Alabama the #4 All Time CFB Program (according to AP)

[TideEngineer08]

Guys, this is nothing more than the AP patting itself on the back for it's own flawed ranking system. The AP has been made irrelevant and in fact made itself irrelevant during the BCS era. It no longer matters. And a major reason is because of the data you'll find reading through this ranking.

Alabama was always seriously under ranked, and usually Notre Dame and Ohio State were over ranked. In other words, the AP had it's favorites and was/is biased.

 

[BamaMoon]

It's just an example of how statistics can be used to prove anything.

As someone said, whoever did it probably used the metrics they used to be able to reach a preconceived notion.

 

[dayhiker]

Any formula that values a #2 ranking the same as it values a #25 ranking might as well be throwing darts. Furthermore, the difference between #25 and #26 (which despite not assigning a ranking can be tracked) should not be greater than the difference between #2 and #25, obviously, but it is. For the purpose of this stupid ranking, being #1 or being #25 at any point other than the final poll is actually as significant as the difference between #25 and #26 (which is nonsense). Think about that for a little while.

Now, how should anyone that's not a complete moron come up with a ranking?
1: They would only use the final poll.
2: They would assign point values to each place in a poll, 25 points for #1, 1 point for #25 ranking.
3: They would assign additional point values for each #1 vote. I'd have to look at the number of voters to come up with that, but it would be similar to the point bonus they already gave.
4: They would also assign points for 26-29 if applicable. .8, .6, .4 and .2 points respectively.

Now that's far from perfect, obviously. But it's a whole two minutes devoted to the topic and still a far better ranking system than the absurdity they came up with.


No, that isn't how they did it. They gave Ohio State points 105 times, for being ranked #1 in the poll. Ohio State didn't win 105 titles, but they got the point bonus for each poll they had a #1 ranking in, even if it ended up meaning nothing. They actually valued being ranked #1 throughout the year more than winning the championship, which is all kind of stupid.

Just look at this part of it. Alabama won 10 AP titles and Ohio State won 5. Yet, because this formula is made by morons, Ohio State (bonus championship points aside) actually got 62 more points for being ranked #1 than Alabama! That makes no sense, Alabama wins twice as many titles, but Ohio State holds the record for meaningless #1 rankings so they rack up points anyway?

I'll have to reread it. I thought they gave 1 point for each week you were ranked, 2 points for each week you were ranked #1, and 10 points for each AP title.

EDIT: I was correct. Here's the cut and paste from teh article. I was saying that the title is weighted 5x higher than being #1 in a poll. I didn't comment on the weight versus a typical ranking because it wasn't relevant to the question that I was asked.

To determine the all-time Top 25, the AP formula counted poll appearances (one point) to mark consistency, No. 1 rankings (two points) to acknowledge elite programs and gave a bonus for AP championships (10 points).

 

[dayhiker]

Any formula that values a #2 ranking the same as it values a #25 ranking might as well be throwing darts. Furthermore, the difference between #25 and #26 (which despite not assigning a ranking can be tracked) should not be greater than the difference between #2 and #25, obviously, but it is. For the purpose of this stupid ranking, being #1 or being #25 at any point other than the final poll is actually as significant as the difference between #25 and #26 (which is nonsense). Think about that for a little while.

Now, how should anyone that's not a complete moron come up with a ranking?
1: They would only use the final poll.
2: They would assign point values to each place in a poll, 25 points for #1, 1 point for #25 ranking.
3: They would assign additional point values for each #1 vote. I'd have to look at the number of voters to come up with that, but it would be similar to the point bonus they already gave.
4: They would also assign points for 26-29 if applicable. .8, .6, .4 and .2 points respectively.

Now that's far from perfect, obviously. But it's a whole two minutes devoted to the topic and still a far better ranking system than the absurdity they came up with.


