CSNbbs

All kinds of debate over PSU vs Ohio State, over including Washington, etc.

But if we had the old BCS, would there be any question about Alabama vs Clemson?

(12-08-2016 10:59 AM)johnbragg Wrote: [ -> ]All kinds of debate over PSU vs Ohio State, over including Washington, etc.

But if we had the old BCS, would there be any question about Alabama vs Clemson?

Actually yes. It would have been Alabama vs Ohio St, and it wouldn't have been close. Ohio St is a clear #2 in both polls, and would have had a unanimous #2 computer rating- they're #2 in 5 of the polls, 3 in the other. hi/low removed gives them an average of 2. Clemson on the other hand has computers of 3,4,3,7,5,4. Their hi/low removed gives them an average of 4.

to give you an idea of how far Ohio St would be ahead, Clemson would have had to be ahead in the AP poll by 61 votes, and the coaches poll by 58 votes to beat Ohio St. Currently, Ohio St is 48 votes ahead in AP and 7 votes ahead in coaches poll.

(12-08-2016 10:59 AM)johnbragg Wrote: [ -> ]All kinds of debate over PSU vs Ohio State, over including Washington, etc.

But if we had the old BCS, would there be any question about Alabama vs Clemson?

Yes. It would have been Ohio State vs Alabama

For whatever reason, the computers are not so high on Clemson.

Massey's composite of computer rankings has 1-6, in order: Alabama, Ohio State, Washington, Clemson, Michigan, Penn State.
http://www.masseyratings.com/cf/compare.htm

the BCS standings would have been:
Alabama 1.000
Ohio St 0.947
Clemson 0.908
Washington 0.885
Penn St 0.823
Michigan 0.811

(12-08-2016 11:50 AM)stever20 Wrote: [ -> ]the BCS standings would have been:
Alabama 1.000
Ohio St 0.947
Clemson 0.908
Washington 0.885
Penn St 0.823
Michigan 0.811

So Clemson is closer to UW than Ohio State, based on BCS Formula

(12-08-2016 11:45 AM)Wedge Wrote: [ -> ]For whatever reason, the computers are not so high on Clemson.

Massey's composite of computer rankings has 1-6, in order: Alabama, Ohio State, Washington, Clemson, Michigan, Penn State.
http://www.masseyratings.com/cf/compare.htm

Clemson had a weaker SOS than Ohio State (#1 vs #7), lost to a 3-loss team (thus lower ranking for Louisville) in a serious blowout. Not really sure how Washington got ranked higher than Clemson though, given that they had a weaker strength of schedule (#7 vs #15). It could be because they lost to a higher ranked team (USC) than who Clemson lost to.

AP is irrelevant because they pulled out of the BCS process a few years ago, so it was the coaches and the now-defunct Harris Interactive polls that were fed into the system. Plus if the voters in either human poll know that their decisions in part affect the national title game matchup, does that change how they cast their votes? Particularly in the final week?

So how the bowls would look if we were still playing by BCS rules today is difficult to determine because one of the polls doesn't even exist.

(12-08-2016 11:03 AM)stever20 Wrote: [ -> ]

(12-08-2016 10:59 AM)johnbragg Wrote: [ -> ]All kinds of debate over PSU vs Ohio State, over including Washington, etc.

But if we had the old BCS, would there be any question about Alabama vs Clemson?

Actually yes. It would have been Alabama vs Ohio St, and it wouldn't have been close. Ohio St is a clear #2 in both polls, and would have had a unanimous #2 computer rating- they're #2 in 5 of the polls, 3 in the other. hi/low removed gives them an average of 2. Clemson on the other hand has computers of 3,4,3,7,5,4. Their hi/low removed gives them an average of 4.

It is hard to tell how it would play out. . On the one hand, Ohio State was not a "clear" number two. The BCS poll did not use the rank as much as it used the points, so how close of a number two mattered, and they were close, especially in the coaches poll, the one that counted, where they were seven points a part. Basically tied (they were not quite as close in the AP, but the AP does not matter).

Now the computers are a problem for Clemson, but there are two things sort impossible to predict if the BCS were still in play. One, is the BCS forced the computers to remove all traces of margin of victory, while once the BCS was dead, the computers started to add it back to a degree. Clemson's computer numbers, as compared to OSU, would have been affected by their number of close games. Second, voters tend to vote different when the outcome matters. Meaning them voting for the final AP or coach's poll today didn't matter, one because the playoffs do their own rankings, and two the top four get in, not just the top two. That changes how people vote. Also the playoff committee itself influenced the voting, as we saw thru the season, something that would not have existed in a BCS world.

