The CardBoard
#5 in USA Today poll - Printable Version

+- The CardBoard (https://thecardboard.org/board)
+-- Forum: C-House! (https://thecardboard.org/board/forum-4.html)
+--- Forum: The CardBoard (https://thecardboard.org/board/forum-5.html)
+--- Thread: #5 in USA Today poll (/thread-12563.html)

Pages: 1 2 3

Re: #5 in USA Today poll - OutsiderFan - 09-06-2016

The thing I hate about Wilner's ranking approach is he says he votes based on how he thinks teams will finish in the rankings, and then he "adjusts" based on season's results. 

Isn't the purpose to rank the teams based on your perception of their quality when you produce your weekly ranking?  I wish someone would ask him to please rank the teams based on how good you think they are now, not where they will finish in 3 months.

Re: #5 in USA Today poll - FarmBoy - 09-06-2016

My understanding of JW's methodology is that the first ballot, pre-season, is based on where he thinks they will end up. Then the games begin and he tries to be results based. For the first couple of weeks, as he notes, the data sample is very poor, so he partially weights his rankings on where he thinks they'll end up and what they have shown to date. In the case of UFDFG, I think he's saying that he views them as a quality team, so Texas beating them proves their worth, however, he doesn't yet prove that UFDFG is not quality, perhaps in part because they were playing on the road. If Texas is a top 10 team, losing in 2OT to them on the road would theoretically mean your team is just as good as them.

It's fun to criticize his methods/outcomes, but I don't see a pollster doing anything better with transparency or any real better methodology to do a week 1 ranking. Every team with a winning record has done it against teams who haven't won a single game yet!

Re: #5 in USA Today poll - CTcard - 09-06-2016

(09-06-2016, 01:51 PM)OutsiderFan link Wrote:The thing I hate about Wilner's ranking approach is he says he votes based on how he thinks teams will finish in the rankings, and then he "adjusts" based on season's results. 

Isn't the purpose to rank the teams based on your perception of their quality when you produce your weekly ranking?  I wish someone would ask him to please rank the teams based on how good you think they are now, not where they will finish in 3 months.

The AP, which runs these rankings, does not make clear just what the purpose of the ranking is - at least in fine detail. The voters are free to vote on lots of different possible meanings. 

Wilner says he bases his preseason ranking on where he thinks they will end up, though curiously he seems to base that upon the common wisdom of how rankings work rather than his own expressed logic of how rankings should work. So, like many (most?) pollsters he gives breaks to teams with easy schedules.

Once the season starts Wilner tries to immediately base everything upon results, explicitly discounting teams with easy schedules.

Where I actually give Wilner credit is that he intends that his preseason rankings NOT influence his in-season rankings. At least on that score, on intention, he is better than most.

Re: #5 in USA Today poll - BostonCard - 09-06-2016

(09-06-2016, 01:25 PM)CTcard link Wrote:[quote author=Boston Card link=topic=15599.msg170261#msg170261 date=1473189671]
Indeed.  We might call him John "Circular Reasoning" Wilner:

I am a little torn on Mr. Circular Reasoning.

I like the fact that he tries to bring logic to his choices, doesn't stick slavishly to what he did the week before, and wants to primarily base his choices on results. But, he doesn't seem to quite understand what he is doing, opening himself up to both circular reasoning and arbitrary data selection.

The key problem right now is that he wants his ballot to be completely results based despite the fact it is week #1, so only a few teams have played reasonable opponents and even among the others we don't really know how good any team is yet.

So there are all sorts of odd things in there.
I'll choose the one affecting Stanford for now.
Quote:Once we account for the teams with marquee victories, the guesswork truly begins with Ohio State, Michigan, TCU, Washington and Stanford in the No. 9 – 13 spots — all teams that I believe are top shelf but didn’t record substantive results.

Well, one of those teams played a Power 5 opponent that was in a bowl game last year, and generally expected to be better this year. Naturally that team came in at the end of this list. Another struggled with an FCS team; they came in mid-group.

