Game Predictions Week 9

Game probabilities for week 9 NFL games are listed below. The probabilities are based on an efficiency win model explained here and here. The model considers offensive and defensive efficiency stats including running, passing, sacks, turnover rates, and penalty rates. Team stats are adjusted for previous opponent strength. Games in which the model disagrees with consensus favorites are highlighted in red.


















VprobVisitorHomeHprob
0.11ARITB0.89
0.20CARTEN0.80
0.51CINBUF0.49
0.56DENDET0.44
0.52GBKC0.48
0.69JAXNO0.31
0.65SDMIN0.35
0.23SFATL0.77
0.71WASNYJ0.29
0.69SEACLE0.31
0.54HOUOAK0.46
0.43NEIND0.57
0.65DALPHI0.35
0.16BALPIT0.84

  • Spread The Love
  • Digg This Post
  • Tweet This Post
  • Stumble This Post
  • Submit This Post To Delicious
  • Submit This Post To Reddit
  • Submit This Post To Mixx

11 Responses to “Game Predictions Week 9”

  1. Brian Burke says:

    It looks like the efficiency model disagrees with consensus point spread favorites in more games than usual this week. There is disagreement in 7 of the 14 games. I'll highlight the disagreements in the original post.

  2. Anonymous says:

    Well, well...

    You have Indy defeating NE in what promises to be the GREATEST regular season game in NFL history.

    That's pretty bold!

    I hope you are right!

    Go Indy! :)

  3. Anonymous says:

    One more thing...

    You've got Det losing to Den.

    Detroit is undefeated at home this year winning games your model had them losing (I think) all three times.

    There is also some hard to statistically quantify BAD blood between Dry Bly (now on Den) and Detroit. He essentially called his tenure on the Lions like playing in Div II. (Det met his trade demands at the end of last year shipping him to Den for Tatum Bell.)

    You need to figure out someway to deal with these intangibles.

    Especially if Det wins.

    If not...


    ...never mind.

  4. Brian Burke says:

    Hey, my model is 3-1 in Detroit's games this year, including predicting 2 upset wins. Just last week CHI was favored by 3 over DET but I had the Lions with a 0.57 probability of winning. The only bust was DET's win over TB. I didn't see the game but was told Tampa outplayed them all game, but the Lions caught a lot of breaks and squeaked by.

    I'm really torn about the NE-IND game. I dislike both teams so much. I wish there were a way they could both lose.

  5. 24frames says:

    Funny stuff :)

    I have been making bets based on your model(completely legal where I live) and I have won a lot of units(since I only make small bets it's not a huge amount of money but it's fun to win :)

    This week I have money on all the upsets. Including, betting on the Colts to win the game. With your model expecting them to win 57% of the time and me getting 2-1 odds there's a lot of expected value in that bet. There's also a lot of value in the Jacksonville - New Orleans game.

    By the way any particular reason for disliking both the Colts and the Patriots? :)

  6. Brian Burke says:

    I'm from Baltimore and was 13 when they left town, so I don't have much affection for the Colts. I can't stand the Patriots just because they're so darn good.

    I'd be wary of the JAX-NO game. In my system this year, all weeks are treated equally, but NO has really improved since their rough start. JAX is playing without Garrard who was playing very solid football. If I could, I'd throw that prediction out the window, but that's one of the limitations of a purely statistical system, so I'll have to stick with what the numbers say.

    NE also scares me because they are scoring many more touchdowns than expected according to their offensive efficiency. They've also shown a consistent ability to win more games than expected given their overall efficiency stats (see my post "Rating Gameday Head Coaches" from June). But I would feel safe saying IND is better than a 2:1 underdog.

  7. NickC says:

    And I thought my system was doing well! I'm jealous of your 74% acccuracy in predicting winners. Using a modified ELO system I'm 47-21 in weeks 4-8 and a slightly better 82-34 for the whole year.

  8. Brian Burke says:

    Nick-We're only 3 games apart. And from what I understand ELO is very simple, which goes to your credit. I have do all kinds of crazy math just to squeeze out the extra 3 wins. And who knows, after this week I could be at 65%.

    ELO is like Sagarin's system, right?
    http://www.usatoday.com/sports/sagarin/nfl07.htm?loc=interstitialskip

    I really appreciate the simplicity there. A 6-year-old I know uses the "team with the better record is going to win" model. And tied records favor the home team. I think his accuracy is better than Football Outsiders and all their secret complicated formulas!

  9. NickC says:

    Sagarin uses something called ELO chess, but I don't know how similar it is to mine. My ELO is basically like this:
    1. Start everyone with rating 0 and add .1896 to the home team. .1896 is 57.5% of the cumulative normal distribution and the % of games home teams have won over the past several years.

    2. Use Arapad ELO's formua to calculate the expected margin of victory

    3. Compare actual margin to expected, and use ELO's update formula to update teams ratings.

    4. after updating all the team's ratings for the week, standardize ratings so the .1896 is still appropriate.

  10. Anonymous says:

    Mr. Burke,

    excellent pix as always.
    Of course, I hope you realize
    you should have listened to me when it comes to DET at home.

    Anon

  11. Brian Burke says:

    Very true!

    Denver's early season performance was exceptional, but not representative. They've been killing me every since.

Leave a Reply