Sorta quasi OT - the BCS Computer pollsters
Page 1 of 12 1234567811 ... LastLast
Results 1 to 13 of 152
  1. #1
    Suspended
    Join Date
    May 2003
    Location
    Lower Alabama
    Posts
    1,142

    Sorta quasi OT - the BCS Computer pollsters

    SI talks to some of the people behind the Computer Polls used in the former BCS calculations.

    https://www.si.com/college-football/2018/07/11/bcs-computer-rankings-polls-formula-sagarin-billingsley
    The Computer Poll Uprising: Creators of the BCS's Most Controversial Component Look Back


    I won't lie, I was a fan of the BCS system, and IMO the formula NEVER made a single mistake AFTER the last major "Corrections" following a pretty big mistake in 2003. But after that - I think the BCS got it "Right" every time.

    I still think it would have made PERFECT sense for the new playoff to simply expand to 4 teams but STILL USE the same PROVEN formula to pick the 4 teams. Instead, "we" chose a secret committee that meets behind closed doors. "The American Way" - total Non-Transparency.

    Anyway, mildly interesting read, IMO.

  2. Advertisement
  3. #2
    BamaNation Hall of Fame B1GTide's Avatar
    Join Date
    Apr 2012
    Posts
    17,186

    Re: Sorta quasi OT - the BCS Computer pollsters

    FWIW, they computer polls were also secret - the actual formulas were not released, even to the BCS committee. So the whole "behind closed doors" CFP argument seems odd since you support the computer rankings.

  4. #3
    BamaNation Hall of Fame CrimsonForce's Avatar
    Join Date
    Dec 2012
    Posts
    11,868

    Re: Sorta quasi OT - the BCS Computer pollsters

    Quote Originally Posted by TomFromBama View Post
    SI talks to some of the people behind the Computer Polls used in the former BCS calculations.

    https://www.si.com/college-football/2018/07/11/bcs-computer-rankings-polls-formula-sagarin-billingsley
    The Computer Poll Uprising: Creators of the BCS's Most Controversial Component Look Back


    I won't lie, I was a fan of the BCS system, and IMO the formula NEVER made a single mistake AFTER the last major "Corrections" following a pretty big mistake in 2003. But after that - I think the BCS got it "Right" every time.

    I still think it would have made PERFECT sense for the new playoff to simply expand to 4 teams but STILL USE the same PROVEN formula to pick the 4 teams. Instead, "we" chose a secret committee that meets behind closed doors. "The American Way" - total Non-Transparency.

    Anyway, mildly interesting read, IMO.
    I'm pretty sure the committee has picked the same 4 teams that the BCS formula would've picked in every playoff. So, the outcome has been the same but just with a little more drama..

  5. #4
    BamaNation Hall of Fame KrAzY3's Avatar
    Join Date
    Jan 2006
    Posts
    7,969

    Re: Sorta quasi OT - the BCS Computer pollsters

    Quote Originally Posted by B1GTide View Post
    FWIW, they computer polls were also secret - the actual formulas were not released, even to the BCS committee. So the whole "behind closed doors" CFP argument seems odd since you support the computer rankings.
    The BCS selection only used the computers as one aspect of their selections though. I actually liked the inclusion of the polls, they followed a sort of logic and reason that we could all understand. I mean most weeks before the polls came out we could have a pretty good idea of where teams would be ranked because we understood how it worked. That's always been harder to follow with the committee. It also preserved the historical importance of the polls (which predate most other methods of selection champions for major sports in America), yet another thing we've seen devalued under the current scheme (along with bowl games).

    The role of the computers though, was literally to guarantee logic was part of the equation. If the polls got too emotional, ranked a team too high or too low because they just lost, or something that didn't necessarily make sense, the computers would step in and correct them, point out where the numbers say the rankings should be. This didn't just work on the final poll, it helped keep the polls in check throughout the season.

