What should CFB polls actually reflect?

Submitted by crg on
Something that has bothered me for many years with respect to college football: the reliance on polls. Essentially glorified popularity contests, the polls have far too much influence on the sport these days IMO. To make matters worse, the major polls (AP and Coaches) tend to give preference to teams that are already ranked. There appears to be a mindset of "what did this team do to justify dropping their ranking" as opposed to " what did this team do to deserve keeping their ranking". I believe a part of this problem is that the poll voters vote on how they believe a team "should" be as opposed to what they have actually done (to be fair, this lessens towards the end if the season, but not entirely). Prime examples just from this week's AP: a 2-2 Ole Miss ranked two spots ahead of a 4-0 Utah and a 3-0 OSU ranked 7 spots above a 4-0 Wisconsin (with better quality wins), although more ridiculous rankings have occurred in prior years.

Ranting aside, here is a question for the board: how should the polls be improved so that they are 1) more fair and more importantly 2) more useful. Or, should we just scrap the polls all together and just have one ranked team at the end (#1), because in the end, there can be only one.
MGoBoard
Football

MichiganStudent

September 27th, 2016 at 10:07 PM ^

Completely agree on the coaches poll. Half (if not more) of the time it's not even the coaches that are ranking these teams as its their assistants or staff members.

I'd rather have the Doctor Pepper guy that invented the College Football Playoff rank the teams.



Sent from MGoBlog HD for iPhone & iPad

Joined: 08/27/2008

MGoPoints: 7223

Mr. Yost

September 28th, 2016 at 12:18 PM ^

...but Houston and OSU beat Oklahoma who looks like shit. Not a highly ranked powerhouse. Not really a quality win.

I mean is an Oklahoma win any better than Virginia Tech or North Carolina or Maryland or West Virginia at this point? No. But teams skyrocket for being a top 5/top 10 team that isn't a top 5 or top 10 team.

Iowa was flirting with the top 10...they lose to North Dakota St. and beat Rutgers 14-7. Not a quality win for anyone right now.

We all know about LSU, USC and Notre Dame.

Meanwhile, you have other teams who've beaten teams who are BETTER than their ranking when they played. Michigan shouldn't get too much for being Colorado, but whoa...looks like that was a fairly decent win rather than a snoozer versus a cupake we should've dominated from the opening kick.

There are a number of teams in that boat as well.

...so now that we have more information, let's lay it out and really see who's got quality wins or bad losses.

Joined: 07/25/2011

MGoPoints: -9898220

TrueBlue2003

September 27th, 2016 at 11:16 PM ^

the polls already mean nothing.  The CFP rankings are what matter and the committee has already shown the ability to disregard biases towards expectations and name recognition and rank teams based on the right mix of quality and performance.  The committee has actually nailed the first two years.

The Coaches and AP poll are fodder for discussions and debate and nothing more.  There does not need to a be a right or wrong way to do it.  They're for entertainment purposes only.

Joined: 01/05/2010

MGoPoints: 38075

Kevin13

September 28th, 2016 at 11:52 AM ^

4 games should be played and first polls should be released in Octobber. Let the teams plays some games so everyone can get a good idea on just how good these teams are, it's too tough to rank a team based off of what they did last year and expectations.

 

The coaches poll is also a joke. I think most votes are by support staff working for the coaches as they don't have the time, nor do they have any idea on just how good these other teams are, without watching them and they don't spend the time doing that.

Joined: 01/26/2016

MGoPoints: 16040

Red is Blue

September 27th, 2016 at 9:46 PM ^

The polls are too dependent on the number of losses without enough consideration of strength of schedule. A school going 9-3 having lost to three top teams and beating several other top teams will almost certainly be ranked below a tEames that goes 11-1 against a cupcake schedule.

Joined: 05/13/2010

MGoPoints: 23889

lhglrkwg

September 28th, 2016 at 1:25 AM ^

Non-conference schedules have gotten significantly better recently apparently as a hope of improving their SOS and CFP chances. Well if a 12-0 Baylor team who schedules absolutely no one in the non-conference makes it in over another 11-1 Power 5 champion who play two top 25 teams in the non-con, you will quickly see schools start to see that they should just schedule the cupcakes again.

The CFP committee really needs to use SOS or it'll be back to a MAC team, an FCS team, and Notre Dame for us which would suck

Joined: 07/03/2008

MGoPoints: 87925

sheepdog

September 27th, 2016 at 9:48 PM ^

Polls are really more like power rankings. If you beat a good team or look impressive, it helps you. If you win and look sloppy it can hurt. True polling needs to take into account more SOS and other metrics, not just votes from coaching staffs that more than likely don't even watch the games.