If we took simply the results as they are, and applied the BCS formula today, yeah Ohio State would be in. But you cannot do that, because the data would likely change if we did a BCS situation. IIRC in the BCS, only two teams who did not win their conference championship played in the BCS title game. One was a rematch of what people considered the two best teams that year, but were in the same conference (hence a non champion). The other was because SOS was the most dominant factor in the BCS at that time. The formula was equal parts SOS, computer rank, and human rank, plus a quality win component, and a loss component. And since both the human and computer polls ALSO use SOS, in addition to the quality win part basically being a component of SOS. that made SOS the most important data point. So when the number one team in the human polls was number 3 in the computers, and the number one team in the computer polls, was number 3 in the human polls, and was also number one in SOS (hence the computer lead), the split kept them number one in the BCS, and left out the number one team in the human polls. Those are the only two instances. So just using the logic that mostly governed the BCS, Ohio State likely would not have gotten the nod in the human polls if being number 2 mattered.

(12-08-2016 11:10 AM)stever20 Wrote: [ -> ]to give you an idea of how far Ohio St would be ahead, Clemson would have had to be ahead in the AP poll by 61 votes, and the coaches poll by 58 votes to beat Ohio St. Currently, Ohio St is 48 votes ahead in AP and 7 votes ahead in coaches poll.

In addition to the above, AP was not a part of the BCS the last several years. The coach's poll was nearly dead even. It's not a simple as looking at the rankings now, hence why it is hard to answer.

(12-08-2016 11:45 AM)Wedge Wrote: [ -> ]For whatever reason, the computers are not so high on Clemson.

Massey's composite of computer rankings has 1-6, in order: Alabama, Ohio State, Washington, Clemson, Michigan, Penn State.
http://www.masseyratings.com/cf/compare.htm

close games, compared to OSU having a lot of blowouts. However the computers under the BCS could not utilize MOV.

(12-08-2016 12:13 PM)AntiG Wrote: [ -> ]

(12-08-2016 11:45 AM)Wedge Wrote: [ -> ]For whatever reason, the computers are not so high on Clemson.

Massey's composite of computer rankings has 1-6, in order: Alabama, Ohio State, Washington, Clemson, Michigan, Penn State.
http://www.masseyratings.com/cf/compare.htm

Clemson had a weaker SOS than Ohio State (#1 vs #7), lost to a 3-loss team (thus lower ranking for Louisville) in a serious blowout. Not really sure how Washington got ranked higher than Clemson though, given that they had a weaker strength of schedule (#7 vs #15). It could be because they lost to a higher ranked team (USC) than who Clemson lost to.

Clemson lost to Pitt and beat Louisville.

Per the Massey composite of rankings, Washington played three teams ranked higher than Clemson's toughest opponent.

Washington's five highest-ranked opponents: (9) USC, (11) Colorado, (13) Stanford, (18) Washington State, (32) Utah.

Clemson's five highest-ranked opponents: (14) Auburn, (15) Louisville, (23) Virginia Tech, (28) Pittsburgh, (39) Georgia Tech.

(12-08-2016 12:54 PM)Wedge Wrote: [ -> ]

(12-08-2016 12:13 PM)AntiG Wrote: [ -> ]

(12-08-2016 11:45 AM)Wedge Wrote: [ -> ]For whatever reason, the computers are not so high on Clemson.

Massey's composite of computer rankings has 1-6, in order: Alabama, Ohio State, Washington, Clemson, Michigan, Penn State.
http://www.masseyratings.com/cf/compare.htm

Clemson had a weaker SOS than Ohio State (#1 vs #7), lost to a 3-loss team (thus lower ranking for Louisville) in a serious blowout. Not really sure how Washington got ranked higher than Clemson though, given that they had a weaker strength of schedule (#7 vs #15). It could be because they lost to a higher ranked team (USC) than who Clemson lost to.

Clemson lost to Pitt and beat Louisville.

Per the Massey composite of rankings, Washington played three teams ranked higher than Clemson's toughest opponent.

Washington's five highest-ranked opponents: (9) USC, (11) Colorado, (13) Stanford, (18) Washington State, (32) Utah.