No, that isn't how they did it. They gave Ohio State points 105 times, for being ranked #1 in the poll. Ohio State didn't win 105 titles, but they got the point bonus for each poll they had a #1 ranking in, even if it ended up meaning nothing. They actually valued being ranked #1 throughout the year more than winning the championship, which is all kind of stupid.

Just look at this part of it. Alabama won 10 AP titles and Ohio State won 5. Yet, because this formula is made by morons, Ohio State (bonus championship points aside) actually got 62 more points for being ranked #1 than Alabama! That makes no sense, Alabama wins twice as many titles, but Ohio State holds the record for meaningless #1 rankings so they rack up points anyway?

PS, note I said I didn't agree with it, I was just commenting on how it worked. Because someone weighs in on the issue, doesn't mean that they agree with it.

 

[LA4Bama]

so if you were ranked number one for nine weeks but lost national championship you got 18 pts? But if you were ranked number two for nine weeks and won national championship you got 10 pts?

I think you forgot, there's one point for every non-first-place appearance in the poll, so on your scenario, both teams gets 19 pts
(9*2)+1
(9*1)+10

 

[NoNC4Tubs]

Was the author an OSU fan?

Sent from the International Space Station.

 

[BamaNation]

Two things:

1) given that the AP poll and those who rely on the AP poll think college football started around 1940, I discount anything that relies on them for historical rankings.

2) as I told an Ole Miss Rebel Black Bear buddy of mine this week, I will take a yearly drubbing at the hands of Ole Miss if the result is the same as the last two years and especially last year. Results are what matter. Most folks don't understand our logic.

 

[B1GTide]

I think that this proves how meaningless weekly polls are, right?

 

[dvldog]

Did they give points for coaches punching players?


Sent from my iPhone using Tapatalk

 

[KrAzY3]

I was saying that the title is weighted 5x higher than being #1 in a poll.

You are correct and I suppose I was overstating things a bit, since one can assume the eventual champion is going to be ranked throughout the season (though some have started unranked I believe), but they did put a system in place in which it was possible to get less points even if you won a championship. I was overemphasizing the 2 points for #1 thing though, since technically it's 1 point for #1 then a bonus point for being the poll because reasons.

I wasn't going after you specifically, just pointing out they far overvalue being ranked #1 at meaningless points in the season. They had no business counting the rankings at all until the final poll in the first place, if they had any intent in actually ranking the best teams. This does create a scenario in which the eventual champion is trailing by several points in their silly ranking, and somehow finishes with less points for a championship season. That's stupid, but I was stating things as though it was 2 points vs 0 which while hypothetically possible isn't necessarily accurate. I can't really accurately calculate how much they might have underscored a champion or overscored a non-champion any particular year without going through year by year polls, so it was a misstatement to arrive to any specific conclusions without doing that.

I think that this proves how meaningless weekly polls are, right?

I am not one who thinks weekly polls are meaningless, I feel they are important. However, I think of it as something that is fine tuned as the season goes on. As the season progresses, SoS and rankings start to mean more and more, as there's more and more data to base those on. So, I don't feel like week one poll is meaningless, but it's basically just an educated guess. No way on earth it should hold the same value as the final poll, that's just silly as hell (and for anything short of the champion they made them both count the same).

 

[RTR91]

I think that this proves how meaningless weekly polls are, right?

Only if you emphasize them more than the actual title.


Sent from my iPhone using Tapatalk

 

[deliveryman35]

Per Wikipedia, our first appearance at the AP no. 1 slot was not until 1961 --while the barn's was in 1957. That really surprised me to say the least.

I'd agree we may not have been the most consistently superior team of all time, but I'll take legitimate, consensus NC's any day of the week over that accomplishment--and so would the other three programs ahead of us on that list, you can rest assured of that.

 

[TideEngineer08]

I think that this proves how meaningless weekly polls are, right?

Especially those of a body of sports writers who where (still are) very biased against the South.

