This big news this week is Georgia's loss, and it throws the Playoff picture into chaos. The Bulldogs fall to #22, as they're now expected to have two losses and are not favored to win the conference championship game if they get there. Almost every other team's odds increases to sop up the 45% the Bulldogs lost.
Why did Georgia's odds fall so far—from 46.3% to 1.6%—from one loss? The answer is that one loss is a killer to a team in the SEC, where there are so many teams waiting to take over as the favorite. To project a Playoff foursome you have to make some assumptions about who is going to go undefeated or have just one loss. Georgia went from a team we favored in every game to a 1-loss team that is likely to suffer at least one more loss, and that's a big distinction. Their odds of going undefeated went to zero, too, and that hurts a lot when it comes to the Playoff.
Florida State remains the most likely choice and probable "#1 seed", adding 15 percentage points to their chances without playing a game. Virginia Tech's loss helped, too, as it's that much more likely they win the ACC championship game against a weakened foe, so their odds of going 13-0 went up 2%.
Oregon's odds of going 13-0 fell from 8.8% to 7.2% but their Playoff odds jumped 11 percentage points. Much of that was from UCLA's continued weakness and we now find Oregon favored in every game. USC's tumble didn't hurt the Ducks, either.
Baylor took over the #1 spot in our power rating, and though their odds of going undefeated dropped (due to strong performances by Oklahoma and TCU) they hold on to the #3 spot here. The surprise is that the #4 spot is occupied by Oklahoma after their solid win over Tennessee. They're well below 50% odds so this week's projection isn't saying there will be two Big Twelve teams in the Playoff, but that one or the other of Baylor and Oklahoma will get in.
It's not impossible for two Big Twelve teams to get in; that scenario would include both Baylor and Oklahoma going undefeated in a dominant fashion then playing a very close game in Norman, perhaps with Oklahoma winning in overtime so a rematch seems necessary. But it's hard to imagine two Big Twelve teams making it ahead of every SEC team. That's why these results should be interpreted as having three solid teams (all above 50%) and a wide-open field for the 4th spot, and as the SEC picture solidifies a true front-runner will emerge for that spot.
SportsRatings College Football Playoff odds chart after games of week 3 (9/14/2014)
Dropped out: #7 Virginia Tech, #10 USC, #23 Florida, #24 Louisville, #25 Indiana
Others receiving "votes": USC (2-1), West Virginia (2-1), Virginia Tech (2-1), Navy (2-1), Florida (2-0), Mississippi State (3-0)
Why isn't Alabama a clear front-runner? The Tide actually drop back a spot, behind Michigan State who already has a loss to Oregon. The reason is that Alabama, as of this week, is now expected to lose at Mississippi, and they still aren't favored to beat Georgia for the SEC title. In other words, the SEC teams are expected to cannibalize each other (and that already started happening with South Carolina's win over Georgia). The projection is at a "perfect storm" point where every SEC team ends with 2 losses or more with only a "fuzzy" championship game winner. So until more concrete results come in, Alabama and Auburn are on the sidelines, though the Tigers made a good-sized jump to #9 without playing a game.
BYU didn't have the strongest showing against Houston but they moved up to #21 in the polls, which is where they need to make their progress. Going undefeated is already figured in to these projections for the Cougars for the most part, so they won't break the top 5 without making the top 5 in the polls. Another non-Power Five longshot, Marshall, got their mojo back and moved up to #8. The Herd are again favored to go 13-0 after a strong win over Ohio.
Two Big Ten teams rebounded with strong performances and jumped back into the thick of things. Ohio State's 66-0 win over Kent State has our power rating believing that the Buckeyes can possibly win the Big Ten and they return at #12. Their odds are roughly equal to Wisconsin's, since the Badgers have a much easier path to a potential conference title collision. Nebraska also had a great showing, 55-19 against Fresno State and they return at #18. The Big Ten suffered more bad losses (Indiana, Maryland, Minnesota, Iowa) so the rest of the conference saw their schedules get even easier.
Missouri kept the SEC at 7 representatives on the chart as Florida drops off (due to needing three overtimes to beat Kentucky). The Tigers belted UCF 38-10 and also gained a bit from Georgia's power rating dropping. The Bulldogs loss to South Carolina demoted them enough that they are expected to lose to Auburn, which puts them at 10-2 straight up, and around 9-3 projected. Texas A&M, Ole Miss, and LSU gained partly from strong wins, partly because Georgia is a potential SEC title game opponent; none of the three plays the Bulldogs in the regular season.
Duke and Utah join the top 25—Duke because they beat Kansas badly enough that they're looking like a solid ACC Coastal contender again, and Utah because so many other borderline contenders bit the dust last week—including USC, who looks more more beatable in the Pac-12 South now. But neither teams has much in the way of good odds; they'll need a big upset here and there to get that.
Next week is very important for Northern Illinois' odds. They play Arkansas, and it's the only game we expect them to lose. If they beat the Razorbacks they'll get partial credit for a projected undefeated season and make the jump into the top ten—where they'd wait along with BYU and Marshall to see if the season is going to be wild enough to clear out a space for one of them.
The most important game next week is Clemson at FSU, which is make or break for the Tigers. Lose as expected and their #14 ranking becomes a ceiling, but if they win they bust through that ceiling and Florida State becomes just another team scrambling for a bid.