1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

Statistical Analysis Projections Don't Like CU

Discussion in 'Colorado Football Message Board' started by Goose, May 20, 2015.

  1. Goose

    Goose Hoops Moderator Club Member Junta Member

    Joined:
    Aug 25, 2005
    Messages:
    10,561
    Likes Received:
    1,548
    Bill C, who does some really great stuff for SBNation, just debuted his season projections for next year and they're less than positive for CU. There's improvement, just not a lot. He has us 75th in the nation - up from 77th last year, but still last in the Pac-12. In fact, there are only 6 power conference teams behind us. But we would be one of the top teams in the MWC - although still behind CSU.

    http://www.sbnation.com/college-foo...61/2015-college-football-rankings-projections
     
  2. TSchekler

    TSchekler Darth's Hero Club Member

    Joined:
    Jan 27, 2015
    Messages:
    11,194
    Likes Received:
    4,265
    Not sure if I buy 5 year performance as a legit statistical measurement of future success in college football. This isn't the stock market.
     
  3. J.R. Ewing

    J.R. Ewing Club Member Club Member

    Joined:
    Jun 17, 2011
    Messages:
    13,880
    Likes Received:
    3,612
    I really like the work that Bill does, and I've traded emails with him a handful of times on some stuff, but I've got to think that we're a little underrated with the small amount of attrition we had, and the large number of injuries. There are some other teams on his list that look pretty over-rated. Washington at 55, BC at 49, Maryland at 56, Oregon St at 74 all jump out to me.
     
  4. absinthe

    absinthe Ambitious but rubbish. Club Member Junta Member

    Joined:
    Dec 1, 2006
    Messages:
    25,885
    Likes Received:
    1,499
    Looking at the numbers one of the things i think he is really under-rating is QB play, it is so key in college football and losing a guy like Mariotta is going to set you back more than they are recognizing in this formula.
     
  5. MiamiBuffs

    MiamiBuffs Wᴉɐɯᴉ qnɟɟs Club Member

    Joined:
    Jul 12, 2005
    Messages:
    36,143
    Likes Received:
    2,025
    Why not? If recruiting is really all its cracked up to be the previous five year trend is a good gauge. The fact is the best coaches in the world can only do so much with players that hit their glass ceiling at a low level. Our recruiting remains lackluster. And thats where we are.
     
  6. TSchekler

    TSchekler Darth's Hero Club Member

    Joined:
    Jan 27, 2015
    Messages:
    11,194
    Likes Received:
    4,265
    Because there is already a recruiting category that should take that into consideration. What tangible effect does performance on the field (W/L) the past 5 years have as an indicator of future success? Using an entire category of past performance to heavily weight projections in 2015 is just a little ambiguous.
     
  7. Duff Man

    Duff Man Moderator Club Member Junta Member

    Joined:
    Jul 11, 2006
    Messages:
    39,701
    Likes Received:
    4,694
    So why do I hear about the mess MM inherited if the past five years are not a good indicator?
     
  8. Darth Snow

    Darth Snow Hawaiian Buffalo Club Member Junta Member

    Joined:
    Feb 1, 2008
    Messages:
    56,940
    Likes Received:
    4,947
    Basically says that the only way CU takes a big step forward is through coaching/staying healthy/surprise recruits. We better hope Leavitt can work wonders on D.
     
  9. TSchekler

    TSchekler Darth's Hero Club Member

    Joined:
    Jan 27, 2015
    Messages:
    11,194
    Likes Received:
    4,265
    5 year performance effects recruiting, but I don't like it being used by itself as a statistical metric to project future performance. Is Oregon's past 5 year run going to translate into another top 3 season this year?
     
  10. MiamiBuffs

    MiamiBuffs Wᴉɐɯᴉ qnɟɟs Club Member

    Joined:
    Jul 12, 2005
    Messages:
    36,143
    Likes Received:
    2,025
    In the notes section he discusses the influence of recruiting rankings on his methodology. Returning starters are also a factor.

    When you put your lower ranked guys on the field against teams with higher ranked, really highly ranked, and topped ranked guys your probability of winning sharply decreases respectively. Its probably you win 1 out of 2 against the higher ranked, 1 out of 4 against the really highly ranked, and you win 1 out of 6 (or worse) against the topped ranked. And Im probably being generous.

    As for wins and losses I'll venture that he was some kind of formula that looks at the quality of opponent and margin of victory in his rankings.
     