[/quote]

Here's the thing that makes no sense.  Although I don't think it makes sense to throw away your priors (pre-season expectations) after only a game, I can respect trying to make the season rankings based on results and willing to adjust quickly to new data (and WIlner is willing to do this).  The issue is that is not what he is doing.  If you are going to make your rankings based on results and not based on your pre-season expectations, then Stanford beating Kansas State is functionally the same as Texas beating Notre Dame since both teams beat power 5 opponents in the first week.  Yet Wilner discounts Stanford's victory over KState (which is fine) and gives Texas credit for beating a "good" Notre Dame.  But that means that his priors (KState is a middle-of-the-road Big12 team while Notre Dame is a top-10 team) are not being thrown away.  In other words, Wilner is claiming to have one methodology, but his actual methodology is not what he says it is.

BC




Re: #5 in USA Today poll - dabigv13 - 09-06-2016

What would be ideal is to do away with preseason rankings and in season rankings until a after, say, the 4th game. It'll never happen, but that would be the fairest way of doing things.

Re: #5 in USA Today poll - FarmBoy - 09-06-2016

Quote:What would be ideal is to do away with preseason rankings and in season rankings until a after, say, the 4th game. It'll never happen, but that would be the fairest way of doing things.

This. Perhaps the coaches who vote could come together and refuse to vote until the 4th week. The AP voters could never do this because they are so replaceable. But if the coaches got together, a small fraternity like that, maybe they could get an agreement not to participate until teams have shown their abilities on the field of play.

Re: #5 in USA Today poll - stupac2 - 09-06-2016

(09-06-2016, 05:50 PM)dabigv link Wrote:What would be ideal is to do away with preseason rankings and in season rankings until a after, say, the 4th game. It'll never happen, but that would be the fairest way of doing things.

I posted this in some other thread, but there are basically two reasons this won't happen or isn't a good idea:

1) Rankings drive traffic, which is why virtually every sports website in the universe either has its own ranking system or reports on others (or both).
2) Without rankings, people would have to use something to evaluate which early season games are meaningful, and the replacement heuristics could be even worse.

Re: #5 in USA Today poll - CowboyIndian - 09-06-2016

(09-06-2016, 02:09 PM)FarmBoy link Wrote:My understanding of JW's methodology

The words "understanding" and "Wilner methodology" should never be found in the same sentence.

Re: #5 in USA Today poll - BostonCard - 09-06-2016

Also, while starting in week 5 might reduce the influence of pre-season priors, with 15 (I think) serial polls, I think the influence of priors is overstated a bit. Last year, Auburn, Georgia and USC all started the season ranked in the top-10 and ended the season unranked.  Conversely, 10 teams who began the season unranked ended the season ranked.  We started the season ranked in the bottom-half of the top-25, and ended #3.  Nor can you really blame the rankings for not winding up in the playoffs; for that you can blame a brutal game against Northwestern and the hiccup against Oregon.  Win one of those two and we are among the final four.  Moreover, I think reputation would still play a big role in the initial rankings, no matter when you started them.

There is no perfect system, and no amount of tinkering with the computer algorithms or changing the method of the polling is going to change the fact that it is difficult to determine rankings when you have 130 teams playing only 12 - 14 games each, with 8 to 9 of those games being within a closed system (a conference) and relatively few out of conference games matching teams of relatively similar quality.  Then there is the fact that chance is underrated as a factor that affects the results.  Even with the playoffs, I would guess that there is some uncertainty (in a mathematical sense) that the system determined the "best" team.

Instead of trying to get the optimum system, I suggest we enjoy football for what it is, a chance for thousands of like-minded fans to root for a team the best they can, enjoying gorgeous Saturday afternoons or evenings in one of the most beautiful locations in the country, or having unprecedented access to watching their team from far away in a picture of unprecedented quality.  It has never been a better time to be a college football fan than today.

Screw the rankings!

BC

Re: #5 in USA Today poll - JJJ - 09-06-2016

(09-06-2016, 01:29 PM)2006alum link Wrote:[quote author=CTcard link=topic=15599.msg170274#msg170274 date=1473193538]

And not quite apropos of anything, looking over the ESPN list of scores for Week #1 is kind of funny in that there is one team that appears twice. And they put in some serious travel miles. Sure hoped Michigan paid up the big bucks.