    While the committee hasn't made any egregious errors yet, the biggest issue is there is nothing at all to keep them in check. Should they do something illogical, there's nothing to point them back in the right direction like the computers did for the BCS.
    Apologies if I said anything that could be construed as a personal insult.
    "Everybody that chooses to go to the game should stay there and support the team for the game." - Nick Saban

  6. #5
    BamaNation All-American
    Join Date
    Feb 2017
    Posts
    2,838

    Re: Sorta quasi OT - the BCS Computer pollsters

    Doesn't the committee use some advanced stats while making their decision? I'm not in love with the committee format but so far I don't think they have done anything egregious

  7. #6
    BamaNation All-SEC
    Join Date
    Sep 2000
    Location
    Arlington, Tx
    Posts
    1,381

    Re: Sorta quasi OT - the BCS Computer pollsters

    And honestly, we don't know what kind of multipliers or algorithms is going on behind the scenes to get to those numbers. They could arrange the formulas in any number of ways in order to favor one side versus the other. Ask any financial analyst to run a projection and he can give you different any number of answers based on what he thinks you want the answer to be (ask me how I know). The same could be done with the BCS formula.

  8. #7
    BamaNation Hall of Fame B1GTide's Avatar
    Join Date
    Apr 2012
    Posts
    17,186

    Re: Sorta quasi OT - the BCS Computer pollsters

    Quote Originally Posted by KrAzY3 View Post
    While the committee hasn't made any egregious errors yet, the biggest issue is there is nothing at all to keep them in check. Should they do something illogical, there's nothing to point them back in the right direction like the computers did for the BCS.
    To be clear - they have NO mistakes. None. Get back to me when they make one.

  9. #8
    BamaNation Hall of Fame B1GTide's Avatar
    Join Date
    Apr 2012
    Posts
    17,186

    Re: Sorta quasi OT - the BCS Computer pollsters

    Just watch this a few times and forget about your doubts about the committee:


  10. #9
    BamaNation Hall of Fame GrayTide's Avatar
    Join Date
    Nov 2005
    Location
    Greenbow, Alabama
    Posts
    13,954

    Re: Sorta quasi OT - the BCS Computer pollsters

    I still do not like the Playoff Committee, just use the BCS formula as before and let it select the best 4 teams.
    "My momma always said you got to put the past behind you before you can move on." Forrest Gump

    "The past is never dead. It's not even past." William Faulkner

    "Get off my lawn." Clint Eastwood in Gran Torino

  11. #10
    Suspended
    Join Date
    May 2003
    Location
    Lower Alabama
    Posts
    1,142
    Thread Starter

    Re: Sorta quasi OT - the BCS Computer pollsters

    Quote Originally Posted by B1GTide View Post
    FWIW, they computer polls were also secret - the actual formulas were not released, even to the BCS committee. So the whole "behind closed doors" CFP argument seems odd since you support the computer rankings.
    True, the actual formulas (except for Colley) were never made public. BUT - Respectfully, the result of EVERY Computer Poll was published EVERY WEEK. And the METHODOLOGY for arriving at the consensus of the computers (which was NEVER a simple "average" or "majority rule") was likewise published.
    We always knew, and we could cross check to see if the final numbers were correctly calculated.

    Do we know how any individual member of the "Double Secret Committee" has voted, at any point during the past 4 years?
    Do we know which members voted UA #4 and which voted us #5 last December? (we know for the AP poll each week, and we know for the final "Coaches Poll" and we knew for the final Harris poll as well).
    Do we know what Criteria (if any) MUST be considered by the committee?
    Do we know the committee arives at it's results? What kinds of numbers are generated? or who they are tabulated?
    Do the members vote by ordinals? or do they assign numerical scores?
    Is the final poll a Simple average? Weighted average? Mean? Median?

    Respectfully, the answer to ALL those questions, is "NO". We just don't know. It's all a SECRET. NOTHING can be "checked" or "verified".

    All we know is the final result.


    Quote Originally Posted by CrimsonForce View Post
    I'm pretty sure the committee has picked the same 4 teams that the BCS formula would've picked in every playoff. So, the outcome has been the same but just with a little more drama..
    Well, RESPECTFULLY - yes and no. It's true that when we run "Proxy" BCS calculations, we almost always get the same result, but there are some problems there.

    First and foremost - 1/3 of the BCS formula - the Harris Interactive Poll, no longer exists. So there's no way to guess how that Poll would have operated week to week - the Harris Poll DIDN'T always agree with the "Coaches" or the "Computers".

