Thread by @iowahawkblog

The Week 13 marble counts for the #CFBMarbleGame are up! Thanks as always, Nate.

https://twitter.com/NateEMiller/status/1596902602914304000

According to Nate's analysis, if the CFP were set by the marble game 6 teams still have a shot at playoff:

TCU: lock, no matter what happens
Michigan: also a lock
Georgia: only out if they lose to LSU and KSU beats TCU
USC: in if they win, out if they lose to Utah

1/2

(actually 7 teams still with chance)

LSU: in if they beat Georgia and USC loses to Utah and TCU beats KSU
Washington: in if Utah beats USC and TCU beats KSU
KSU: in if they beat TCU and (LSU beats Georgia or Utah beats USC)

All other teams are mathematically eliminated.

2/2

Modest proposal:

1. Scrap conference title games and playoff committee
2. 16 team playoff, seeded by marble game
3. First 2 rounds played on campus of higher seed
4. Losers in first 2 rounds still eligible for later bowl games

Marble Game algorithm, in its entirety:

All FBS teams start season with 100 marbles, plus 10 bonus marbles for every P5 team on schedule.

Win at home/neutral site, take 20% of opponent's marbles; win on the road, take 25%.

Fractional marbles are rounded to nearest integer.

Say what you want about marble algorithm but it's simple, transparent, and objective. No subjective "eye tests" or "style points," and unlike mysterious "computer polls" anybody with a calculator can verify the marble counts.

Marble Game algorithm is biased only in the sense it awards bonus starting points for P5 teams; that can be regulated by parameter; e.g., 200 starting marbles for everybody + 10 bonus for each P5 game.

Highest marble rank non P5 this year are #21 Tulane and #22 Troy.

Anyway, this is the 4th or 5th year the Marble Game has been tracked and by end of season it largely mimics the human & computer polls, especially top 4. Non-P5 teams regularly show up in top 16, as high as #8.

If this is my only legacy, I will proudly take it

https://twitter.com/iamjasond/status/1597230014977961985

I get a lot of "SEC should count more" or at least make conference distinctions in pre-season schedule bonus marbles. OK, here's a proposal:

Take average end of season marble count by conference, divide by 12 & round.

This becomes next year's bonus schedule marbles:

So in 2023 every FBS team starts with 100 marbles, plus 20 bonus marbles for every SEC team on schedule, 19 for every B12, 17 for every B10 or P10, 7 for every MAC, etc etc.