Home General
Hey folks - as a member of the DawgNation community, please remember to abide by simple rules of civil engagement with other members:

- Please no inappropriate usernames (remember that there may be youngsters in the room)

- Personal attacks on other community members are unacceptable, practice the good manners your mama taught you when engaging with fellow Dawg fans

- Use common sense and respect personal differences in the community: sexual and other inappropriate language or imagery, political rants and belittling the opinions of others will get your posts deleted and result in warnings and/ or banning from the forum

- 3/17/19 UPDATE -- We've updated the permissions for our "Football" and "Commit to the G" recruiting message boards. We aim to be the best free board out there and that has not changed. We do now ask that all of you good people register as a member of our forum in order to see the sugar that is falling from our skies, so to speak.
Options

ESPN's FPI....Are you kidding me?!

MinnesotaDawgMinnesotaDawg Posts: 552 ✭✭✭✭✭ Graduate

I try not to get hung up on irrelevant rankings too much....certainly the AP and Coaches poll have their flaws and biases. But with Georgia's relevancy in the playoff discussion this year, I'm certainly more sensitive to the discussion of rankings this year. That's why I continue to bristle at the reference to ESPN's FPI rankings in this debate. Over the course of last weekend's games (Thurs through Sat.) I saw the side-by-side comparison graphic of the committee's Top 5 with the FPI Top 5 on no less than four different game broadcasts. This wouldn't be so troubling, if the FPI wasn't so bizarre in its rankings...having Georgia at number 5 behind one-loss Ohio State (2) and Penn State (3) last week. Out of curiosity, I just checked out the FPI rankings AFTER this weekend's games...

  1. Alabama
  2. Ohio State (still!?)
  3. Penn State (****??)
  4. Washington (huh?)
  5. Notre Dame
  6. Georgia (down one spot)

I guess that the FPI is a model, that uses a ton of game simulations based on actual and projected results...but it seems more like the rankings of an old man living in a basement in Youngstown, Ohio. He doesn't actually follow college football, but did read the sports page over the course of a few weeks last December..... Weird.

«13

Comments

  • Options
    MacDawg15MacDawg15 Posts: 3,057 ✭✭✭✭✭ Graduate

    That crap is biased due to panel members. Let's just shut the "experts" up and win. We control what we do.. I have no idea what playoff rankings are needed RIGHT NOW for, except to stir controversy. Big 10 and ND lovers smh

  • Options
    KaseyKasey Posts: 28,890 mod

    Bear and Stanford Steve have been against Georgia all season. These things are nuts. It’s why we play the games.

    Loooooove being overlooked though

  • Options
    JMFDJMFD Posts: 219 ✭✭✭ Junior

    Who cares what ESPCNN turds think!!

  • Options
    MacDawg15MacDawg15 Posts: 3,057 ✭✭✭✭✭ Graduate

    What's so crazy is if you look at Bear's and some other "experts" on Gameday's picks; they hover around 50-60% accurate, and these are people that do nothing but scout football teams lol. How are they "experts"?

  • Options
    MIghtydawgMIghtydawg Posts: 984 ✭✭✭✭✭ Graduate

    @MacDawg15 said:
    That crap is biased due to panel members. Let's just shut the "experts" up and win. We control what we do.. I have no idea what playoff rankings are needed RIGHT NOW for, except to stir controversy. Big 10 and ND lovers smh

    FPI is computer based. It must be that the computers are biased.

    The computer rankings are only as good as the information put in. This is part of the reason that the BCS was eliminated. Now it is 13 people sitting in a room ranking teams.

  • Options
    MacDawg15MacDawg15 Posts: 3,057 ✭✭✭✭✭ Graduate

    I didn't catch him last week. But I remember the previous week he was like 10-6 or 9-7. Bottom line is anything can happen week to week. Knowing the pulse of a team imo let's us know who's a favorite etc. That's why it's great to be a DAWG because nobody can dispute how focused this team is and the unseen recruiting wins in our history that we're being blessed with. Besides Bama and a few more, I don't see but a handful of programs moving 100% in the right direction with us. (Big picture)

  • Options
    MIghtydawgMIghtydawg Posts: 984 ✭✭✭✭✭ Graduate

    You need to also understand what FPI is--The Football Power Index (FPI) is a measure of team strength that is meant to be the best predictor of a team's performance going forward for the rest of the season. FPI represents how many points above or below average a team is. Projected results are based on 10,000 simulations of the rest of the season using FPI, results to date, and the remaining schedule.

