Phil
Stats Geek
- Joined
- Aug 25, 2011
- Messages
- 4,495
- Reaction Score
- 5,951
Check the linked Google sheet, you will see that the top ten teams (12/31/2025) are:
Team
1 Connecticut
2 Texas
3 South Carolina
4 UCLA
5 LSU
6 TCU
7 Maryland
8 Michigan
9 Oklahoma
10 Iowa St
I then added the Massey rating value as of 31 December, and the value as of 2 January, the day after the large number of games. There is also a Delta column which is how much the rating changed between those two dates.
Let's first focus on the questions about LSU, Maryland, and Michigan.
Look at the graph on the Google sheet which I will try to embed here as well. Please note that I did a lower truncation of the values at 8.0 to highlight the changes.
Each of them lost, so not surprisingly the red bar, reflecting the rating on 2 January is lower.
Now look at how much the bars dropped for those three teams (or look at the Delta column). The Maryland and Michigan decreases were substantial, while the LSU decrease was even larger. I can imagine someone saying that the LSU result was a closer call, but LSU is rated higher which means it's a bigger deal to be upset.
And now for the interesting part — if the LSU loss was so significant, much more so than the Maryland or Michigan loss, why on earth what I argue that the LSU ranking should be unchanged.
The simple answer is that while ordinal ranking a suggests that LSU at 5 is just a little bit better than TCU at 6, the rating shows that there is a bigger disparity. It is the case (as some others have stated) that there are five top teams, than a gap, then everyone else. LSU's rating is 8.9 while TCU's is 8.53, a gap of 0.37. That single loss chewed up more than half the gap, But they had enough of a margin that it still didn't drop them below TCU even though TCU gained a tiny bit in those three days.
Maryland and Michigan had Massey ratings of 8.49 only a hair's breath above Oklahoma's 8.48. So while their losses were worth 0.17 and 0.14 rating points, a smaller difference than in the case of LSU, that movement drop them much further.
Maryland's movement in rating points was less than LSU's movement but dropping from 8.49 to 8.32 moved them past Iowa State, USC, Iowa, Michigan, Kentucky, and Vanderbilt.
I don't for a second think the voters will buy into this. If only due to worries about fan reactions, they won't claim that they moved LSU down in their estimation but not far enough to drop below the #6 team. I think they will move more than that but that doesn't mean they are right.
Team
1 Connecticut
2 Texas
3 South Carolina
4 UCLA
5 LSU
6 TCU
7 Maryland
8 Michigan
9 Oklahoma
10 Iowa St
I then added the Massey rating value as of 31 December, and the value as of 2 January, the day after the large number of games. There is also a Delta column which is how much the rating changed between those two dates.
Let's first focus on the questions about LSU, Maryland, and Michigan.
Look at the graph on the Google sheet which I will try to embed here as well. Please note that I did a lower truncation of the values at 8.0 to highlight the changes.
Each of them lost, so not surprisingly the red bar, reflecting the rating on 2 January is lower.
Now look at how much the bars dropped for those three teams (or look at the Delta column). The Maryland and Michigan decreases were substantial, while the LSU decrease was even larger. I can imagine someone saying that the LSU result was a closer call, but LSU is rated higher which means it's a bigger deal to be upset.
And now for the interesting part — if the LSU loss was so significant, much more so than the Maryland or Michigan loss, why on earth what I argue that the LSU ranking should be unchanged.
The simple answer is that while ordinal ranking a suggests that LSU at 5 is just a little bit better than TCU at 6, the rating shows that there is a bigger disparity. It is the case (as some others have stated) that there are five top teams, than a gap, then everyone else. LSU's rating is 8.9 while TCU's is 8.53, a gap of 0.37. That single loss chewed up more than half the gap, But they had enough of a margin that it still didn't drop them below TCU even though TCU gained a tiny bit in those three days.
Maryland and Michigan had Massey ratings of 8.49 only a hair's breath above Oklahoma's 8.48. So while their losses were worth 0.17 and 0.14 rating points, a smaller difference than in the case of LSU, that movement drop them much further.
Maryland's movement in rating points was less than LSU's movement but dropping from 8.49 to 8.32 moved them past Iowa State, USC, Iowa, Michigan, Kentucky, and Vanderbilt.
I don't for a second think the voters will buy into this. If only due to worries about fan reactions, they won't claim that they moved LSU down in their estimation but not far enough to drop below the #6 team. I think they will move more than that but that doesn't mean they are right.