  11. buffedup

    buffedup Cooler than a Popsicle Stand. Club Member

    Joined:
    Dec 18, 2006
    Messages:
    7,580
    Likes Received:
    845
    I'm at a point nowadays where I just don't care about the statistics. I'm sure that Colorado was ranked pretty highly based on these stats going into 2006 and we know how that went. All Colorado needs to do is lineup on game days and win the damn game.
     
    FLounder likes this.
  12. Colobuffs21

    Colobuffs21 Club Member Club Member

    Joined:
    Jun 14, 2011
    Messages:
    538
    Likes Received:
    52
    I second this w/ #7 fUCLA as well
     
  13. tante

    tante Club Member Club Member

    Joined:
    Jul 21, 2009
    Messages:
    25,787
    Likes Received:
    1,359
    That is not how the stock market works either.
     
  14. Buffnik

    Buffnik Real name isn't Nik Club Member Junta Member

    Joined:
    Mar 20, 2009
    Messages:
    80,685
    Likes Received:
    16,384
    I think it's fair and reasonable.

    We all want to be the team that bucks the trend lines. It happens all the time, positively and negatively. But generally the teams that were good are the teams that are good this year. The teams that have recruited well based on the rankings are the teams that have the best on-field results. And vice versa.

    If I'm looking from a national perspective, I see CU as an improving program that hasn't shown it can win and about which the recruiting numbers say is outgunned.

    Nothing to get upset about here.

    They've spent a few years building the program a certain way. Usually that won't result in a breakthrough season. Sometimes it does. But if I'm setting odds or betting on things, I'm not going to say that any random program is going to improve results a ton based on continuity. I need to see something more than that.
     
  15. TSchekler

    TSchekler Darth's Hero Club Member

    Joined:
    Jan 27, 2015
    Messages:
    11,194
    Likes Received:
    4,265
    Can't tell if this is in jest. I probably should have said investment portfolio, not stock market.
     
    Last edited: May 20, 2015
  16. tante

    tante Club Member Club Member

    Joined:
    Jul 21, 2009
    Messages:
    25,787
    Likes Received:
    1,359
    It's not. Stocks follow a random walk and past performance doesn't predict a future price. Let me put it this way. The 5 year trend for apple shows a price increase but that doesn't guarantee the price will go up tomorrow or next week.
     
  17. TSchekler

    TSchekler Darth's Hero Club Member

    Joined:
    Jan 27, 2015
    Messages:
    11,194
    Likes Received:
    4,265
    My point was that one of the main performance indicators that the regular investor looks at when choosing stocks, etfs, money managers, managed account portfolios, etc. is 3, 5, 10 year trends. I didn't mean to make this comparison overly complicated.
     
  18. tante

    tante Club Member Club Member

    Joined:
    Jul 21, 2009
    Messages:
    25,787
    Likes Received:
    1,359
    Not really. They use their return to get clients but dig a hell of a lot deeper when building a portfolio.
     
  19. BlackNGold

    BlackNGold Club Member Club Member

    Joined:
    Jul 8, 2005
    Messages:
    4,499
    Likes Received:
    605
    That is depressing. Of course he is using recruiting rankings which we know do not matter....

    I hope his model is wrong...
     
  20. TSchekler

    TSchekler Darth's Hero Club Member

    Joined:
    Jan 27, 2015
    Messages:
    11,194
    Likes Received:
    4,265
    Sigh. Ok.
     
  21. Bufffan68

    Bufffan68 Club Member Club Member

    Joined:
    Sep 8, 2014
    Messages:
    4,223
    Likes Received:
    749
    Meh. As long as we win 4-6 this season I'm happy and don't care where we land on some chart.
     
  22. sackman

    sackman Club Member Club Member

    Joined:
    Jul 8, 2005
    Messages:
    48,495
    Likes Received:
    4,499
    So, he's using recruiting rankings as a predictor of this years' on-field performance and we're surprised he has us ranked so low?

    whaaa?

    We've said all along that in order for us to recruit better, we need to play better than we've recruited.
     
  23. Scotch

    Scotch Registered User Club Member Junta Member

    Joined:
    Jul 9, 2005
    Messages:
    9,809
    Likes Received:
    431
    :pcguru:
     
  24. DiggerBuffs

    DiggerBuffs Well-Known Member

    Joined:
    Jan 3, 2014
    Messages:
    1,114
    Likes Received:
    240
    Statistical analysis based off past data will typically show you who the top dogs regularly are. After that, it is a crap shoot and can lead to many difference scenarios. I take a 9/10-horse race as a for instance and ...