Sent from MGoBlog HD for iPhone & iPad

Joined: 06/15/2011

MGoPoints: 3277

TrueBlue2003

September 28th, 2016 at 12:10 AM ^

argue SoS isn't accounted for but complain about Ole Miss being ranked ahead of Utah when it's exactly their SoS that has them ranked ahead (as they definitely should be).  They've played a BRUTAL schedule so far, losing to two (legitimate) top 12 teams in close-ish games and DESTROYING a borderline top 25 team.  They're ranked right about where they should be slightly closer to top 12 than 25th.  Utah hasn't really played anyone with best wins being very close home wins against 1-3 BYU and 1-3 USC (both of whom are probably top 40ish teams but that still just makes Utah slightly better than 40th probably).

FWIW, every unbiased, numerical rating that takes into account SoS and performance rates Ole Miss wayyyyyy higher than Utah with Ole Miss usually around 10th and Utah usually in the 30s.

Joined: 01/05/2010

MGoPoints: 38075

ElBictors

September 28th, 2016 at 12:37 AM ^

"Strength of Schedule" should also consider the outcome.  Your interpretation is precisely how the SEC Echo Chamber operates.  You assert Ole Miss is better than Utah because of polling and factors aside from actual game outcomes.

 

You don't know Utah couldnt beat Ole Miss

Joined: 03/27/2015

MGoPoints: 8446

ElBictors

September 28th, 2016 at 11:04 AM ^

Nope. You just can't rely on pollsters being honest. Houston at 11-1 is not better than a 10-2 MICHIGAN in any way based on SOS ...so an 11-1 MICHIGAN is far more tested and tried than that same 11-1 Houston. The problem is the BS the SEC honks have made people believe over the past decade - that somehow if you barely beat Vandy or Kentucky, it's a big win ...just cuz

Joined: 03/27/2015

MGoPoints: 8446

ElBictors

September 27th, 2016 at 9:52 PM ^

My favorite bet in Vegas is the CFB National Title bets where you start with Alabama at 3:2 or whatever and then go down the line to an Oregon State being 100:1 and then the "Rest of the Field" bet where you can take any program outside the Top 50 or so with lines.

 

That has to be the biggest sucker bet ever.

 

Joined: 03/27/2015

MGoPoints: 8446

ElBictors

September 28th, 2016 at 12:45 AM ^

I wasn't posting in reply to a thread about how to bet online, dick.  Because I know how to do that.

I was replying precisely to the notion that ranking a Top 25 and other polling precludes about 50 programs from ever winning the national title.  And any fool willing to bet on the "field" outside of the Top 50 programs is a sucker.

But please, continue with the point you were making or ...just quit drinking and go to bed.

 

dick.

 

Joined: 03/27/2015

MGoPoints: 8446

mbrummer

September 27th, 2016 at 10:00 PM ^

Polls were designed to sell papers and bring viewers to games on television.

Why would anyone in the West or South want to watch a Michigan - Notre Dame game?  But you put #3 Michigan and #5 Notre Dame and it portends to mean something.

Currently, it is the same.  Let say Ole Miss plays South Carolina.  Would you watch?  Now if I put #19  Ole Miss and #12 South Carolina, it becomes much more interesting  to a viewer outside those fanbases.

They are callled the AP, Associated Press,and USA Today Coaches' Poll.   They sell papers. If you think of sports section as an although symbiotic free advertisement for the sports teams they cover, it becomes much easier to dismiss. They are symbiotic because they sell the sports teams they cover,and the sports pages used to help selll newspapers.

I'm sure people will suggest no polls until October or November.  This isn't realisitc because it flies in the face of the exact reason of their existence.  

Since, they are inherently flawed, there is no fixing them.

Joined: 07/29/2008

MGoPoints: 7872

crg

September 27th, 2016 at 10:16 PM ^

Historically, this assessment is completely correct. However, with the implementation of the BCS rankings (and now the CFP), the major polls are taken into consideration upon a ranking system that actually means something (less now in CFP than BCS, but still present). Therefore, the flaws in those polls can propagate into real consequences for some teams.