 

[JBama_in_PCOLA]

Any formula that values a #2 ranking the same as it values a #25 ranking might as well be throwing darts. Furthermore, the difference between #25 and #26 (which despite not assigning a ranking can be tracked) should not be greater than the difference between #2 and #25, obviously, but it is. For the purpose of this stupid ranking, being #1 or being #25 at any point other than the final poll is actually as significant as the difference between #25 and #26 (which is nonsense). Think about that for a little while.

Now, how should anyone that's not a complete moron come up with a ranking?
1: They would only use the final poll.
2: They would assign point values to each place in a poll, 25 points for #1, 1 point for #25 ranking.
3: They would assign additional point values for each #1 vote. I'd have to look at the number of voters to come up with that, but it would be similar to the point bonus they already gave.
4: They would also assign points for 26-29 if applicable. .8, .6, .4 and .2 points respectively.

Now that's far from perfect, obviously. But it's a whole two minutes devoted to the topic and still a far better ranking system than the absurdity they came up with.


No, that isn't how they did it. They gave Ohio State points 105 times, for being ranked #1 in the poll. Ohio State didn't win 105 titles, but they got the point bonus for each poll they had a #1 ranking in, even if it ended up meaning nothing. They actually valued being ranked #1 throughout the year more than winning the championship, which is all kind of stupid.

Just look at this part of it. Alabama won 10 AP titles and Ohio State won 5. Yet, because this formula is made by morons, Ohio State (bonus championship points aside) actually got 62 more points for being ranked #1 than Alabama! That makes no sense, Alabama wins twice as many titles, but Ohio State holds the record for meaningless #1 rankings so they rack up points anyway?

Yeah, I don't get it at all.


Sent from my iPhone 6+ using Tapatalk

 

[JBama_in_PCOLA]

Anything to show Alabama in a bad light I guess.


Sent from my iPhone 6+ using Tapatalk

 

[selmaborntidefan]

I'm gonna be putting up several posts here, partly thinking out loud and partly making an argument.

If you had asked me in 2010 who the greatest college football program of all-time was, I would have said Notre Dame. Seriously. In fact, it has long been considered pretty much indisputable that the TOP TWO programs of all-time were the Irish and the Tide. However, I have argued since January 7, 2013, that Alabama now owns that designation.

Let's agree (because of the whole controversy thing) to set aside pre-1936 ONLY FOR COUNTING TITLES. I'll agree to that, okay (just for now). Let's consider what an all-time team would have to have to be considered the greatest program of all-time. Keep in mind that I have no idea how ALL of the details will play out, I just know some of them.

1) Wins

No comment necessary here. An all-time great program is going to have a lot of wins. Michigan is number one all-time with 915. They're followed by Texas, Notre Dame, Nebraska, Ohio State, Oklahoma, Alabama. I'm assuming this is with our 'vacated wins' subtracted.

So four of the top five selected are here. Notre Dame has 882, Alabama and OU 850, and everyone else is in between. Not a substantial difference.

2) National Championships

A lot of wins is going to lead to a lot of national titles. What's the post-1936 count?

Alabama 10 (note that we don't count 1973, either)
Notre Dame 8
Oklahoma 7
USC 5
Ohio State 5

3) National Championships In Different Decades

So that we're talking long term dominance, a team should have titles across multiple decades, demonstrating in particular the changes that occur and adapting to them.

Alabama - 60s, 70s, 90s, 00s, 10s
Notre Dame - 40s, 50s, 60s, 70s, 80s
Ohio State - 40s, 50s, 60s, 00s, 10s
Oklahoma - 50s, 70s, 80s, 00s
USC - 60s, 70s, 00s

THREE of the five schools have won titles in FIVE different decades.

4) Bowl Wins

The immediate objection is that Notre Dame didn't play bowl games prior to 1970 (for the most part) so how can we count these. The problem is that bowl games - even while considered exhibitions - virtually always featured match-ups of highly ranked and well thought of teams. Ohio State also gets punished just a bit because of the Big Ten 'no repeat' rule.