Clemson's five highest-ranked opponents: (14) Auburn, (15) Louisville, (23) Virginia Tech, (28) Pittsburgh, (39) Georgia Tech.

Oh yes, my mistake, I mixed them up with FSU.

(12-08-2016 12:49 PM)adcorbett Wrote: [ -> ]

(12-08-2016 11:03 AM)stever20 Wrote: [ -> ]

(12-08-2016 10:59 AM)johnbragg Wrote: [ -> ]All kinds of debate over PSU vs Ohio State, over including Washington, etc.

But if we had the old BCS, would there be any question about Alabama vs Clemson?

Actually yes. It would have been Alabama vs Ohio St, and it wouldn't have been close. Ohio St is a clear #2 in both polls, and would have had a unanimous #2 computer rating- they're #2 in 5 of the polls, 3 in the other. hi/low removed gives them an average of 2. Clemson on the other hand has computers of 3,4,3,7,5,4. Their hi/low removed gives them an average of 4.

It is hard to tell how it would play out. . On the one hand, Ohio State was not a "clear" number two. The BCS poll did not use the rank as much as it used the points, so how close of a number two mattered, and they were close, especially in the coaches poll, the one that counted, where they were seven points a part. Basically tied (they were not quite as close in the AP, but the AP does not matter).

Now the computers are a problem for Clemson, but there are two things sort impossible to predict if the BCS were still in play. One, is the BCS forced the computers to remove all traces of margin of victory, while once the BCS was dead, the computers started to add it back to a degree. Clemson's computer numbers, as compared to OSU, would have been affected by their number of close games. Second, voters tend to vote different when the outcome matters. Meaning them voting for the final AP or coach's poll today didn't matter, one because the playoffs do their own rankings, and two the top four get in, not just the top two. That changes how people vote. Also the playoff committee itself influenced the voting, as we saw thru the season, something that would not have existed in a BCS world.

If we took simply the results as they are, and applied the BCS formula today, yeah Ohio State would be in. But you cannot do that, because the data would likely change if we did a BCS situation. IIRC in the BCS, only two teams who did not win their conference championship played in the BCS title game. One was a rematch of what people considered the two best teams that year, but were in the same conference (hence a non champion). The other was because SOS was the most dominant factor in the BCS at that time. The formula was equal parts SOS, computer rank, and human rank, plus a quality win component, and a loss component. And since both the human and computer polls ALSO use SOS, in addition to the quality win part basically being a component of SOS. that made SOS the most important data point. So when the number one team in the human polls was number 3 in the computers, and the number one team in the computer polls, was number 3 in the human polls, and was also number one in SOS (hence the computer lead), the split kept them number one in the BCS, and left out the number one team in the human polls. Those are the only two instances. So just using the logic that mostly governed the BCS, Ohio State likely would not have gotten the nod in the human polls if being number 2 mattered.

Polls-
So I looked at the obvious comparison in 2011 between Alabama and Oklahoma St as this was the closest comp IMO.

IN AP-
Alabama 1418 .9453
Okla St 1400 .9333

But in Harris-
Alabama 2723 .9471
Okla St 2654 .9231

so Alabama had a MUCH larger lead in Harris than they did in AP.

In the coaches poll Alabama was up by 32 votes.

So, I don't think you can automatically say that Clemson would have been better off because they were using Harris vs AP.

Now the computers. In the 4 computers that don't use MOV, Clemson was 3,4,3,4(while all 4 had Ohio St #2). So if they replicate that in the other 2, they have an average of 3.5.

So the margin that Clemson would have needed to be ahead of would be only 3/4 of what I said. So about 49 AP and 47 coaches.

Now yes, some votes would have been altered if it had been the BCS. But I don't think it would have been enough. Definitely the original hypothesis is incorrect that it would be a no brainer with Alabama vs Clemson.

About the committee impacting the polls. I think if anything, Clemson got helped by the committee, not hurt.

(12-08-2016 01:19 PM)stever20 Wrote: [ -> ]So I looked at the obvious comparison in 2011 between Alabama and Oklahoma St as this was the closest comp IMO.

IN AP-
Alabama 1418 .9453
Okla St 1400 .9333

But in Harris-
Alabama 2723 .9471
Okla St 2654 .9231

so Alabama had a MUCH larger lead in Harris than they did in AP.

In the coaches poll Alabama was up by 32 votes.

So, I don't think you can automatically say that Clemson would have been better off because they were using Harris vs AP.