    It is not ranking UGA as the 6th best team in the country. Rather the ranking is a reflection of the Dawgs upcoming schedule, which is ranked as the 9th toughest remaining in the country. Here is UGA's FPI resume, which is pretty strong.

    http://www.espn.com/college-football/team/fpi?id=61&year=2017

  • Options
    Denmen185Denmen185 Posts: 7,407 ✭✭✭✭✭ Graduate

    @MacDawg15 said:
    I didn't catch him last week. But I remember the previous week he was like 10-6 or 9-7. Bottom line is anything can happen week to week. Knowing the pulse of a team imo let's us know who's a favorite etc. That's why it's great to be a DAWG because nobody can dispute how focused this team is and the unseen recruiting wins in our history that we're being blessed with. Besides Bama and a few more, I don't see but a handful of programs moving 100% in the right direction with us. (Big picture)

    Maybe we are looking at different things as I was referring to the 3 picks on Game Day only. There he was something around 17-9-1 I think for the season. You appear to be talking about some wider predictions for the week only.

  • Options
    MacDawg15MacDawg15 Posts: 3,057 ✭✭✭✭✭ Graduate

    He WAS doing a top 25 pick em. Don't know if that's still going down or not.

  • Options
    MacDawg15MacDawg15 Posts: 3,057 ✭✭✭✭✭ Graduate

    But I know the 3 your talking about.

  • Options
    MacDawg15MacDawg15 Posts: 3,057 ✭✭✭✭✭ Graduate

    FPI is computer based. It must be that the computers are biased.

    The computer rankings are only as good as the information put in. This is part of the reason that the BCS was eliminated. Now it is 13 people sitting in a room ranking teams.

    Well after the PSU loss and OSU spanking in Iowa those are the dumbest computers I've heard of. Neither team should have a prayer to get in the CFP! Up to THIS day, the 4 best teams in the country are us, Bama, ND, and CU

  • Options
    MIghtydawgMIghtydawg Posts: 984 ✭✭✭✭✭ Graduate

    That is not what it is saying. It is not ranking the top teams for the season. It is predicting who has the best chance of winning going forward.

  • Options
    MacDawg15MacDawg15 Posts: 3,057 ✭✭✭✭✭ Graduate

    I understand what the poll's intent is, and what your saying. I'm JS those are the 4 best teams I've seen week in and week out. With Wisky possibly being there determining how they finish.

  • Options
    donmdonm Posts: 10,241 ✭✭✭✭✭ Graduate

    @MIghtydawg said:

    @MacDawg15 said:
    That crap is biased due to panel members. Let's just shut the "experts" up and win. We control what we do.. I have no idea what playoff rankings are needed RIGHT NOW for, except to stir controversy. Big 10 and ND lovers smh

    FPI is computer based. It must be that the computers are biased.

    The computer rankings are only as good as the information put in. This is part of the reason that the BCS was eliminated. Now it is 13 people sitting in a room ranking teams.

    Yeah, 13 people who picked Ohio State last year and left PSU out. There may be more bias in those 13 people than any computer model around. Who the heck is more biased than people? They are merely pretending to be objective as in the "eye" test.

  • Options
    MinnesotaDawgMinnesotaDawg Posts: 552 ✭✭✭✭✭ Graduate

    @MIghtydawg said:
    That is not what it is saying. It is not ranking the top teams for the season. It is predicting who has the best chance of winning going forward.

    No, that's not right...It's NOT just a prediction of who has the best chance to win going forward. There is actually a different calculation about who has the best chance to win going forward called the "Win Out %" (again, based on it's FPI model). FWIW, Penn State (FPI #3) has the best chance to win out (88.5%), then it's Clemson (FPI #7) with a win-out percentage of 65.7%.

    It actually IS supposed to be a ranking of who the BEST teams are based on theoretical head to head matches. Put another way, the model ranks teams based on how the way it predicts the results of hypothetical games. Example, if asked to predict Georgia (6) v. Washington (4) now, FPI model favors Washington. If asked to predict Clemson (7) v. Georgia (6), FPI model favors Georgia.

  • Options
    BamaDawgBamaDawg Posts: 2,523 ✭✭✭✭✭ Graduate

    Not to belittle anyone else's concerns and I hate to repost an earlier post of mine, but it seems very appropriate here:

    Don't care!

  • Options
    Denmen185Denmen185 Posts: 7,407 ✭✭✭✭✭ Graduate

    @MinnesotaDawg said:

    @MIghtydawg said:
    That is not what it is saying. It is not ranking the top teams for the season. It is predicting who has the best chance of winning going forward.