    FAVORITE/BEST ODDS-Usually 2/5 to 5/2 odds (Top 8 teams): A great horse is a great horse: Alabama, Ohio State, Oregon, FSU All Have Top 10 ranking average over last 5 years. In a 10 race horse card, these guys would be the favorite in any race. OSU would be considered a 2/5 horse with the likelihood of a national title repeat.

    This group usually is always in the money and wins the majority of the races. For an Alabama they are like all the horses that win 2 of 3 legs of the Triple Crown.

    The Place and Show choices-Usually 5/2 to 6:1 odds (9-19)-This group is the Baylor's, Georgia's, Auburn's, Ole Miss, LSU's, Michigan State, USC, TCU type. These horses are always in the mix but the conditions can totally throw them for a loop. They struggle in their own conference against the Top dogs and always lose that one game that keeps them from being the best. They may even be on the edge of the Top 4 but trip up somewhere.

    That horse on the cusp or fading from glory/4th choice-Usually a 6:1-12:1 odds (20-40): Boise State, Utah State, Arkansas, Mizzou, Clemson, Notre Dame, ASU, Arizona, etc.

    Absolutely no one believes this group can win the race but they have a lot of supporters that think they are close and will put money on them to maybe surprise and win their conference or push for a BCS bowl. None believe they are championship level though but hey maybe they can push the big guy and luck into a win. Also the type that is either coming back from many years ago glory or falling away from glory.

    5th/6th horse-15:1-20:1 odds (41-65): Georgia Tech, Tennessee, Mississippi State, UCLA, Utah State, etc.
    This group is nice. They may sneak into a conference title game or have shown promise. They make the Top 25 here and there but toil in the middle. One week they are great and take 2nd/3rd and upset a trifecta bet, the next they are losing a close game against the 4th/5th horse. A lot of times this horse was at the top but hasn't done anything for awhile and lost it.

    7th/8th-25:1 or worse (66-100): CU, lowest tier Power 5 schools and average non Power 5 schools:
    This group has little hope. They are usually outclassed, still trying to find the right race distance and need to find their groove. Occasionally this 7th or 8th horse finally finds that groove and finds their way into a trifecta bet because things fall into place. I take CSU as a for instance last year, Vanderbilt in 2013/2012, San Jose State in 2012, Cincy 2011. KU would be that 100 team that is in Power 5 so they can't be as bad as bottom 20 teams.

    9/10 No shot (101-120): This horse only beats a 7/8 team in a blue moon. They are thrown in to pace a race or slow down a favorite. Otherwise, they are worthless. These are the non-Power 5 group that should be at the FCS-Div 1AA level almost. Think Buffalo, Eastern Michigan, Idaho. Even at their worse CU was not here but damn close in Embrees 2012 campaign.

    Any of these horses other than the 9/10 can move up or down. CU has seen all occasions. I will do a last 50 or so years to show similarities to an old race horse:

    1963-1973 (Crowder): Started as a 5/6 horse after coming off that 2, 3/4 guy that gad some great years under Ward and Grandelius but 3 years of crap made him that 5/6 horse. From 1965-1967 he went to 3/4 and even the 2 horse. '68 brought him down to 5/6 horse but quickly spent 69-72 back in that 2-4 range.

    1973-1978 (Mallory): Same start as Crowder, but quickly was back into the 2-4 range. Left as the 5th horse and led to...

    1979-1981 (Fairbanks): Horse fell apart and by the end was 6th at start but ended as the 7th/8th horse.

    1982-1994 (McCartney): Sat in the 7th/8th range. His 2nd year made people think he could move to the 5th/6th horse but quickly got knocked back in Year 3. 1985-1987 cemented him into the 4th horse. Another big year in 1988 had people using CU in their trifecta bets as the 2nd-3rd choice. 1989 made CU a favorite for 1990 and after the national title, 1991-1996 saw CU as Favorite with some 2nd/3rd years but were considered an FSU/Oregon type that is normally on top.

    1995-1998 (New-weasel):Started on top of the world but signs showed that he was headed towards the 2/3 range at the end of '96. '97 saw CU move to a 4th choice type and kind of hang there while being a sneaky exacta/trifecta wager in Slick Rick's last season.

