’t really a bug, but rather the program choking on what appeared to be a unique situation in where the possible loops spiraled out of control. It took 20 minutes just to go through IND’s paths. I figured out a way to abbreviate the path searching though so it performs more quickly now. However, this makes me a lot more concerned with trying to actually apply Boga’s idea. It appears it would be able to save more data, but the cost could be a lot more processing time.

    My first priority right now is fleshing out my site and moving it to a new domain name.

  10. Boga says:

    doktarr,

    My original comments were in the last topics thread

    http://beatpaths.com/?p=281#comments

    What Moose said above was right.

    Boga

  11. doktarr says:

    hmm… sounds interesting. Considering all the loops at once could retain more data.

    Is this approach still removing links by reducing the weight of all the loops containing that link by an equal amount? The nice thing about that approach is that it guarantees that the total weight of all links going in and out of a given team will equal the differential in their record.

Leave a Reply

Your email address will not be published. Required fields are marked *