Game Probabilities - Week 8

Weekly game probabilities are available now at the nytimes.com Fifth Down. This week I also lead-in with a re-post of an article from earlier in the year on firing coaches and regression to the mean.

  • 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

24 Responses to “Game Probabilities - Week 8”

  1. Anonymous says:

    Win Chance GAME Win Chance
    0.65 Houston at Buffalo 0.35
    0.10 Cleveland at Chicago 0.90
    0.17 Seattle at Dallas 0.83
    0.41 St. Louis at Detroit 0.59
    0.73 Denver at Baltimore 0.27
    0.07 San Francisco at Indianapolis 0.93
    0.40 Miami at Jets 0.60
    0.58 Giants at Philadelphia 0.42
    0.54 Jacksonville at Tennessee 0.46
    0.08 Oakland at San Diego 0.92
    0.16 Carolina at Arizona 0.84
    0.26 Minnesota at Green Bay 0.74
    0.12 Atlanta at New Orleans 0.88


    Brian explained how his predictions work in this post.

  2. Anonymous says:

    Here they are, all 36 of them.

    E. CAROLINA -6.......W
    DUKE +7
    N. ILLONOIS -10
    KENT +1
    SMU +17
    UTEP -6 1/2
    AIR FARCE -3 1/2
    IDAHO -3 1/2
    FRESNO ST. -15 1/2
    SAN DIEGO ST -17
    WYOMING +16 1/2
    OREGON +3 1/2
    M. TENN ST +1 1/2
    OHIO -7 1/2
    TOLEDO -5
    MISS ST +3 1/2
    TEXAS TECH -7
    HOUSTON -6 1/2
    MICHIGAN ST -5
    OREGON STATE -9 1/2
    CINCINATI -17
    GEORGIA +16
    NORTH TEXAS -10 1/2
    OLE MISS -3 1/2
    SC +5
    FLORIDA ST. -7 1/2
    MIAMI-FLA -7
    IOWA ST +7 1/2
    TROY ST. -16
    WISCONSIN -6 1/2
    MICHIGAN -7
    CONNECTICUT -7 1/2
    NEBRASKA -13
    HOUSTON -6 1/2
    PENN ST -15
    NAVY -6 1/2

    LAST WEEK'S 63.8% RIGHT ATS IS THE BAR.

    BEST TOO YA!!!

    JMED2 in GA

  3. Anonymous says:

    Hey JMED2 in GA,

    Just wondering if you have win % probability for your picks. It would be nice if you could rate your picks in order of strenght or something. Anyways, do enjoy seeing your plays.

  4. Heather says:

    Hello,

    I was introduced to this blog via Fifth Down the first week the probabilities were posted there (and now am a regular reader). I have looked through so many past posts I don't remember where I saw it, but I know that Brian found that the stronger team wins 70% of the time. I was wondering if anyone knows whether this varies based on how close (or far) the probabilities are from each other. For example, is it more likely that the SF @ Indy game and CLE @ CHI will go to the stronger team than the JAC @ TEN or NYG @ PHI?

    Thanks.

  5. Anonymous says:

    Here they are, all 36 of them.

    BOY, you people want all the SECRETS, dontcha?

    Ha-ha,,,OK

    E. CAROLINA -6.......W...was a 99% ATS pick
    DUKE +7......30% Chance of Win/51% Cover ATS
    N. ILLONOIS -10...80% Chance of Win/75% Cover
    KENT +1.....56% Chance of Win/59% Cover
    SMU +17.....14% Chance of Win/65% Cover
    UTEP -6 1/2..72% Chance of Win/99% Cover ATS
    AIR FARCE -3 1/2...75% COW/99% Cover ATS
    IDAHO -3 1/2....48% COW/51% Cover ATS
    FRESNO ST. -15 1/2...88% COW/70% ATS
    SAN DIEGO ST -17.....83%COW/99%ATS
    WYOMING +16 1/2...7% COW/90% ATS
    OREGON +3 1/2.....52% COW/51% ATS
    M. TENN ST +1 1/2...49% COW/98% ATS
    OHIO -7 1/2......63% COW/99% ATS
    TOLEDO -5........67% COW/98% ATS
    MISS ST +3 1/2...31% COW/69% ATS
    TEXAS TECH -7....68% COW/89% ATS
    HOUSTON -6 1/2...70% COW/98% ATS
    MICHIGAN ST -5...56% COW/60% ATS
    OREGON STATE -9 1/2..76% COW/89% ATS
    CINCINATI -17......91% COW/97% ATS
    GEORGIA +16........8% COW/55% ATS
    NORTH TEXAS -10 1/2...78% COW/89% ATS
    OLE MISS -3 1/2....65% COW/70% ATS
    SC +5..........29% COW/65% ATS
    FLORIDA ST. -7 1/2....84% COW/88% ATS
    MIAMI-FLA -7.......59% COW/96% ATS
    IOWA ST +7 1/2.....40% COW/51% ATS
    TROY ST. -16.......77% COW/50% ATS
    WISCONSIN -6 1/2...63% COW/65% ATS
    MICHIGAN -7........91% COW/98% ATS
    CONNECTICUT -7 1/2...70% COW/97% ATS
    NEBRASKA -13.......71% COW/98% ATS
    HOUSTON -6 1/2.....70% COW/98% ATS
    PENN ST -15........92% COW/97% ATS
    NAVY -6 1/2........72% COW/96% ATS