But given the plethora of bowl games today, I think you have to limit these as well. It's preposterous to put the modern-day Music City Bowl on par with the 1964 Sugar Bowl. So I would only count bowl appearances in the Big Four prior to 1987 and then include the Fiesta Bowl, starting with Miami-Penn State.




Problems With The Weekly Counting

There are some immediate problems with the weekly counting of AP top-ranked teams. Here's a rather obvious one - how many times were these teams PRE-SEASON number ones?

The AP didn't release a preseason poll until 1950. Notre Dame was ranked there first in 1950....and the Irish then proceeded to go 4-4-1. But they still get points under the 'AP only' system.


Alabama was pre-season #1 four times - 1966, 1978, 2010, and 2013. And 1966 shows all that's wrong with using AP only ratings as Alabama dropped without doing anything wrong in September.....but Notre Dame did NOT drop after their playing for a tie against Michigan State. And Alabama gets a bunch of points for 12 straight polls at the top in 2013.......only to not even the win the division.

Oklahoma was pre-season #1 TEN times.....meaning they pocketed easy points in 1955-56-57-74-75-77-85-86-87-03-11

Notre Dame (1950, 1953, 1954, 1967, 1971)

Ohio State (1958, 1962, 1969, 1970, 1980, 1998, 2006, 2015)

USC (1963, 1973, 1979, 2004, 2005, 2007, 2012)

So OU starts at number one and gets points. They get to run up a BUNCH of points in a year like 2003......when they get blown out in the Big 12 title game and still get a shot at the whole bowl of wax....and lose again. Same with Ohio State in 2006, where they got all the way to the title game against Florida......and got clobbered.


So because the AP overrated a team like Notre Dame in 1950 or (to a much lesser extent) USC in 2007, it doesn't matter. They get points for it.

 

[Bamabuzzard]

My guess is a lot of the reasoning behind putting Bama #4 is to bring all the Bama fans out of the woodwork to get all ticked off, click on the article, go to message boards, talk about it for days on end. Bama moves the needle. Plain and simple. This had a purpose no doubt.

We all know, even the idiots who put the list together, that championships is the measuring stick.

"YOU PLAY. TO WIN. THE GAME!!!" And Bama has won the game. It's not even debatable at this point.

 

[CajunCrimson]

Every time something like this comes out, my first thought is...

clutter.....

 

[BamaInBham]

2015 point total:

Clemson 21
OSU 24
Bama 25

Clemson does not get credit for remaining num 1 after beating OU - they don't have poll after the semifinals.

Of course we all know that OSU did not even make the playoffs.

4 times Bama won the NC being ranked num 1 only 1 time. The most this happened to anyone else was 2 times.

A team that is ranked num 1 for the first 13 weeks then misses out on the playoff by losing their last game, then loses their bowl might end up out of the top 10 but would garner more points than the eventual NC - 28-25. That is not far fetched either, in fact similar things have happened.

Their method puts too much emphasis on preseason supposition and totally shortchanges the ultimate achievement of winning the NC.

IMO, a far greater indication of teams' performance would be to do something like the following:

1) Don't start counting points til the 5th week of the season

2) Use a weighted system giving a little more weight to the polls as the season progresses. E.g., give 10% more weight than the first poll, to next week's poll. (I.e., avoid the fractions). Thus, the final poll would be 110% or a bit more than twice as valuable as the first poll counted.

3) Give points based on position. For example start off with 25 for 1, 24 for 2...1 for 25

4) Give a significant bonus to the NC. E.g., using the point system above, a bonus of 200 points above the 2nd highest point total.


As all know, there is no perfect system, but this would be easy to do and more importantly, IMO, a far more accurate assessment. Having said that, I'm sure there are other methods yet better.

 

[B1GTide]

Maybe I am too simple minded, but for me the only way to rank programs is by the number of championships that they have won - conference and national championships. Awarding a program points for looking good to media members during the season is stupid, unless you are one of those media members.