Nice spot. It's worth noting Hawaii played in Australia, too, so talk about jet lag. Isn't that something like 12 hour difference? Makes our Northwestern wake-up call from last year seem like nothing by comparison.
[/quote]

UH received $750k for playing Cal down under and $1mln for playing Michigan. The program has been in the red for years so they'll do whatever it takes.

http://www.hawaiinewsnow.com/story/30653755/hawaiical

Re: #5 in USA Today poll - stupac2 - 09-06-2016

(09-06-2016, 06:57 PM)Boston Card link Wrote:There is no perfect system

Sure there is, make the playoff large enough that every good team gets in. Last year proved that 4 is not enough, Stanford was certainly a better team than either MSU or Oklahoma proved to be.

Re: #5 in USA Today poll - CTcard - 09-06-2016

(09-06-2016, 06:14 PM)FarmBoy link Wrote:Perhaps the coaches who vote could come together and refuse to vote ...

I believe they did this years ago, and kept it up ever since.



p.s. the idea of no polls until the 4th, or 6th, or whatever week has been floated since the beginning of polls. I am sure it will start any time now.

Re: #5 in USA Today poll - BostonCard - 09-06-2016

(09-06-2016, 08:12 PM)stupac2 link Wrote:[quote author=Boston Card link=topic=15599.msg170327#msg170327 date=1473213456]There is no perfect system

Sure there is, make the playoff large enough that every good team gets in. Last year proved that 4 is not enough, Stanford was certainly a better team than either MSU or Oklahoma proved to be.
[/quote]

A) While we could make a case that we were better than Oklahoma or MSU, I don't think we can make a case that we were better than Alabama or Clemson, so I don't think there is a great injustice in not getting into the playoff.
B) If you are arguing that once in the playoff anything can happen, then that undermines your assertion that a larger playoff will be better to determine the "best" team.

If you draw the line at 8, then the 9th ranked team will argue that they were better than the 7th or 8th and should have been included.  If you draw the line at 16, the 17th best team will complain.  If at 32... you get the idea.  Before you know it, you have 64 teams, and some are basically average.

You are ultimately facing a dilemma.  The more teams you let in, the greater the chance that the "best team" gets upset along the way.  The fewer teams you let in, the greater the chance that the "best team" doesn't get in.

I'm all for just accepting that the flaws inherent in any system.

BC

Re: #5 in USA Today poll - TheFarm07 - 09-06-2016

(09-06-2016, 09:15 PM)Boston Card link Wrote:[quote author=stupac2 link=topic=15599.msg170338#msg170338 date=1473217958]
[quote author=Boston Card link=topic=15599.msg170327#msg170327 date=1473213456]There is no perfect system

Sure there is, make the playoff large enough that every good team gets in. Last year proved that 4 is not enough, Stanford was certainly a better team than either MSU or Oklahoma proved to be.
[/quote]

A) While we could make a case that we were better than Oklahoma or MSU, I don't think we can make a case that we were better than Alabama or Clemson, so I don't think there is a great injustice in not getting into the playoff.
B) If you are arguing that once in the playoff anything can happen, then that undermines your assertion that a larger playoff will be better to determine the "best" team.

If you draw the line at 8, then the 9th ranked team will argue that they were better than the 7th or 8th and should have been included.  If you draw the line at 16, the 17th best team will complain.  If at 32... you get the idea.  Before you know it, you have 64 teams, and some are basically average.

You are ultimately facing a dilemma.  The more teams you let in, the greater the chance that the "best team" gets upset along the way.  The fewer teams you let in, the greater the chance that the "best team" doesn't get in.

I'm all for just accepting that the flaws inherent in any system.

BC
[/quote]

True, there will always be a team that gets left out who will complain that they are better than the last team let in no matter the size of the playoff field, but if you expand the field too large, the nth team that gets left out doesn't really have a legitimate argument that they have a decent chance of winning the whole tournament/playoffs, while if you expand the field slightly (from 4 to 6 or 8, for example), you can argue that the 2 or 4 additional teams are at or near the level of the top 4 teams and deserve a chance to prove their worth on the field, rather than be penalized by an untimely loss or a different schedule.