    LATE ADDITIONS:

    CENTRAL FLORIDA -7....65% COW/96% ATS
    TEXAS 9 1/2......79% COW/89% ATS
    GEORGIA TECH -12 1/2....82% COW/90% ATS

    IF PLAYING THE SPREAD, PLAY THOSE 90%+ ABOVE ATS FOR BEST RESULTS.

    LAST WEEK'S 63.8% RIGHT ATS IS THE BAR.

    GOOD LUCK ALL...

    JMED2 in GA

  6. Jeff Clarke says:

    JMED,

    I'm not sure I understand your numbers.COW means "Chance of Winning". ATS means "Against the spread". Correct? How can a favorite have a higher ATS number than an ATS number? If they cover the spread, they won, but not necessarily vice versa. So shouldn't the ATS number always be lower for favorites.

    I'm also highly skeptical (to say the least) of any prediction system that routinely predicts games at 95%+ probability. How are you getting these numbers?

  7. Jeff Clarke says:

    Brian,

    Your system only uses the current year stats, right? I guess I'm a little bit worried about sample size issues.

    Look at a team like Denver that went from having one of the worst defenses in the league to the best with basically the same personnel. You have to think that is at least partially a sample size related statistical fluke. I know McDaniels might have helped them more than we thought. That big a jump doesn't seem right to me. Your model consistently gives Denver a larger win pct than the Vegas lines or other models. It has been right so far. Still, I can't help but wonder if it is overextrapolating. Have you done any work in looking at how the model handled previous year's surprise teams? I keep thinking some regression to the mean is in order.

    I'm also sort of surprised by the SF/Indy prediction. There is no doubt that Indy is the better team and should be favored. This is an average team playing at a great team. Although it hasn't been true so far this year, I keep thinking that the "Any given Sunday" rule still applies in the NFL. Some regression to the mean is always to be expected with both good and bad teams. Its hard to see a 93% win prob. even when it is really good vs really bad. SF isn't even really bad. They seem to me to be almost the definition of average (perhaps slightly below). I'm curious about your thoughts on this.

  8. Anonymous says:

    Jeff,

    Yes, Chance of Winning = COW and ATS means Against the Spread.

    How can a Favorite have a higher ATS number versus an ATS number? That's easy, because its 2 different calculations.

    Example: PENN ST: 92% COW/97% ATS
    My computer model says Penn State has a 92% chance of winning the game outright. That's the COW. Winning the game even by ONE point.
    Now, the ATS is 97%. That just means PENN STATE is a -15 point favorite that has a 97% chance of covering that spread.

    Now, as an aside, the OVER/UNDER on this game is 47 POINTS. 47-15=32. So, 32+1=33. 33 is the minimum number points PENN STATE would have to score in a perfect closed system game to win this bet. However, a 16-0 game would produce the same result on your ATS bet, a Winning Wager. So I take it one step further: I take 16/33=.4848 out of .50 (I just multiple it by 2 for aesthetics on a 1-100 scale). That's how I get a 97% ATS for this game.

    At the end of the day,. it's a computer model.

    JMED2 in GA

  9. Anonymous says:

    Heather

    Are you referring to a comment from BB, that the winner covers ATS 70% of the time?
    I recall something similar from two years ago when I read this site cover to cover, over and over. I did keep track my self, and for the ’07 – ’08 NFL seasons the game winner also covered ATS, 318 for 383 or 83% of the time. Looks easy..right…??

  10. Heather says:

    Anonymous,

    Yes, I was refering to the post you mention though actually, I am in a pickem group so I don't use spread, just confidence points. I have tried to use BBs prob's while incorporating the idea that the better team doesn't always win. I think if I picked my confidence points based on his prob's I'd be better off. Do you think this is a good method? Thanks

  11. Anonymous says:

    At the end of the Day, all models & predictions are all about probabilities.

