![]() ![]() |
![]() |
![]() |
|
|
LinkBack | Thread Tools | Display Modes |
|
|
#91 (permalink) | |
|
Veteran
Join Date: May 2012
Location: Shanghai, China
Age: 32
Posts: 1,066
Rep Power: 192540
|
Re: 12/8 A-10 Games
Quote:
The basketball team don't have the players to be competitive this season, especially with Bill Edwards out for the year. Sullivan transferring was a big blow, too. We'll see what kind of players Cooper can bring in. It will never be the same anyway without Charlie Coles. Nice win for Illinois over Gonzaga in Spokane. They could be for real. Last edited by bprichard; 12-08-2012 at 09:46 PM. |
|
|
|
|
| Sponsored Links | |||
Advertisement | |||
|
|
#95 (permalink) |
|
Star
Join Date: Nov 2003
Posts: 2,727
Rep Power: 226644
|
Re: 12/8 A-10 Games
antboy - sorry - haven't been on in a while. You're not really missing anything with regards to the weighting. Conference opponents do count more because they get overweighted in the opponents opponents component. This is why res and anyone else who has looked at the RPI carefully understands that the conference OOC performance is critical to good RPI (so that every team should schedule just at the level where they go something like 9-3 OOC). Nevertheless, the distribution of the opponents opponents component of the RPI is a very tight distribution around 0.500 relative to the distribution of either of the other two components (record and opponents record). Every team's opponents opponent component looks like the entirety of D1 (0.500) with slight overweight of the sum total won-loss of their conference. So, strong conferences have a slightly greater than 0.500 opponents opponents while weak conferences have less than 0.500 opponents opponents. The net impact is to reward strong conference members whose record is similar to those from weak conferences. The same happens, to a larger degree, with the opponents record component since that components is even more dominated by overall conference performance when comparing teams from different conferences. The point of all of this, though, is that the difference between 1.7% and 2.25% as a weighting for an individual opponent's record simply will not be particularly significant in your final rank order in RPI. On the other hand, the sum total of the increased weighting for each conference opponents will matter.
Last edited by UMass87; 12-10-2012 at 11:43 AM. |
|
|
|
|
|
#96 (permalink) |
|
Veteran
Join Date: Oct 2005
Posts: 1,477
Rep Power: 2220
|
Re: 12/8 A-10 Games
Ok, so the argument is that one team in conference is not that much more important than a single team out of conference, though they are a little more important. But having multiple teams in conference with strong OOC records will add up to a larger difference.
I can see that point. For the original point, it's ok to root Miami over Dayton since that game by itself will have little impact, but don't do so for every A10 team. On the other hand, I'd also bring up the other issues: the importance of conference strength and how the committee tends to use the rpi (SOS only) to choices made on the bubble teams, both of which would add to that minor difference. I.e. it pays to root for conference mates OOC for many reasons if you want to make rational choices. But if you really dislike a team, it's probably still not enough to overcome that. |
|
|
|
![]() |
| Currently Active Users Viewing This Thread: 1 (0 members and 1 guests) | |
| Thread Tools | |
| Display Modes | |
|
|