Home› Hardcore Husky Board

Howdy, Stranger!

It looks like you're new here. Sign in or register to get started.

Sign In Register
Welcome to the Hardcore Husky Forums. Folks who are well-known in Cyberland and not that dumb.

Another Pre-Season Analytic Projection

DoogCouricsDoogCourics Member Posts: 5,739
in Hardcore Husky Board
Rob Bowron enjoyed Bill Connelly's S&P+ but felt there were a few things he could improve on. He created Beta_Rank. I found it interesting so I decided I'd throw it on here for anyone interested. Think of it as the free version of the Pick Six Predictions I threw down in the WAM.

If you're interested in knowing more about it and reading how he comes up with his projections:

https://www.sharpcollegefootball.com/post/what-is-beta_rank


First, explanation of terms:









Next, our shitty offensive history:



And, our dominant defensive history:





Now, our team dashboard and schedule analysis:








Lastly, the 2019 conference comparison:






· Share on FacebookShare on Twitter
«134

Comments

  • DoogCouricsDoogCourics Member Posts: 5,739
    GrundleStiltzkin said:

    Might as well not even play the games then

    Then what would I watch all fall?
    · Share on FacebookShare on Twitter
  • FremontTrollFremontTroll Member Posts: 4,744
    He hates us.

    Luckily his model is total garbage.

    At least preseason. Maybe it will improve although I doubt it since the returning production portion seems to be the most solid part.
    · Share on FacebookShare on Twitter
  • ArcArc Member, Swaye's Wigwam Posts: 113 Swaye's Wigwam
    GrundleStiltzkin said:

    Might as well not even play the games then

    Might as well not even talk about the team on internet forums then.
    · Share on FacebookShare on Twitter
  • DoogCouricsDoogCourics Member Posts: 5,739
    edited August 2019
    FremontTroll said:

    He hates us.

    Luckily his model is total garbage.

    At least preseason. Maybe it will improve although I doubt it since the returning production portion seems to be the most solid part.

    Reading through his website, in season he hits 80% predictions on who will win games, but preseason projections are just below 60% hit rates. So it definitely is more accurate in season as the data begins to compile.

    He admits himself that the data skews towards returning production in his preseason models.

    Not sure exactly who he is but he does pods with the Arizona 247 mod, and he claims he is an econometrician.
    · Share on FacebookShare on Twitter
  • FremontTrollFremontTroll Member Posts: 4,744
    Not attacking you always appreciate you and your service.

    Looking at bad models is still interesting. Will follow it to see if it is better in-season.
    · Share on FacebookShare on Twitter
  • dtddtd Member Posts: 5,097 Standard Supporter
    Returning starts has been proven to be one of the least correlated of the correlated stats(to winning).
    · Share on FacebookShare on Twitter
  • Mad_SonMad_Son Member Posts: 10,194
    edited August 2019
    FremontTroll said:

    DoogCourics said:

    FremontTroll said:

    He hates us.

    Luckily his model is total garbage.

    At least preseason. Maybe it will improve although I doubt it since the returning production portion seems to be the most solid part.

    Reading through his website, in season he hits 80% predictions on who will win games.

    Not sure exactly who he is but he does pods with the Arizona 247 mod, and he claims he is an econometrician.
    If you just pick the favorite in every game you will be about 80%. The question is how he does against the spread. If he doesn’t publicize that it’s probably not good.

    Put it this way I noticed he loves Utah so I looked at one game- Utah vs. BYU.

    He has it at Utah -13.5. Spread is actually -4.5.

    If his model were correct the Kelly Criterion says that line is so far out of whack he should bet 40% of his bankroll on it.
    80% straight up is really good if it is achieved early in the season (which they of course can't). Otherwise yeah, it's about par for the course.

    http://www.thepredictiontracker.com/ncaaresults.php?orderby=wpct desc&type=1&year=18

    https://hardcorehusky.com/discussion/53705/what-you-should-know-about-upsets-81-simple-facts-that-will-change-how-you-view-the-world/p1
    · Share on FacebookShare on Twitter
  • DoogCouricsDoogCourics Member Posts: 5,739
    dtd said:

    Returning starts has been proven to be one of the least correlated of the correlated stats(to winning).

    This.

    Because I believe 40% of last year’s snaps actually return on defense since we rotate so much. That’s definitely not dead last in the country.

    Plus half the defense has started multiple games, they just were full time starters.
    · Share on FacebookShare on Twitter
  • theknowledgetheknowledge Member, Swaye's Wigwam Posts: 5,236 Founders Club
    When do we have enough blue chip players or a strong enough team philosophy to qualify for the alpha rank? DDY wants to know.
    · Share on FacebookShare on Twitter
  • FremontTrollFremontTroll Member Posts: 4,744
    dtd said:

    Returning starts has been proven to be one of the least correlated of the correlated stats(to winning).

    He didn’t use returning starts though.
    · Share on FacebookShare on Twitter
  • dncdnc Member Posts: 56,812
    A whole thread about Beta Rank and not one person has tagged @NorwegianHusky?

    This place is slipping.
    · Share on FacebookShare on Twitter
Sign In or Register to comment.