    Second is the problem of cross-contamination, due to the "sheep" mentality of most of the Sportswriters (and many of the SID's and other functionaries who actually vote the "Coaches" poll). - NONE of them want to be singled out for being "Wrong" - at least not at the end. So once the first "Double Secret Committee" poll comes out - the AP and Coaches polls ALWAYS fall into line within a couple weeks. This is due to the FEAR of most sportswriters of someone asking them in January why they voted for the "Wrong" team.

    So what happens is, by the time the final Four teams are selected, the AP and Coaches polls have each become nearly UNANIMOUS in Mirroring the Committee Poll, again, out of "fear" of being called out for being "wrong".

    Here's just ONE example - last year, week 11 the first Committee poll is released, and 9-0 Wisconsin is ranked #8. The same week, the Coaches poll had Wisconsin at #3.
    The next week Wisconsin wins again and goes to 10-0. And they didn't "Just" win - they beat a RANKED team, On the Road, by 24 points. But the Coaches moved them DOWN to #4???
    Why?
    Well, we all KNOW why, don't we? Because no one wants to be put on the spot by the media and asked why they voted differently than the "All Powerful OZ Committee".

    So what you have is a classic "Self Fulfilling Prophecy". Happens every year.

    The point being, this kind of "Cross contamination" really invalidates any serious attempt to divine what the ACTUAL BCS might have done.

  12. #11
    BamaNation Second Team bamaltc's Avatar
    Join Date
    Dec 2010
    Location
    Florence, AL
    Posts
    128

    Re: Sorta quasi OT - the BCS Computer pollsters

    Quote Originally Posted by GrayTide View Post
    I still do not like the Playoff Committee, just use the BCS formula as before and let it select the best 4 teams.
    Agree. Most people agree the BCS got it right most of the time with the top 2. If it were the top 4, it would have been even better.

  13. #12
    BamaNation Hall of Fame B1GTide's Avatar
    Join Date
    Apr 2012
    Posts
    17,186

    Re: Sorta quasi OT - the BCS Computer pollsters

    Quote Originally Posted by TomFromBama View Post
    True, the actual formulas (except for Colley) were never made public. BUT - Respectfully, the result of EVERY Computer Poll was published EVERY WEEK. And the METHODOLOGY for arriving at the consensus of the computers (which was NEVER a simple "average" or "majority rule") was likewise published.
    Computer polls are meaningless the first few months of the season - ask any one of these guys who created these models and they will tell you that. There simply is not enough data. So, yes, they were published, but they were also meaningless until November. And, as a result, the computer rankings change wildly until November. Not sure how this does anything to make you feel warm and fuzzy.

    Result - you have secret formulas producing meaningless data for months and that makes you feel better when we get to November, even though the November ranking are wildly different from the September and October rankings. But, hey, at least they released the results of their secret polls weekly.

    The bottom line - when it matters, both publish their results every week, and both are very open about the methodology used. You are not going to be satisfied unless you get to sit in the room with them while they make their decisions. But you take off the tin foil hat while computer geeks play with their secret formulas every week, trusting them completely.

    Either wear the hat or take it off. Please. I can't decide if you are really crazy, or maybe just trying to convince us that you are crazy.

  14. #13
    BamaNation All-American tusks_n_raider's Avatar
    Join Date
    May 2009
    Posts
    3,836

    Re: Sorta quasi OT - the BCS Computer pollsters

    The committee has picked the correct 4 teams all 4 years with ZERO mistakes.

    Their #4 pick has been validated TWICE by ultimately winning the NCG.

    Last year was certainly entertaining for me. I knew they would pick us #4. There was no logical argument for the alternative.... yet so many people were freaking out about it.

    I'm not going to dig up the old threads but if anyone is so inclined to search you will see dozens of posts by me trying to reassure everybody before the selection that the committee was on the up and up and would make the right decision. It fell on deaf ears mostly but I tried...lol

    I'm more than happy with the Playoff committee and have no urge for the BCS to ever come back.

    Now the New CFP NC Trophy is another matter.... THAT thing STILL looks gawd-awful. The Crystal Ball is 100000000X's better.

Bookmarks

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  

 

TideFansStore.com: Get YOUR gear!

[CLICK] GET BAMA Calendar