    No, that's not right...It's NOT just a prediction of who has the best chance to win going forward. There is actually a different calculation about who has the best chance to win going forward called the "Win Out %" (again, based on it's FPI model). FWIW, Penn State (FPI #3) has the best chance to win out (88.5%), then it's Clemson (FPI #7) with a win-out percentage of 65.7%.

    It actually IS supposed to be a ranking of who the BEST teams are based on theoretical head to head matches. Put another way, the model ranks teams based on how the way it predicts the results of hypothetical games. Example, if asked to predict Georgia (6) v. Washington (4) now, FPI model favors Washington. If asked to predict Clemson (7) v. Georgia (6), FPI model favors Georgia.

    Part of the reason for Penn State having a higher win out % is because they are not expected to play in the championship game hence, even though rated behind OSU they have a much higher % of not losing in 3 games than OSU has of losing in 4.

  • Options
    MinnesotaDawgMinnesotaDawg Posts: 552 ✭✭✭✭✭ Graduate

    @Denmen185 said:

    @MinnesotaDawg said:

    @MIghtydawg said:
    That is not what it is saying. It is not ranking the top teams for the season. It is predicting who has the best chance of winning going forward.

    No, that's not right...It's NOT just a prediction of who has the best chance to win going forward. There is actually a different calculation about who has the best chance to win going forward called the "Win Out %" (again, based on it's FPI model). FWIW, Penn State (FPI #3) has the best chance to win out (88.5%), then it's Clemson (FPI #7) with a win-out percentage of 65.7%.

    It actually IS supposed to be a ranking of who the BEST teams are based on theoretical head to head matches. Put another way, the model ranks teams based on how the way it predicts the results of hypothetical games. Example, if asked to predict Georgia (6) v. Washington (4) now, FPI model favors Washington. If asked to predict Clemson (7) v. Georgia (6), FPI model favors Georgia.

    Part of the reason for Penn State having a higher win out % is because they are not expected to play in the championship game hence, even though rated behind OSU they have a much higher % of not losing in 3 games than OSU has of losing in 4.

    Yep, that's right. Win-out % and FPI ranking are a separate thing. If Penn State somehow found it's way back into the Big 10 championship game....it's championship % likelihood (another separate FPI measure) would go way up, while it's win-out % would go way down.

    Georgia and Auburn have (by pretty far) the lowest "win-out %" in their Top 10 most likely because of Alabama being ahead on the schedule--definitely for Auburn, likely for UGA in the SECC.

  • Options
    Denmen185Denmen185 Posts: 7,407 ✭✭✭✭✭ Graduate

    @MinnesotaDawg said:

    @Denmen185 said:

    @MinnesotaDawg said:

    @MIghtydawg said:
    That is not what it is saying. It is not ranking the top teams for the season. It is predicting who has the best chance of winning going forward.

    No, that's not right...It's NOT just a prediction of who has the best chance to win going forward. There is actually a different calculation about who has the best chance to win going forward called the "Win Out %" (again, based on it's FPI model). FWIW, Penn State (FPI #3) has the best chance to win out (88.5%), then it's Clemson (FPI #7) with a win-out percentage of 65.7%.

    It actually IS supposed to be a ranking of who the BEST teams are based on theoretical head to head matches. Put another way, the model ranks teams based on how the way it predicts the results of hypothetical games. Example, if asked to predict Georgia (6) v. Washington (4) now, FPI model favors Washington. If asked to predict Clemson (7) v. Georgia (6), FPI model favors Georgia.

    Part of the reason for Penn State having a higher win out % is because they are not expected to play in the championship game hence, even though rated behind OSU they have a much higher % of not losing in 3 games than OSU has of losing in 4.

    Yep, that's right. Win-out % and FPI ranking are a separate thing. If Penn State somehow found it's way back into the Big 10 championship game....it's championship % likelihood (another separate FPI measure) would go way up, while it's win-out % would go way down.

    Georgia and Auburn have (by pretty far) the lowest "win-out %" in their Top 10 most likely because of Alabama being ahead on the schedule--definitely for Auburn, likely for UGA in the SECC.

    It's just a mathematical formula. UGA v AU is basically 50/50 so if nothing changes this week in terms of rating, the chances for both teams will change. If AU wins it will go to 15% (from 7.5) as it one less game and UGA will go to 26.4% (13.2) for the same reason. A UGA win means AU will likely go to 30% (unless Bama also losses) as they will not be in another 50/50 game with UGA and UGA will actually be lower than 26.4 (24 or thereabouts) as they will include a 100% chance of facing Bama (25% or thereabouts) rather than a 30% chance of facing Auburn (50% chance).

    Bottom line is all we can say at this point is that FPI says we have a 13.2% chance of finishing 13-0 but what do they really know.

Sign In or Register to comment.