    Bottom line, nobody gets them all right, EVER!

    JMED2 in GA

  12. ben says:

    Heather,

    I have used his win percentages in my pickem with confidence points for the past two years and I average a couple of weekly highscores and have won the end of the year totals both years.

    This year I have a weekly win and am in 3rd place overall.

    So yes, use his percetages I think that they are better than Vegas and WAY better than my own sense of how good teams are.

  13. Jeff Clarke says:

    Are you referring to a comment from BB, that the winner covers ATS 70% of the time?
    I recall something similar from two years ago when I read this site cover to cover, over and over. I did keep track my self, and for the ’07 – ’08 NFL seasons the game winner also covered ATS, 318 for 383 or 83% of the time. Looks easy..right…??



    This is true but also somewhat misleading. When the underdog wins, they cover the spread 100% of the time. There are also a lot of games with spreads less than 3 points. Vegas is in essence saying: There really is no favorite. Team A will win the game only about 54% of the time. Team B will win the game 46% of the time. If we move 4% away from the favorite and to the underdog, it will be an even 50/50.

    Look at that 83% number another way. It is saying that 17% of the time the game ends up in the middle. Those 17% will always be when the favorite wins. So you know that if the favorite covers the spread 50% of the time and wins without covering 17%, the favorite will win 67% of the time. This isn't really a surprise and its basically another way of stating a number we all generally knew already.

    Of course, this year has been different so far. The favorites are winning far more often than normal and covering the spread more often than normal. Whether this is the start of a whole new anti-parity trend in the NFL or a fluke created by a fairly small sample size is open to debate.

  14. Anonymous says:

    There are LIES, damn LIES, and then there's Statistics. FDR

  15. Anonymous says:

    Anybody drop a parlay or teaser card based on my Picks this week?

    JMED in GA

  16. TheEdge says:

    Anybody drop a parlay or teaser card based on my Picks this week?

    no. get your own website.

    btw if a team only has a 92% chance of winning by 1 point they can't have an even higher chance of winning by 15.

  17. Anonymous says:

    How does this website account for a scenario like this:

    Vikings play weak teams early in the year. They use a limited playbook and don't show off their best stuff. They beat the Lions et. al., but not convincingly.

    They look weak according to this site's analysis, but then they play awfully competitively against much better teams (Ravens, Steelers, Packers at Lambeau).

    Maybe they kept their best stuff to themselves against the dregs of the leagues; then aired it out against the top teams.

    I hate to say it, but the NFL coaching staffs are a couple of steps ahead of this website. Maybe several.

    After all, they've got millions (a quarter of a billion even?), invested in their project.

  18. Anonymous says:

    "How does this website account for a scenario like this:

    Vikings play weak teams early in the year. They use a limited playbook and don't show off their best stuff. They beat the Lions et. al, but not convincingly."

    I think that Brian's model would account for this at least a little bit in that it incorporates the difficulty of a team's schedule of opponents. He does this on the assumption that stats will be higher against weaker opponents. While your scenario reverses this assumption, your scenario would still be partially reflected in this varivable because the impact of a weak opponent would otherwise be greater.

    Also, I don't understand the point of your comment about NFL coaching staff being a couple of steps ahead of this website. I doubt coaches care about people performing statistical analysis like this site does, and either way I don't see why NFL coaches would care about trying to make their stats weaker than they should be. Don't they just want to win?

  19. Anonymous says:

    "btw if a team only has a 92% chance of winning by 1 point they can't have an even higher chance of winning by 15."

    Ummmm, No, that's totally incorrect.

    You are obviously IGNORANT and never took a Logic 101 Class in College.

    I suppose you've never heard of an AND/OR & BUT/IF scenario either, so...Why even argue about it.

    And if you think winning a football game and covering a spread is a Mutually exclusive argument, you're an IGNORANT and not worth my time.

    JMED2 in GA

  20. Anonymous says:

    btw,,,I think you're right, I will go to another website.

    JMED2 in GA

  21. Anonymous says:

    JMED 2
    -first off don't worry about the ignorant people on here,the only person ignorance affects is themselves!
    -second, this is Mr. Burks site and if that is respected and he doesn't mind you posting a few numbers on here once a week then that would be great. Are you still planning on posting them or could you provide the info to where you will be posting them? Last week happens, its called sports but would like to keep reviewing your picks!
    Thanks

  22. Anonymous says:

    is the guy who post his college picks gonna put those up this week?

  23. Anonymous says:

    JMED 2 GA
    -are you going to be posting your college picks or can you share where you are now posting those?
    Thanks

  24. Anonymous says:

    Any college picks for tomorrow?

Leave a Reply