Rpi - cougarguard.com — unofficial BYU Cougars / LDS sports, football, basketball forum and message board

jay santos · 01-29-2007, 04:11 PM

BYU is RPI #33 in my RPI calculations.

If I have time, I might run my own projections for where BYU will finish. BYU nearly always climbs in the RPI through the conference season, so I expect that to happen this year, as well. I think we actually have a good shot at making tourney without winning MWC championship, but it might be hard to get three teams--and AFA and UNLV are near locks right now.

For Geeks Only: It's extremely frustrating nailing the RPI calculation. I'm not matching the official NCAA RPI but I'm close. I might spend an hour or two today trying to figure out why my RPI's don't match.

Indy Coug · 01-29-2007, 04:28 PM

Does your formula account for games that were scheduled, but not played (like some of the Colorado college teams)?

jay santos · 01-29-2007, 04:32 PM

Quote:

Originally Posted by Indy Coug

Does your formula account for games that were scheduled, but not played (like some of the Colorado college teams)?

No, only actual game results. The next thing I'm going to test which I think my problem might be is in how to calculate SOS.

Does opp winning % = total wins by opp's /total games by opp's or is it an average of each opp's W/L? I think that's my problem. Right now I'm just summing all games and giving weighted avg instead of straight avg.

Indy Coug · 01-29-2007, 05:12 PM

Quote:

Originally Posted by jay santos

No, only actual game results. The next thing I'm going to test which I think my problem might be is in how to calculate SOS.

Does opp winning % = total wins by opp's /total games by opp's or is it an average of each opp's W/L? I think that's my problem. Right now I'm just summing all games and giving weighted avg instead of straight avg.

IIRC, it should be the sum of all opponents wins divided by sum of all
opponennts games, less the games directly against that team.

This also applies to opponent-opponent's W-L%.

Example: Team A is 20-0. Their opponents cumulative W-L is 250-200. The opponent winning percentage should be (250 - 20)/(450 - 20) = (230/430)

I did some research at the request of BYU around 4 years ago about the likely difference in RPI results if BYU scheduled some a couple of lower-tier D-1 teams instead of a couple of D-II teams, and I seemed to reconcile with www.collegerpi.com's RPI using the above methodology.

01-29-2007, 04:11 PM	#1
jay santos Senior Member Join Date: Jan 2006 Posts: 6,177	Rpi BYU is RPI #33 in my RPI calculations. If I have time, I might run my own projections for where BYU will finish. BYU nearly always climbs in the RPI through the conference season, so I expect that to happen this year, as well. I think we actually have a good shot at making tourney without winning MWC championship, but it might be hard to get three teams--and AFA and UNLV are near locks right now. For Geeks Only: It's extremely frustrating nailing the RPI calculation. I'm not matching the official NCAA RPI but I'm close. I might spend an hour or two today trying to figure out why my RPI's don't match.

01-29-2007, 04:28 PM	#2
Indy Coug Senior Member Join Date: Jan 2006 Location: Between Iraq and a hard place Posts: 7,569	Does your formula account for games that were scheduled, but not played (like some of the Colorado college teams)?