Re: #5 in USA Today poll - BostonCard - 09-06-2016

(09-06-2016, 10:41 PM)TheFarm07 link Wrote:True, there will always be a team that gets left out who will complain that they are better than the last team let in no matter the size of the playoff field, but if you expand the field too large, the nth team that gets left out doesn't really have a legitimate argument that they have a decent chance of winning the whole tournament/playoffs, while if you expand the field slightly (from 4 to 6 or 8, for example), you can argue that the 2 or 4 additional teams are at or near the level of the top 4 teams and deserve a chance to prove their worth on the field, rather than be penalized by an untimely loss or a different schedule.

That is, in essence, the dilemma, I mentioned.  If you allow more teams into the playoff, it increases the chances that the "best team" gets into the playoffs, since such a team might have had an untimely loss in the regular season.  But it also decreases the chances that the "best team" wins the championship, since it is easier for a team to be upset if it has to play three times than if it has to play twice, since teams can have untimely losses in the playoffs too.  Any system requires tradeoffs, and I am not convinced that it the "best team" will be left out of a four team playoff any more often than it being upset when playing one more playoff game.

To me the better argument for a playoff involves giving every P5 conference champion a seat, plus the best non-P5 team (such as BYU or Houston, or, ugh, UDFG) if you stop at 6, and two wildcards if you stop at eight.  There is a nice aura of inclusiveness there, but I bet it doesn't get any closer to crowning the best team as champion than a system where the best 4 P5 conference champions make the playoff.

On the other hand it means another game, and another chance for injury, for the players.  Increasing the number of games from 15 (12 regular season games plus a conference championship plus two playoff games) to 16 might not seem like much, but remember that not too long ago, teams were playing 12 games total (11 regular season games, 1 bowl game).  At some point we have to remember these guys are supposed to be students too, not professionals.

Last year was about the worst possible scenario for Stanford vis a vis the playoffs, being the odd P5 team out.  And yet I was not bothered one bit.  Loved the Rose Bowl and loved kicking the crap out of Iowa.  I don't think the playoffs would have been a better experience, to be honest.

BC

Re: #5 in USA Today poll - stupac2 - 09-07-2016

The thing is that trying to determine the "best" team is an actively pointless goal. It's so poorly defined that I don't understand why ostensibly smart people even discuss it. There is a reason that literally every other team sport I'm aware of determines a championship based on:

1) Group teams into geographical clusters.
2) The winners (determined somehow) out of those clusters advance to the tournament.
3) The best records of the non-winners (again, determined somehow) also advance to the tournament.
4) Seeding is done either by records or by informed panels producing a consensus.
5) There's a defined path to winning the tournament.

College football is the only sport that doesn't do that and I find it baffling. I find it even more baffling that people defend this approach, especially fans of the team that has been completely screwed over by it at least twice in the past 8 years.

Finall, what I meant by us being better than Oklahoma/MSU is that we wouldn't have put up clunkers of games against Alabama/Clemson. Would we have beaten the latter two? Maybe! It's impossible to know without playing the game, which is all that I want the opportunity to do. Could we have beaten Auburn in 2010? I think so! Alabama in 2011/2012? Probably not but who knows? I want to see those games!

Re: #5 in USA Today poll - TheFarm07 - 09-07-2016

(09-07-2016, 08:29 AM)stupac2 link Wrote:Finall, what I meant by us being better than Oklahoma/MSU is that we wouldn't have put up clunkers of games against Alabama/Clemson. Would we have beaten the latter two? Maybe! It's impossible to know without playing the game, which is all that I want the opportunity to do. Could we have beaten Auburn in 2010? I think so! Alabama in 2011/2012? Probably not but who knows? I want to see those games!

I agree, while watching us demolish Iowa in the Rose Bowl was very pleasurable, I was still miffed that we didn't have a chance to prove ourselves against Alabama, Clemson, etc. Since we are never going to play them in the regular season, I would like the opportunity to face them and have an opportunity to stuff their East Coast bias/SEC arrogance in their faces. Of course, we run the risk of falling flat on our faces, but I think it's still worth it to prove to the rest of the country that we are a good football team.

Re: #5 in USA Today poll - jacketree - 09-07-2016

Yeah, the point of any playoff or championship is not to determine the "best team", but the champion.

