I agree with ELA .. but here is dumb question, how could we do tiers for home/away .. ie Nebraska is an absolute beast at home, but not so much on the road? Or this year do we say home team should beat someone 2 tiers higher or certain teams that are absolute beasts at home should get a 2 tier win probability?Teams that are home beasts: Nebraska, Mich St, Purdue, Northwestern(?)
- 1 - Illinois, Michigan, Purdue, MSU
- 2 - Wisconsin, Maryland, Oregon
- 3 - OSU, UCLA, Nebraska
- 4 - Iowa, PSU
- 5 - Northwestern, Indiana, USC
- 6 - Rutgers, Washington
- 7 - Minnesota
I think we should just go with our traditional set-up. For review that is:
- A team should win all home games except those against teams at least 2 tiers better.
- A team should lose all road games except those against teams at least 2 tiers worse.
Using tier-3 as the example:
UNL, UCLA, and tOSU should:
- Lose both H&A to the tier-1 teams: IL, M, PU, and MSU
- Split a series H&A with the tier-2, tier-3, and tier-4 teams: UW, UMD, OR, each other, IA, PSU
- Win both H&A vs the tier-5, tier-6, and tier-7 teams: NU, IU, USC, RU, UDUB, MN
Thus, in a full, 34 game double-round-robin they would each project to go 19-15.
Then I'll subtract the 14 games not played (7 home and 7 road) to come up with a projection.
We should keep an eye on the "upsets" to make sure our model is working. If there are a bunch of instances of home teams beating teams two tiers better then maybe we need to adjust and say that you should win all home games except those against teams at least 3 tiers better.
One thing I want to point out early just to get it on everyone's radar is that just like in football, the larger conference means that there WILL be bigger variations between schedules. In the past the schedule usually only altered the projection by a game or two but now it could be MUCH more. If Ohio State misses home games against a bunch of the tier-1 teams, that will improve tOSU's projected record because they would have lost those. Conversely, if UCLA misses road games against a bunch of bottom feeders that will diminish UCLA's projected record because they would have won those.
All three tier-3 teams would project to go 19-15 on a 34 game double-round-robin but we are only playing 20 games:
- 7 teams home only
- 7 teams away only
- 3 teams H&A
When I back out those 14 games not played it will be more than 40% of the full potential schedule. As a practical matter it will not actually be THIS bad but theoretically you could have a humongous variance between the tier-3 teams where:
- One misses 4 projected wins and 10 projected losses and their season projection is 15-5*
- One misses 8 projected wins and 6 projected losses and their season projection is 11-9^.
*Since you play all teams at least once, you can't miss both losses against the tier-1 teams so the maximum number of projected losses that you could miss would be 10 leaving four losses to the tier-1 teams.
^Since you play all teams at least once, you can't miss both wins against the tier 5-7 teams so the maximum number of projected wins that you could miss would be 8 leaving leaving six wins over the tier tier-5 through tier-7 teams.