    1999-2005 (Barnett): What a ****ing roller coaster! Barnett started as the 4th horse in '99 and quickly went to the dreaded 5th/6th in '00. A resurgence in 2001 led CU to a 2nd/3rd horse and brought promise. 2002 stayed that course. A rough 2003 made CU go back to the 4th horse with fears of moving back to 5th/6th. 2004 and most of 2005 kept them there and maybe sniffing the money horses asses but the ending led them to wondering if 5th/6th type was up next.

    2006-2010 (He Who Shall not be named): The first game was enough for me to see we were hosed. There was no doubt, we were 5th/6th. The 2nd season brought that promise back and while still in that group but dead even with that 4th horse for our next race certainly! False hope, after the 2008 season led the Buffs to stay 5th/6th the following two years showed we were closer to the 7th/8th horse. Hawk love left and while the record showed 5th/6th the state of the program showed we were closer to being that forgotten guy that hasn't done **** in years.

    2011-2012 (Embree): Yep, we are definitely the 7th horse after 2011 but people still hold out some hope and still bet because this is CU after all. They won a national title 20 years ago. Oh ****, 2012 yep we should be shot and put down.

    2013-current (MacIntyre): Started out as the 8th horse that still gets a bet to money and the way things were ran, the way the first season played out, we were somewhere closer to 6th horse than the 8th horse after 2013 right? 2014 was like getting to 2nd base repeatedly but never getting to even get to 3rd. AKA the Rockies. Not bad enough to be the worse horse in the race but we were definitely the 7th guy and 8th in a good field.

    So that leads to 2015, I see CU somewhere along the lines of last year's NC State, Arkansas, Kentucky, Tennessee. They have struggled for many years while showing some hope at times, but still nothing special. They play in solid conferences and can win out of division, but are lucky to do better than get a couple wins in conference and finish higher than near the bottom in the Pac 12. That's why I see CU as the 5th/6th horse that will go 4-0 in non-conference and 2-7 to 3-6 in conference.

    So with my horse analogy, even the crappiest CU teams of all times weren't completely worthless. They would have been competitive in the Sun Belt, MAC, MWC or CUSA but be 3-9 or 4-8 that year. If we had played in those conferences this year, CU would have been around the Utah State, Air Force, level. This year CU would challenge for the title in most any Non-Power 5 conference in '15.
     
    FChairbanks and Lukeasaurus like this.
  25. Buffnik

    Buffnik Real name isn't Nik Club Member Junta Member

    Joined:
    Mar 20, 2009
    Messages:
    80,685
    Likes Received:
    16,384
    I declare Digger the winner (no matter what he might have written in that post).

    Mtn is ducking his head in shame and awe right about now.

    :smile2:
     
  26. FLounder

    FLounder The Buffs will rise again! Club Member

    Joined:
    Dec 14, 2011
    Messages:
    6,180
    Likes Received:
    464
    So I saw something RK posted about the buffs being set at 4.5 wins this year, besides the betting sub-thread, how and where can I bet on this.
     
  27. MiamiBuffs

    MiamiBuffs Wᴉɐɯᴉ qnɟɟs Club Member

    Joined:
    Jul 12, 2005
    Messages:
    36,143
    Likes Received:
    2,025
    I dunno. I think Sacky nailed it.

     
  28. DiggerBuffs

    DiggerBuffs Well-Known Member

    Joined:
    Jan 3, 2014
    Messages:
    1,114
    Likes Received:
    240
    If we end up with only 4 wins, this season is a bust. Need 5 bare bones minimum. Another 3 that are less than 7 either our way or the other. 3 more competitive but a score and change. Only two where we have no chance, Oregon and whoever else that may be.


    Sent from my iPhone using Tapatalk
     
  29. FChairbanks

    FChairbanks muthaka slayer Club Member

    Joined:
    Jul 8, 2005
    Messages:
    9,669
    Likes Received:
    707
    this may be the best post on allbuffs....ever.
     
  30. DiggerBuffs

    DiggerBuffs Well-Known Member

    Joined:
    Jan 3, 2014
    Messages:
    1,114
    Likes Received:
    240
    Or the most ridiculously long winded babbling crap and proof I am lucky I am married or I may be a horse gambling addict. I could have also made this a poker final table so yeah that would be an issue too.


    Sent from my iPhone using Tapatalk
     

Share This Page