I have a fool-proof proof plan to reorganize "big boy" college football into eight ten-team conferences that finishes with an eight-team playoff among only the round-robin conference champions.  Eleven or twelve game regular seasons; no conference title games. Relegation a la English soccer features prominently.  Sorry Wake Forest; until you can pull a Leicester City, you're out. I'd love to think that plan is sitting on an important desk in a bright red folder titled "LET'S DO THIS!!" However, whenever I float the idea people think I'm nuts.

But if we go to a 12-team playoff with some first round byes, then anyone could be the National Champion in any given year.  All on the field.  Polls and committees wouldn't mean squat other than maybe to seed and determine home fields.

And if you really are the best team, you won't lose any playoff games right?  (I'm looking at you 2007 New England Patriots!)

Re: #5 in USA Today poll - OutsiderFan - 09-07-2016

(09-07-2016, 01:01 PM)jacketree link Wrote:Yeah, the point of any playoff or championship is not to determine the "best team", but the champion.

I have a fool-proof proof plan to reorganize "big boy" college football into eight ten-team conferences that finishes with an eight-team playoff among only the round-robin conference champions.  Eleven or twelve game regular seasons; no conference title games. Relegation a la English soccer features prominently.  Sorry Wake Forest; until you can pull a Leicester City, you're out. I'd love to think that plan is sitting on an important desk in a bright red folder titled "LET'S DO THIS!!" However, whenever I float the idea people think I'm nuts.

But if we go to a 12-team playoff with some first round byes, then anyone could be the National Champion in any given year.  All on the field.  Polls and committees wouldn't mean squat other than maybe to seed and determine home fields.

And if you really are the best team, you won't lose any playoff games right?  (I'm looking at you 2007 New England Patriots!)

Create a website to explain and promote your idea.  Just make sure it has high production value and is easily understood.  Maybe partner with the Dr. Pepper pitch man?

Re: #5 in USA Today poll - ColoradoTree - 09-07-2016

(09-06-2016, 11:16 PM)Boston Card link Wrote:Last year was about the worst possible scenario for Stanford vis a vis the playoffs, being the odd P5 team out.  And yet I was not bothered one bit.  Loved the Rose Bowl and loved kicking the crap out of Iowa.  I don't think the playoffs would have been a better experience, to be honest.

BC

I agree with this statement, though I'd also prefer a playoff format that ensures the P5 champions all get in, whether that's 6 (incl. 1 wild card) or 8 (incl. 3 wild cards). 

But I didn't feel cheated last year.  Not by any stretch.  I thought we would have beaten Michigan State and Oklahoma, and it would have been an exciting matchup versus Alabama or Clemson.  Of course I wanted to be in the mix so our guys could prove themselves against the other top teams.  But we could have left no doubt by beating Oregon.  We were in control of our own destiny at that point, I believe, even with the Northwestern loss, and would have squeaked in ahead of probably Oklahoma.  The Rose Bowl was a perfect capstone to a great, if imperfect, season. 

Now, if Stanford misses out this year as a 1-loss champ to another 1-loss champ, or as a 2-loss champ to another 2-loss champ, then I'll be ticked.  That's why I want Clemson/FSU to fail at every turn, and all the Big 12 and Big Ten teams to beat each other senseless, so we remove as many roadblocks to that national test as possible.  And for every Pac-12 team to win every non-conference game.  And for ND to win every game they play against everyone but us (except maybe SC). 

That's why I disagree with those who say it's too early to worry about this stuff.  In some games, like Ole Miss-FSU, that might be true.  It's hard to know which team we wish will have lost when we're at the end of the season, but to me, it's not too early for generalities, like rooting for the Pac-12 to win every possible non-conference game. Where any team from the ACC, SEC, Big Ten, or Big 12 plays a non-P5 team, I'm rooting for the non-P5 team (unless it's Houston, who is recognized as a possible playoff threat and could well go undefeated now because of the weakness of its remaining schedule).  I figured Ole Miss would take a loss or two in conference play pretty readily, while I thought FSU's only threat was Clemson, so I wanted FSU to take a loss there.  I don't think Wisconsin will make its way through its schedule unscathed, so why not root for them to knock off LSU, which was a popular playoff pick?  I figured Auburn was more likely than Clemson to take a loss or two in conference play, so I wanted Clemson to lose that game.Â