Perhaps the BEST B1G Forum anywhere, here at College Football Fan Site, CFB51!!!

Once in 50 years means in efffect, never, then rare, pretty rare, pretty often, and nearly all the time (which I doubt of course).

I think you need more choices between #4 and #5.  Your options are:

• Maybe once in 50 years = <2%
• Once in about 20 years = ~5%
• Once a decade = ~10%
• Every five years or so = 20%
• Every year but one = ?  The percentage here is dependent on out of how many years you mean.  Ie if you mean every year but one out of five this is 80% while if you mean every year but one out of 10 then 90% etc.  

IMHO, the jump from ~20% to ~80% is far too large.  

In the other thread @OrangeAfroMan pointed out that there are some REALLY good teams that would have been #5 or #6 seeds or higher over the course of the past couple decades.  Just within the CFP era all of the following teams would have been #5 or #6 seeds and I think that any of them could win the whole thing without needing any major miraculous upsets of vastly superior teams:

• 2019 UGA (#5 in final CFP rankings):  They were 11-2 and the two losses were 20-17 to USCe and 37-10 to LSU in the SECCG.  LSU was just a buzz saw but other than that UGA could stand toe-to-toe with anyone so winning the NC wouldn't have been out of the question, they just needed LSU to have an off night somewhere along the line.  
• 2019 Florida (#9 in final CFP rankings):  They were 11-2 and the two losses were 24-17 to UGA and 42-28 to LSU so same analysis here as for UGA.  
• 2018 UGA (#5 in final CFP rankings):  They were 11-2 and the two losses were 36-16 to LSU and 35-28 to Bama in the SECCG.  This team was capable of standing toe-to-toe with anyone.  
• 2018 tOSU (#6):  This team had one major upset loss (49-20 to Purdue and finished 13-1, they were good enough to win the NC.  
• 2018 M (#7):  I'm not sure how to analyze this one.  Prior to playing tOSU they were 10-1 and their only loss was a close game in South Bend but then tOSU and UF just slaughtered them in the last two games so . . .
• 2017 tOSU (#5):  This team lost early to Oklahoma.  It was 31-16 but note that the Buckeyes lead 13-10 deep in the third quarter before Baker Mayfield put on his superman cape and the Sooners exploded for three unanswered TD's.  Much like 2018, this tOSU team had a blowout loss to a vastly inferior team (55-24 @ Iowa in 2018) but that seems like an aberration considering that the Buckeyes finished 12-2 with a B1GCG win and a Cotton Bowl win over USC.  
• 2017 Wisconsin (#6):  Their only loss was competitive (27-21) to tOSU in the B1GCG so maybe . . .
• 2017 Auburn (#7):  The Tigers lost three games but they were by eight points to Clemson, but four points to LSU, and in the SECCG to a Georgia team that they had previously defeated.  That seems like a team that was close enough to potentially beat any of those teams and thus win the NC.  
• 2017 PSU (#9):  The Nittany Lions lost by one point in Columbus then they let the Buckeyes beat them twice when the lost the very next week in East Lansing.  They won all the rest of their games and a team good enough to go into Columbus and push the Buckeyes that hard was good enough to potentially win it all.  
• 2016 Michigan (#6):  The Wolverines ended up 10-3 with no hardware but in the three losses they trailed by a combined 2 points at the end of regulation.  They lost at Iowa 13-14, lost in OT in Columbus, and lost 33-32 to FSU in the Orange Bowl.  They were good enough to stand toe-to-toe with anyone so they could have won it all.  
• 2016 Oklahoma (#7):  They lost early to Houston which may have just been an early season issue then lost 45-24 at home to Ohio State but after their 1-2 start they went on a 10-game tear and looked the part of an NC contender by season's end.  
• 2016 Wisconsin (#8):  I'm not sure about this one.  They were competitive in all three losses (by a TD each @M, vs tOSU and in the B1GCG to PSU).  
• 2015 Stanford (#6):  They lost early to Northwestern and late to Oregon but both were close and by the end of the season they were firing on all cylinders with a win over Notre Dame followed by absolute demolitions of USC (PACCG) and Iowa (RoseBowl).  
• 2015 tOSU (#7):  The 2015 Buckeyes had a single loss to MSU by a FG and that kept them out of the B1GCG so they ended up on the outside looking in but IMHO they were better in 2015 than they were in 2014 when they did win the NC and they were possibly the best team in the country in 2015.  In their last two games they absolutely slaughtered Michigan and Notre Dame who each finished 10-3.  
• 2015 FSU (#9):  In retrospect the Seminoles were clearly on the way downhill in 2015 after their NC in 2013 and CFP appearance in 2014 but they still had enough talent to show up and play like a legitimate NC Contender in any given week such as the week they went into the Swamp and obliterated the Gators.  The problem was that they were also discombobulated enough to show up and play like absolute crap in any given week such as the week they went to Atlanta and lost to a horrible GaTech squad.  
• 2014 Baylor (#5):  I actually don't think that Baylor could have won it all nor that they were the B12's best team which was:
• 2014 TCU (#6):  Baylor was technically the League Champion because they won the H2H but that was a three point win in Waco.  TCU's loss to Baylor was MUCH better than Baylor's two-TD loss to a mediocre WVU team and hindsight tells us that in the Bowls Baylor lost to the B1G's second best team while TCU annihilated Ole Miss (who had beaten Bama).  

