Kenpom handicaps the SEC

Discussion in 'Nuttin' but Net' started by REM08, Jan 2, 2014.

  1. REM08
    Offline

    REM08 Well-Known Member

    Joined:
    Jun 1, 2010
    Messages:
    4,861
    Likes Received:
    198
    Trophy Points:
    63
    Location:
    Arkansas
    Ratings Received:
    +706
    He mentions it being a close call between UK and Florida but his system leaned Kentucky. He simulates the season 10,000 times and the SEC's results look like this:

    http://kenpom.com/blog/index.php/weblog/entry/conference_race_simulations_part_iii

    Not trying to start a fight here, just always think kenpom's stuff is interesting. I'd actually, at this point, lean Florida over UK because of the consistency required to have the best record over 18 games in the league. I can make arguments both ways but that's the one that pulls me the most.

    I was surprised how relatively close UT is in the simulations.

    For what its worth, here are last years pre-conference SEC simulations:

    • Informative Informative x 4
    • Optimistic Optimistic x 1
  2. g8rboy
    Offline

    g8rboy Well-Known Member

    Joined:
    Apr 10, 2007
    Messages:
    1,961
    Likes Received:
    199
    Trophy Points:
    63
    Location:
    Spring Hill, FL
    Ratings Received:
    +907
    vandy has fallen hard, huh?
  3. REM08
    Offline

    REM08 Well-Known Member

    Joined:
    Jun 1, 2010
    Messages:
    4,861
    Likes Received:
    198
    Trophy Points:
    63
    Location:
    Arkansas
    Ratings Received:
    +706

    No kidding. But when a good coach like Kevin Stallings leaves....wait, what? He's still there?
    • Funny Funny x 4
    • Like Like x 1
  4. ThePlayer
    Offline

    ThePlayer VIP Member

    Joined:
    Apr 3, 2007
    Messages:
    28,376
    Likes Received:
    526
    Trophy Points:
    113
    Ratings Received:
    +2,428
    You can't miss his bald head...nor the smug look on his face.
  5. ApexNC
    Online

    ApexNC Well-Known Member

    Joined:
    Apr 8, 2007
    Messages:
    8,272
    Likes Received:
    343
    Trophy Points:
    83
    Location:
    NC
    Ratings Received:
    +1,466
    Two interesting things to me:

    1) Statistically, Mizzou has almost no chance 440/10,000
    2) Top two in the acc are Syracuse and Pitt….LMAO
  6. HallGator
    Offline

    HallGator Administrator VIP Member

    Joined:
    Apr 3, 2007
    Messages:
    45,098
    Likes Received:
    1,223
    Trophy Points:
    113
    Location:
    Outer Limits
    Ratings Received:
    +4,648
    Big drop off after Florida and Kentucky.
  7. BengermanV
    Offline

    BengermanV Well-Known Member

    Joined:
    Apr 22, 2009
    Messages:
    4,545
    Likes Received:
    262
    Trophy Points:
    83
    Ratings Received:
    +883
    Ken Pom is fantastic. Agree with pretty much everything he says here. Going to be a wild race!
  8. themistocles
    Offline

    themistocles Well-Known Member

    Joined:
    Apr 8, 2007
    Messages:
    12,849
    Likes Received:
    96
    Trophy Points:
    48
    Ratings Received:
    +462
    Yes, MASSIVE - with Ky - 3765, UF - 3277 and Tenn in third at 1213 it is almost 3 to 1 UF and KY over Tenn.

    The one thing I think is in Florida's advantage is experience, it is utterly impossible to statistically estimate the impact of having 4 four year seniors on a team.

    Interesting that Missouri was only 448, I think he isn't giving their coach credit.
  9. MadduxFanII
    Offline

    MadduxFanII Well-Known Member

    Joined:
    Apr 10, 2007
    Messages:
    5,350
    Likes Received:
    111
    Trophy Points:
    63
    Ratings Received:
    +469
    Those SEC projections seem incredibly close. I don't know enough about statistics to say if the difference there is statistically significant (I suspect it is), but Ken Pom's system obviously doesn't see much difference between the two schools.
  10. MadduxFanII
    Offline

    MadduxFanII Well-Known Member

    Joined:
    Apr 10, 2007
    Messages:
    5,350
    Likes Received:
    111
    Trophy Points:
    63
    Ratings Received:
    +469
    I do like all the schools that don't win a single projected season.

    Sorry South Carolina, Mississippi State, Vanderbilt, there are literally no parallel universes in which you can win the conference.
  11. GatorLurker
    Offline

    GatorLurker Well-Known Member

    Joined:
    Apr 8, 2007
    Messages:
    2,958
    Likes Received:
    139
    Trophy Points:
    63
    Ratings Received:
    +483
    Ken Pom's predicted winner would vary from day to day. Some days UF was ranked higher and most recently UK after beating the Cards who have a profile that over-rates them by his metrics and hence UK got a big bump.

