Crayton
Heisman

Posts: 5,313
Joined: Feb 2019
Reputation: 379
I Root For: Florida
Location:
|
RE: 2025 CFP Race (North Texas to the playoff!)
(11-02-2025 12:18 PM)bryanw1995 Wrote: (10-30-2025 06:42 PM)Crayton Wrote: (10-30-2025 04:04 PM)bryanw1995 Wrote: (10-30-2025 03:32 PM)Crayton Wrote: Generally a P2 teams needs 10 wins to get an at large, an M2 needs 11, a G5 12. If we use 50% as the breakpoint, here are the projected exceptions to those rules:
93% 11-1 Navy
80% 10-2 BYU
74% 10-2 Pittsburgh
71% 10-2 Miami
54% 10-2 Utah
42% 10-3 Ohio State
29% 10-2 Ole Miss
24% 10-2 Indiana, Nebraska, Vanderbilt, Texas A&M
24% 11-2 Georgia Tech
16% 10-2 Ohio State
12% 11-2 Virginia
8% 11-1 Houston
1% 11-2 Houston
I'm confused, why is 10-2 OSU so much less likely to get an invite than 10-3 OSU? Losing the B1G CCG shouldn't hurt much, but it has to hurt at least a little bit, right? Also, are you saying that Indiana, Nebraska, Vanderbilt and A&M are all at a 24% to chance to make the CFP at 10-2? That's confusing when we consider the current and remaining SOS for the 4 schools:
A&M: 14 current, also 14 remaining SOS
Vanderbilt: 32, 8
Nebraska: 67, 23
Indiana: 42, 33
I'd also question how an 11-2 GT that lost their CCG + another game, just barely keeps squeaking by substandard opponents, and has a SOS of 82, 43 would get in over either Vanderbilt or A&M.
All of that to say nothing of BYU, Pitt, Utah and Miami:
BYU: 49, 24
Pitt: 91, 16
Miami: 36, 56
Utah: 53, 50
And even Ole Miss ahead of Vandy or A&M seems odd:
Ole Miss: 19, 59
https://www.espn.com/college-football/fpi/_/view/resume
If I'm misreading what you wrote, sorry about that. Are you saying that these are the probabilities that these teams finish with those records?
You’re reading it right. These are the probabilities, given that record, that a team makes the playoff. Strength of Record is the primary way teams are sorted, but there are 3 additional considerations.
1. The biggest additional consideration is that we’re not strictly comparing apples to apples; these teams are not compared against the same set of 12 other potential playoff teams. A 10-2 P2 team would be ranked #3, #4, or worse in their conference. In a scenario where that same team goes 10-3, they were ranked #1 or #2 in their conference, meaning they had ~2 extra at large spots not taken up by conference mates to cushion the blow of a CCG loss (which, yes, would be minor but still something), making the playoff more often than if they were 10-2.
2. How you finish matters. Do you drop your last two games? Does your star player get injured? Human committees tend to have a recency bias to them. A 10-2 A&M or Ohio State likely ended the season with a loss. If there is another 10-2 team like Miami or Pitt who ended with some good wins (and a generally strong eye-test), the committee is going to err on the side of the hot hand (or hot “brand” though I refuse to model that).
3. The underlying FPI that ESPN uses is different than my own. It looks like I have the P4 more as more equal with each other than ESPN does. That could also be a function of me looking at season-ending numbers (after more .500 play within each conference) while the numbers you cite are current/future SOR after only 9 weeks. Not sure.
I know there were a few specific examples included in your post, but I may dive deeper into a handful of those next week. Explorations like that can help uncover ways to change and improve my model.
We don't need to use ESPN's SOS metric. How about Sagarin's?
A&M: 11
OSU: 22
BYU: 49
Indiana: 50
Alabama: 4
Georgia: 8
Ole Miss: 23
Utah: 43
Oregon: 44
Pitt: 64
etc etc
Sagarin and ESPN seem to be pretty close right now, and with the new instructions to the CFP Committee to factor in SOS, I think that is likely to be given more weight this year. Ie, Indiana gets less credit by blowing out bad teams by 50 than they did last year, and OSU gets more credit for the #22 SOS than they did last year. Indiana's one good win, on the road @ Oregon was nice, but they've mostly played cupcakes otherwise. OSU, by contrast, has played a much more difficult schedule but hasn't blown out their (on avg) much tougher opponents by as many points. This can also be used to compare schools of different conferences, but it's even more instructive when comparing, say, Vanderbilt to Georgia to Ole Miss if all 3 end with the same record. Maybe that just means that the Committee gets them in right order from 8-10, but it also could mean that it gets them in the right order from 10-12, meaning that 1 or 2 of them are left out, and it's the CORRECT one or 2 that get left out instead of the team that won their last game sneaking in despite a weaker overall resume.
I pulled out my SOR and FPI-equivalent Top 25s to see how they compare to ESPN's. For the most part they are in line. And so I believe the majority of the differences in ascertaining whether, for example, 10-2 Utah is better/worse positioned to make the playoff than 10-2 Vanderbilt is due to (a) the few differences highlighted below, and/or (b) the fact that I am looking at season-end values, not current.
It is worth noting that ESPN defines their "current" and "future" SOS differently from each other and that their SOR metric is keyed to the strength of their #13 team, rather than shifting depending on the position being evaluated.
One of the big differences I see is that differences between Texas and Tech Tech. These impact the SOR, respectively, of Vanderbilt and Utah.
Power Ranking (ESPN's FPI)
1 1817 Indiana (2)
2 1778 Ohio St (1)
3 1735 Notre Dame (5)
4 1727 Utah (7)
5 1722 Texas Tech (11)
6 1713 Oregon (3)
7 1711 Texas A&M (9)
8 1708 Alabama (4)
9 1706 Miami (12)
10 1702 USC (10)
11 1685 BYU (14)
12 1666 Georgia (6)
13 1663 Florida St (24)
14 1662 Ole Miss (13)
15 1656 Vanderbilt (17)
16 1648 Missouri (18)
17 1649 Oklahoma (16)
18 1646 Washington (23)
19 1644 Iowa (20)
20 1639 Michigan (19)
21 1640 Texas (8)
22 1636 Illinois (28)
23 1627 Pittsburgh (31)
24 1627 Tennessee (15)
25 1623 Louisville (26)
Strength of Record (ESPN's)
1 Ohio St (3)
2 Indiana (2)
3 Texas A&M (1)
4 BYU (4)
5 Alabama (5)
6 Georgia (6)
7 Ole Miss (7)
8 Texas Tech (10)
9 Louisville (13)
10 Notre Dame (15)
11 Texas (9)
12 Utah (23)
13 Oregon (8)
14 Oklahoma (11)
15 Michigan (14)
16 Virginia (17)
17 Miami (18)
18 USC (20)
19 Vanderbilt (12)
20 Washington (21)
21 Illinois (19)
22 North Texas (33)
23 Georgia Tech (16)
24 Cincy (25)
25 Iowa (26)
|
|