That is not what I said. I pointed out the AP was not as close, but it was irrelevant. The primary thing is, when it doesn't matter if you are 1-4, you will vote differently then if you have to be 1 or 2. We saw that specifically in 2006 and 2011 the year you are talking about. What you are missing is not the final vote, it is how the final vote compared the week before, when OK St went up 122 points, when the voters move Oklahoma St WAY up, and many dropped Alabama to 4 or 5 - which no one really thought they should be ranked that low, they were just trying to deflate their ranking specifically to try to get them out of the number 2 position - because it mattered.

(12-08-2016 01:19 PM)stever20 Wrote: [ -> ]Now yes, some votes would have been altered if it had been the BCS. But I don't think it would have been enough. Definitely the original hypothesis is incorrect that it would be a no brainer with Alabama vs Clemson.

That is no small thing. It the two times the scenario came up, where the team perceived to be number two, but did not win their conference, had an active campaign to remove them. Once it worked (2006), once it did not (2011). I didn't count 2002 since the voters had Oklahoma a clear number 3, so no need to have a campaign, but the unexpected issue of the computers and SOS moving them up.

(12-08-2016 01:19 PM)stever20 Wrote: [ -> ]About the committee impacting the polls. I think if anything, Clemson got helped by the committee, not hurt.

Could be. I was stating in general the committee affects the polls, so in a committee-less world, you cannot assume the final human poll rankings would be where they ended up in a committee world.

Okla St was never up 122 votes. Yes, it got closer. But because of the margin in the computers, Oklahoma St didn't need to just pass Alabama. They needed to be way ahead of Alabama. And to put in context. Even if you give Clemson a unanimous 3 rating in the computers, they would have been further behind in the computers than Oklahoma St was. Clemson would have needed a bigger gap with Ohio St than Oklahoma St needed with Alabama.

Also, you can say AP was irrelevant but the Harris Poll always acted much more like AP than the coaches poll. So you can't just say that well AP doesn't matter, and well coaches were only 7 point gap. Media views things a lot differently than the coaches do.

I think we can both agree that it wouldn't be a no-brainer either way for Clemson or Ohio St.

(12-08-2016 12:49 PM)adcorbett Wrote: [ -> ]So just using the logic that mostly governed the BCS, Ohio State likely would not have gotten the nod in the human polls if being number 2 mattered.

Huh? We can't know what the "Harris Poll" would have done, but assuming they would have had OSU #2, like the Coaches and AP poll did, then Ohio State would have easily trumped Clemson both in the early years where the computer formula was more important, and in the later years when the polls were weighted more heavily.

Whether 2000 or 2008 or 2013, any BCS formula that was actually used would have made Ohio State #2.

(12-08-2016 01:50 PM)stever20 Wrote: [ -> ]Okla St was never up 122 votes. Yes, it got closer.

I didn't say that. I said they went up 122 POINTS. They went from 1245 points in the week before the CCG's at the five spot, and went up to 1367 the next week to get the three spot.

(12-08-2016 01:56 PM)quo vadis Wrote: [ -> ][quote='adcorbett' pid='13885595' dateline='1481219385']
So just using the logic that mostly governed the BCS, Ohio State likely would not have gotten the nod in the human polls if being number 2 mattered.

No one made an assumption of what the Harris might have been. He used the AP poll as an example, and I pointed out the AP was not used. The point made, was if the choice was to choose who is number 2, as opposed to it not mattering if you are ranked 2, 3, 4, then how you vote would change. That is what I said, and even the person I am discussing with, agrees with that. This was not a point being argued.

I think they would have put Clemson in the game. I don't think they would have put a non-conf champ in the top 2.

OSU wouldn't have the extra data point that Clemson had. Plus it's been stated already that they factored in the Harris Poll 1/3, so I don't think the human element would have rewarded one less game played, regardless of the win over OU.

Clearly Clemson would have gotten in having played game #13 and having scheduled Auburn OOC as opposed to Washington's OOC.

some of Oklahoma St's improvement in the polls was that a team directly ahead of them, Virginia Tech, lost.

looking at things closely....
Alabama in Harris on 11/27. 2756 Coaches 1411.
Alabama in Harris on 12/4. 2723. Coaches 1399.

So there were only 33 votes lost in Harris and only 12 in coaches. So yes, there was some difference. BUT no where near enough there. And like I've said the computer gap this year would have been much larger(instead of .95 to .93, it would be at least .96 to .92, if not more)