Joined: 04/30/2015

MGoPoints: 75054

Hail Harbo

September 27th, 2016 at 10:23 PM ^

Polls don't exist for the convenience of determining who's the best and who should play whom, they are a private venture designed to sell a product.  In days of yore it was the AP and UPI wire services that were selling the polls to the various newspapers.  And why did the newspapers want those polls?  Why, to sell newspapers.  Polls used to be published on what was typically a slow news day, Tuesday. Why?  To make money.

BTW, polls used to only include the top twenty.  40 years ago along came USA Today with its own poll of 25, and 25 it's been ever since.  Why did USA Today go with 25 instead of the traditional 20?  You probably guessed it, to make more MONEY.

Joined: 06/18/2015

MGoPoints: 13428

mbrummer

September 27th, 2016 at 10:41 PM ^

The polls weren't and aren't  going to change though.  BCS was a disaster, they changed their formula every year.

I believe the AP dropped out of it, because at least the press knew it was flawed.

I  don't think the CFP takes into account the polls.  Lets me honest, only the last one matters.  If they do, the final polls are usually decently legitimate.  

The basketball committee has been ignoring the polls for years.

Joined: 07/29/2008

MGoPoints: 7872

crg

September 28th, 2016 at 9:06 AM ^

But we don't "know" that. The CFP operates behind closed doors and is a black box. Yes, they provide a few weeks of rankings before the end and maybe a few statements to defend their positions, but ultimately they can decide to use whatever criteria they want with any weighting they want.

Joined: 04/30/2015

MGoPoints: 75054

DGM06

September 27th, 2016 at 10:05 PM ^

Starting the polls later in the season would help. No ranking that is used for determining playoff participants should ever depend on preseason expectations. The early season polls are only helpful in boosting TV ratings; would a Louisville/Clemson game have national appeal without those small numbers next to each team's name?

Joined: 08/03/2010

MGoPoints: 655

Lionsfan

September 27th, 2016 at 10:17 PM ^

I have no problems with how the polls are currently set up. They only exist because of football junkies, and if they suddenly stopped next year until Week X, the internet would be flooded with everybody trying to rank teams their own way.

And honestly, they're just numbers anyways. The only ones that matter are the CFP rankings, which don't start until Week 9 anyways. And the CFP committee members haven't shown they care about how the polls when it's all said and done

Joined: 06/03/2011

MGoPoints: 2528

DualThreat

September 27th, 2016 at 10:18 PM ^

... is that they try to merge two distinct concepts into one product:

> Which team is more deserving

vs

> Which team is probably actually better

If you divorce these two concepts into seperate polls, you can actually have more meaningful debate and discussion.

For example, here is my current top 5 "deserving" poll, based soley on perceived quality (ranking) of teams beaten:

1 - Alabama (as defending champs until they lose)

2 - Wisconsin

3 - Louisville

4 - Texas A&M

5 - Ohio State

 

And here is my top 5 "actually better" poll.  This is basically a who would win on a neutral field poll.  Each team higher would beat a team lower right now.  Personal opinion only of course.

1 - Ohio State

2 - Alabama

3 - Clemson

4 - Louisville

5 - Florida State

(Michigan would be #6 in this poll - again my opinion only.)

 

FWIW - When you have single elimination playoffs, more often then not you're crowning a "most deserving" champion, not necessarily a "who is actually best" champion.  Think the NCAA basketball tournament as a prime example.  Coincindentily, the same sport (basketball) also provides one of the best examples of a "who is actually best" champion - the NBA playoffs where each matchup is a series of games.

 

Joined: 05/04/2012

MGoPoints: 781

Lan Jiao

September 27th, 2016 at 11:13 PM ^

And different voters cast their ballots using both of these methods. My guess is that the Coaches would tend to lean more toward the "Who is Best" method and the AP leans toward "Most Deserving," but I have no way of proving that.

Most Deserving - you hit it on the head.

Who is Best???

1a. Death Star Tuscaloosa

1b. Death Star Columbus

3. Loovull

4. M

5. Stanford/Clemson are interchangeable to me right now. Clemson with a high upside.

Joined: 08/30/2016

MGoPoints: 193

bronxblue

September 27th, 2016 at 10:22 PM ^

They shouldn't come out before October for starters, but I also think they should reflect which teams would win more times than not against each other.  Doesn't mean it's a given, but the #4 team in the country should be able to beat the #8 team on a neutral field, and so on down the line.  The bigger the gap, the more frequent you expect the lower-seeded team to lose, but that's about it.  And if they are equal teams, rank them as such.  We've seen that where two teams receive the same number of #15 votes, and they just tie for it.  That's fine.

Joined: 11/22/2008

MGoPoints: 171617