So by my calculations the #5 and higher seeds had the following number of plausibly legitimate NC Contenders:

• 2 in 2019
• 3 in 2018
• 4 in 2017
• 3 in 2016
• 3 in 2015
• 2 in 2014
• 17 total
• 2.8 average per year

My thinking is that the top-4 seeds will almost always all be legitimate contenders so you are going to have an average of about 6.8 legitimate contenders per year.  Thus, each one has an average of a 14.7% chance to win it all.  Then I'd make two adjustments:

• The 14.7% is an average for each legitimate contender but they aren't all equal.  The #1 seed is better than the #2 and the #2 is better than the #3, etc so the #5+ have significantly less than a 14.7% chance each.  
• The rest of the 5+ teams (the ones not included here as legitimate NC Contenders) still *MIGHT* pull a few upsets and win something like once every 50 years.  This second adjustment partially offsets the first.  

Thus, my percentage chance would be a little over 20% (equals 14.7%*2.8 which is 41.2% divided by two to reflect the fact that 1-4 are generally better than 5+).  

20% is every five years so I chose "every five years or so".  

but, are you adding the extra game?

Honestly, I left that off because I'm not sure how it will play out.  

Consider this example:
Suppose that #4 and #5 are essentially equal and interchangeable but that #4 beat #5 in a close game (home game for #4) and thus made the CCG while #5 missed the CCG.  So here are their paths to the NC after the regular season ends (a lot of assumptions here but we don't know yet so just go with it):

Conference Championship Weekend:

• #4 plays in the CCG at a neutral site
• #5 has the week off

First round games (mid December?):

• #4 has the week off.  
• #5 gets a home game against the worst team in the 12-team CFP field.  This should be an EASY win.  

Quarterfinals (bowl games on or around NYD?)

• #4 plays #5 at a neutral site

Semifinals (around the time of the current NCG?)

• #4/5 winner plays #1/8/9 winner at a neutral site

NCG:

• #4/5 winner plays #2/3/6/7/10/11 winner

Note that #5's path is actually easier than #4 and there is the possibility that #4 will be rusty on NYD after a month off while #5 will have had a game mid-December.  

Obviously it isn't always going to work out like that.  if #5 lost the CCG then they do have an extra game and fatigue is going to be a potential factor.  I just don't know.  

Let's imagine the #5 team is really good, it happens.  So, they'd have probably a 90% chance of beating #12, and then perhaps a 55% chance of beating #4, and a 50% chance of beating #1/2.

That's about a 20% chance, one in five, IFF that #5 team is really good.  I don't think some of those UGA teams were very good.

So, if perchance that #5 team is really good every say 4 years, that boils down to a 5% chance per year.

I presume the #4 seed plays the winner of #5vs12 right?

Let's imagine the #5 team is really good, it happens.  So, they'd have probably a 90% chance of beating #12, and then perhaps a 55% chance of beating #4, and a 50% chance of beating #1/2.

That's about a 20% chance, one in five, IFF that #5 team is really good.  I don't think some of those UGA teams were very good.

So, if perchance that #5 team is really good every say 4 years, that boils down to a 5% chance per year.

I presume the #4 seed plays the winner of #5vs12 right?

That's all good, but when the 5 would be good would be just as important:  'on average' is very different than 'randomly'.  The 5 could be very good every 4 years on average, but in any 20-year sample, it would be good in clusters and not be good in clusters.  So it may be good more often early in a decade and not at all in the rest of a decade.  So that 5% per year is going to be tempting fate in year 7 and we won't have to wait 20 years.

IMHO, the jump from ~20% to ~80% is far too large. 

Eh, negligible, because they are both too high

Wow....the fact that 'every 50 years' has any votes blows my mind.  You must really REALLY believe in the rankings, lol.  It's just comically false. 

But it seems odd to hear that, coming from you... You're the one that wants to make sure that only the "best" teams get into the playoff at all, if there need be a playoff.

If you don't trust the rankings enough to discern that the #5 or #6 team will win that "rarely", then how can we assume that the rankings that determined the pre-1997 MNC, the two BCS entrants, or the 4 CFP entrants are useful?