    Ken Pom is really good, but his methods totally over-rated UF last year. I emailed him a couple of times saying "UF is pretty good, but we ain't the best nor second best team so WTF is wrong with your methods?" He agreed, but didn't offer any explanation about the problems with his methods. This year Louisville and Pitt as well as Wisconsin are over rated by his algorithm. I am thinking that it might have a bias toward slow pace as a good thing.
  12. sixoburn
    Offline

    sixoburn Active Member

    Joined:
    Apr 9, 2007
    Messages:
    1,295
    Likes Received:
    6
    Trophy Points:
    38
    Ratings Received:
    +20
    Gatorlurker: Kenpom (and most advanced stats) care more about scoring margin than win/loss. UF, for whatever reason, absolutely crushed teams last year when they won, especially early in the season. This looks really good in his system nd any advanced stat that takes into account margin of victory (they all should, it is the best predictor of future success). The difference between a 15 point win and a 2 point win is greater than a 2 point in and a 5 point loss. I believe his stats take into account pace, and what he is looking at is actually points per posession. If you read basketball prospectus last year, which was a nice site while it lasted, you saw that UF was posting all time high margin of PPP differential (scored versus allowed) and this made UF look really, really good. It helps that UF has had senior laden teams recently (man, it feels good to write that. I remember in the early-mid 2000s we had so many transfers we would graduate like one player. Now we seem to be plus on transfers... good job billy) and so we should play well early in the year, except when half the team is injured or suspended. I don't believe there is any bias towards pace, just that at a slower pace the same margin of victory is going to be more impressive in his system than the same margin of victory in a high possession fast paced game. It took fewer possessions to build the lead and so winning 55-50 is going to be better than 82-77, assuming similar scoring efficiencies.
    • Like Like x 1
  13. GatorLurker
    Offline

    GatorLurker Well-Known Member

    Joined:
    Apr 8, 2007
    Messages:
    2,958
    Likes Received:
    139
    Trophy Points:
    63
    Ratings Received:
    +483
    Yeah, I know this. So, isn't it a flaw in his system?
  14. gator7_5
    Offline

    gator7_5 Well-Known Member

    Joined:
    Apr 9, 2007
    Messages:
    12,744
    Likes Received:
    220
    Trophy Points:
    63
    Ratings Received:
    +905
    meh, we haven't been near full strength for the stats he's basing this on. UF in a landslide.
  15. corpgator
    Offline

    corpgator Well-Known Member

    Joined:
    May 9, 2010
    Messages:
    6,319
    Likes Received:
    164
    Trophy Points:
    63
    Ratings Received:
    +792
    His system always overates slower teams in terms of pace like us and even more so Wisconsin. Also if you at the team in top 10 all you have to do is look at sos to see who is overrated right now
  16. GatorRade
    Offline

    GatorRade Well-Known Member

    Joined:
    Apr 3, 2007
    Messages:
    7,305
    Likes Received:
    343
    Trophy Points:
    83
    Ratings Received:
    +1,181
    I think the question is "How do you know that they aren't the second best team?" I actually think that they were a top 2-3 team last year. I made a post last year in response to this idea that KenPom is biased toward slow pace. Not true, I think. If it has a bias, it is one against variance. Winning 10 by 50 points and losing 10 games by 1 point would probably put you as the top KenPom team, even though you are a 50% team.

    That said, his algorithm is never going to be perfect, and everyone should know that.
  17. GatorLurker
    Offline

    GatorLurker Well-Known Member

    Joined:
    Apr 8, 2007
    Messages:
    2,958
    Likes Received:
    139
    Trophy Points:
    63
    Ratings Received:
    +483
    I also think that Ken Pom's metrics undervalue UF this year.

    YMMV.
  18. GatorRade
    Offline

    GatorRade Well-Known Member

    Joined:
    Apr 3, 2007
    Messages:
    7,305
    Likes Received:
    343
    Trophy Points:
    83
    Ratings Received:
    +1,181
    His model certainly wasn't impressed with the close win against Richmond. Gators fall to 17th today.
  19. GatorLurker
    Offline

    GatorLurker Well-Known Member

    Joined:
    Apr 8, 2007
    Messages:
    2,958
    Likes Received:
    139
    Trophy Points:
    63
    Ratings Received:
    +483
    Probably the losses by UCONN and Memphis more than anything else.
  20. corpgator
    Offline

    corpgator Well-Known Member

    Joined:
    May 9, 2010
    Messages:
    6,319
    Likes Received:
    164
    Trophy Points:
    63
    Ratings Received:
    +792
    Those 2 losses of past opponents and the Richmond win contributed. College basketball home court advantage accounts for about 5 points, so it was only a 4 point win at a neutral site, when his system pegged us as 9-10 pt winners going in.

Share This Page