[stupac2]

The fivethirtyeight site is a bit odd:

https://fivethirtyeight.com/features/our...-favorite/

Alabama has only a 14% chance of winning the SEC, but a 41% chance of making the playoffs, whereas Florida has a 41% chance of winning the SEC, but only an 18% chance of making the playoff.  Since it is nearly guaranteed to win the SEC EAst (it has only Vanderbilt and SOuth Carolina left in its conference schedule), the logic here is that Florida has about a 40% chance of beating the SEC West champ (which could be LSU, Alabama, or Ole Miss).  I guess the thinking is that Florida might win the SEC but lose to Florida State (FEI predicts a 65% chance of winning against FSU), and that if Ole Miss makes it to the conference championship game and loses to Florida, the committee would pick a one-loss Alabama who did not make it to their conference championship over a two loss Florida and  a two loss Ole Miss.  Still, those probabilities don't seem large enough to account for the discrepancy between Alabama's SEC title chances and their playoff chances.

BC

IIRC Silver said that the bonus for winning your conference varies in the simulation rounds, from nothing to quite large, with the average being somewhat small. I think it's likely that they're underweighting it.

[stupac2]

The fivethirtyeight site is a bit odd:

https://fivethirtyeight.com/features/our...-favorite/

Alabama has only a 14% chance of winning the SEC, but a 41% chance of making the playoffs, whereas Florida has a 41% chance of winning the SEC, but only an 18% chance of making the playoff.  Since it is nearly guaranteed to win the SEC EAst (it has only Vanderbilt and SOuth Carolina left in its conference schedule), the logic here is that Florida has about a 40% chance of beating the SEC West champ (which could be LSU, Alabama, or Ole Miss).  I guess the thinking is that Florida might win the SEC but lose to Florida State (FEI predicts a 65% chance of winning against FSU), and that if Ole Miss makes it to the conference championship game and loses to Florida, the committee would pick a one-loss Alabama who did not make it to their conference championship over a two loss Florida and  a two loss Ole Miss.  Still, those probabilities don't seem large enough to account for the discrepancy between Alabama's SEC title chances and their playoff chances.

BC

IIRC Silver said that the bonus for winning your conference varies in the simulation rounds, from nothing to quite large, with the average being somewhat small. I think it's likely that they're underweighting it.