PDA

View Full Version : Gu rpi 15 sos 17



DixieZag
12-23-2011, 09:15 AM
Just looked over at the CBS.com real time rpi and was very surprised to see this, both the RPI and the SOS are much higher than I would have thought, given that the "name" teams on our schedule (Notre Dame, Butler, Arizona) are not what we figured they would be when scheduled.

I would not care at all about what a computer ranks us as except for the fact that the committee seems to use it and though I know it will fall as we get into league play, maybe it will not as much as it has in the past. It might be that the computers recognize something that WE certainly have recognized throughout the year - though we have not beaten a top 25 team (0-1 MSU), we have only played ONE truly bad team Hawaii (300+) and the next worst is AFA at around 200. Obviously, Arizona is our best win (46) but much less obviously, Oral Roberts is our second best, and Butler third - and the key seems to be that the rest of the group, WSU, EWU, etc. are not below 200 but closer to 100.

Its all sort of counter-intuitive in a lot of ways. But, WE know that our schedule has not been as easy as others may be thinking, and we are backed up by the computers. Just some food for thought.

maynard g krebs
12-23-2011, 10:28 AM
Fifty percent of RPI is opponents' w/l record, so whether they are ranked 100 or 300 matters only half as much, or 25%, since that is based on THEIR opponents' record, which is half of their rpi. If that makes sense.

What's important is that all but 3 opponents to now have solid winning records, and the 11 teams played have a combined winning % well into the 60s.

As of now, the WCC has a combined record of 65-42. Take out GU's record and conf opponents have a record of 56-40, which means if the Zags can get through conf play w/ only a couple losses, RPI should stay pretty good.

There is a lot written here about how WCC teams' weak schedules hurt GU's RPI, and I question whether that is the case. The reason is that if WCC teams all played, say, Monson's Long Beach team's sched, they'd mostly be about 4-7 at this point, or worse, instead of about 7-4 on average. Again, their SOS would be stronger, but when GU plays them, that would only account for 25% of GU's RPI rating for that game, as opposed to their poorer w/l record, which would count for 50% of GU's RPI for that WCC game. Hope that came across in a coherent manner.

DixieZag
12-23-2011, 10:42 AM
Fifty percent of RPI is opponents' w/l record, so whether they are ranked 100 or 300 matters only half as much, or 25%, since that is based on THEIR opponents' record, which is half of their rpi. If that makes sense.

What's important is that all but 3 opponents to now have solid winning records, and the 11 teams played have a combined winning % well into the 60s.

As of now, the WCC has a combined record of 65-42. Take out GU's record and conf opponents have a record of 56-40, which means if the Zags can get through conf play w/ only a couple losses, RPI should stay pretty good.

There is a lot written here about how WCC teams' weak schedules hurt GU's RPI, and I question whether that is the case. The reason is that if WCC teams all played, say, Monson's Long Beach team's sched, they'd mostly be about 4-7 at this point, or worse, instead of about 7-4 on average. Again, their SOS would be stronger, but when GU plays them, that would only account for 25% of GU's RPI rating for that game, as opposed to their poorer w/l record, which would count for 50% of GU's RPI for that WCC game. Hope that came across in a coherent manner.

That helped a ton. And I see your point, teams like WSU (8-4) and Notre Dame (8-5) and Arizona 8-4 are teams with a lot of wins. Even USF has 9 wins, even though they lost a bad one to Holy Cross last night, additionally SMC obviously has a lot with 10.

The people on the BYU board are suddenly realizing that they do not have a quality win yet - though no real bad losses either, (USU, Baylor) and they are suddenly figuring out that they cannot afford many losses in the WCC, most are very confident that they will not lose more than 2 (they seem to accept the possibility of losing at GU and/or SMC - a home loss seems impossible and a loss to a team other than GU/SMC is not even considered.

I know it will be extremely difficult to win in Provo, I don't think that they know how extremely difficult it will be to win 3/3 in SCU/USF/LMU

bullzag23
12-23-2011, 10:58 AM
That helped a ton. And I see your point, teams like WSU (8-4) and Notre Dame (8-5) and Arizona 8-4 are teams with a lot of wins. Even USF has 9 wins, even though they lost a bad one to Holy Cross last night, additionally SMC obviously has a lot with 10.

The people on the BYU board are suddenly realizing that they do not have a quality win yet - though no real bad losses either, (USU, Baylor) and they are suddenly figuring out that they cannot afford many losses in the WCC, most are very confident that they will not lose more than 2 (they seem to accept the possibility of losing at GU and/or SMC - a home loss seems impossible and a loss to a team other than GU/SMC is not even considered.

I know it will be extremely difficult to win in Provo, I don't think that they know how extremely difficult it will be to win 3/3 in SCU/USF/LMU

They'll get their wake up call soon enough. An opening loss at SMC would be a nice welcome to WCC play.

FieldHouseFishHouse
12-23-2011, 12:41 PM
There are two ways to look at the RPI, both of which are boring. Don't say I didn't warn you.

The first and most common way is to cite a teams RPI - "Utah State can't get an at-large with an RPI above 50". The best way to get a "good RPI" is to play (and to a lesser extent, beat) teams with good W-L records. Two teams in the RPI top 10 have 2 losses (So Miss and No Iowa). Long Beach State ranks #9 with a 4-5 record!!! From this perspective, the RPI game is a scheduling game. You don't have to win every game (though winning% is 25% of the RPI) and you don't have to play teams with high RPIs (opponents win% is twice as important as quality of opponent's opponents). On an even deeper level, a good opponent can completely negate a bad opponent, as only the aggregated opponent win% is used in this formula...but this is getting even more inside baseball.

The second way RPI is used (and from my understanding this is the way it is used by the selection committee) is to cluster quality wins/bad losses etc - "GU has 2 top 100 wins and their only losses are to top 50 teams". From this perspective, it does matter precisely who you play and who you beat. Additionally, the actual RPI of the opponents is important, rather than their winning%+(opponent's winning %)/2.

In an environment where, for instance, "good wins" are highly valued these two ways of looking at the RPI can start to come into conflict. In the past we have seen teams with achieve a modest RPI via a great winning percentage (a result of playing and beating average to below average competition). These teams are sometimes denied NCAA bids. In the RPI calculation, good wins cancel out bad losses without prejudice (this is a simplification, but oh well). However, in practice, a team with 2 marquee wins and two questionable losses can often swing a bid because they scheduled tougher and proved they could hang with top competition.

Anyway, perhaps this is just a long way of saying what everyone already know - The RPI formula can't really capture the qualifications and accomplishments of a college basketball team.

maynard g krebs
12-23-2011, 02:26 PM
There are two ways to look at the RPI, both of which are boring.

Classic.:000tens: