kshoe Posted January 2, 2004 Share Posted January 2, 2004 I noticed today on Ken Pomeroy's RPI sight that CUSA was ranked the 6th highest conference, ahead of the Big 12 which is sitting at 8th. Now I realize that it is still early and RPIs don't mean that much to individual teams right now but they do mean a lot to conferences. Since non-conference action is nearly over there will be few opportunitites for conferences to improve this positioning with outside victories. So the next time some announcer or one of our friends to the West tries to tell us that the Big 12 is a far superior conference than CUSA, just point them to the facts. Obviously a Billiken victory on Saturday would only improve the situation. For what its worth the Bills currently rank 136 on the Pomeroy sight which doesn't include margin of victory/defeat but USA Today has us ranked 103 in RPI type land (point margin doesn't matter) but we rank 63 in Sagarin's computer rankings that do include margin (i.e. losing by 1 and 13 to Arizona and G-Tech is a good thing) http://www.usatoday.com/sports/sagarin/bkt0304.htm http://kenpom.com/rpi.php Quote Link to comment Share on other sites More sharing options...
Basketbill Posted January 2, 2004 Share Posted January 2, 2004 nm Quote Link to comment Share on other sites More sharing options...
slufanskip Posted January 2, 2004 Share Posted January 2, 2004 I am still confused how any system can not count margin of victory and claim to be reliable ... Is a 1 pt loss equal to a 30 pt loss or vice-versa. To stop teams from running up the score you just put a cap on margin of victory ... ie a win by 25 counts the same as a win by 50. Official Billikens.com sponsor of H. Waldman Quote Link to comment Share on other sites More sharing options...
kshoe Posted January 2, 2004 Author Share Posted January 2, 2004 as most of our wins were very close games and the Sagarin model had us about 20 spots lower than the RPI but I agree that, in general it is unreliable to just look at W-L. I think we would all agree, howeve that we would rather beat Louisville by 1 and lose to them by 30 than lose to them by 10 points each time. Quote Link to comment Share on other sites More sharing options...
kwyjibo Posted January 2, 2004 Share Posted January 2, 2004 There are different indexes for different uses. Most of the ratings use margin of victory because their paying audience are bettors, so margin of victory is relevant. ALso, indexes that use margin of victory and home court advantage are trying to be predictive which is different than what the RPI is trying to do. While margin of victory does contain some useful information, I have no problem leaving it out either (it also is biased toward teams that score more points regardless of how well they play defense). The NCAA uses RPI for the simple reason that they want to promote sportsmanship as a goal. Beating the crap out of people by leaving your starters in late is really bad for the sport on many levels. The RPI is NOT trying to PREDICT anything. It is simply one attempt at objectively measuring the relative merits of the teams' performances at the END of the season. The RPI is designed in such a way that it becomes much more reliable with more observations (all indexes are more reliable with more info, it is just that RPI is not designed to be reliable "along the way"; I check it of course because you can see where things are headed a bit). There are a lot of indexes that use margin of victory (I look at and [link:www.kenpom.com|Pomeroy] because I like their approaches and they update their sites frequently. There is also a guy ([link:www.vaporia.com/sports/]Wobus) who ranks the rankings (looks at all the statistical indexes and polls and see who most correctly predicts results) and last year Massey did the best (RPI is mediocre as a predictor but last year polls were the worst!). Right now Markov ratings have SLU 74th (that is the lowest they get I believe). BTW, Massey ranks SLU's defense 8th in D-I--that is great. Quote Link to comment Share on other sites More sharing options...
slufanskip Posted January 2, 2004 Share Posted January 2, 2004 I understand why they don't I just don't agree with it. There is no way I'll ever be convinced that you can claim you have an accurate picture of how good a team is as oppossed to one another if you don't count margin of victory in some way. The rest of the polls are nice ... but they don't influence the NCAA selection comittee as the RPI does. I also understand that the RPI is better than a lot of the polls but why would they not make it as accurate as possible. As I stated they could easily put in some kind of system that stops the value of a big win at say 20-25 pts. Sagarin is not very accurate either because of the way they rank their SOS. In other words beating teams ranked #8, #3, and #222 is not worth as much as beating teams ranked #52, #48 and #65. The game at 222 brings them down too far. If I am incorrect in this assessment let me know ... but I have studied the football Sagarin rankins in depth and believe they are ranked that way. Official Billikens.com sponsor of H. Waldman Quote Link to comment Share on other sites More sharing options...
billiken_roy Posted January 2, 2004 Share Posted January 2, 2004 the top six conference thing is huge. it might get us 5 teams in the tourney which conversely might put us on the bubble as it seems to me we will be fighting for 4th to 8th place. you are correct, we have to win these games vs the top conference opponents. Quote Link to comment Share on other sites More sharing options...
billiken_roy Posted January 2, 2004 Share Posted January 2, 2004 i think the reason they dont consider "margins" in their numbers is because a-holes like billy tubbs would run up the score every time he can and then justify it with the need to win big for his ratings. Quote Link to comment Share on other sites More sharing options...
kwyjibo Posted January 2, 2004 Share Posted January 2, 2004 >There is no way I'll ever be convinced that you can claim >you have an accurate picture of how good a team is as >oppossed to one another if you don't count margin of victory >in some way. I agree margin of victory tells you something useful but think it should be left out of tournament selection. With 30 games played any deficiency in the index from not having it in is mostly gone, however as a mid-season aid to forecasting games I would want it in. There is also an inherent "pro-offense" bias that I do not like. Say the high-flying, run-and-gun Hares and defensively minded Tortoises are in some objective way equal. The Hares and Tortoises are exactly 10% better than AnotherTeam but the Hares win 110-100 and the Tortoises win 55-50. If MOV is a factor then the Hares are artificially ranked better than the Tortoises. And as Billiken fans we tend to root for the Tortoises. Personally, my first gripe with RPI is the influence of the schedule is 75% of the formula (50% is opponents record and 25% is opponents' opponents). They could have made it some other mix (like a more sensible 40-30-30 or even 50-25-25) but the decision was made smartly by the biggest conferences who knew they would have better looking profiles for their marginal teams than winning teams in less competitive conferences. My second gripe is that home court is not directly worked into the formula (but is in the subjective profiling the committee uses). Let me say that the use of computer rankings for football is far more problematic than basketball because there is not enough data. With only 10-11 games played the rankings will have unacceptable margins of error to determine championship slots by. The BCS is a modified format (the computers take too much credit/blame) but the real crime isn't the computers it is the fear of losing revenue/interest with a play-off. It does sound like there will be playoff after the bowls some time in the near future but there will be injustices then as well. Personally, pre-BCS was worse when only polls judge champions and champions could duck big bowls. Quote Link to comment Share on other sites More sharing options...
slufanskip Posted January 2, 2004 Share Posted January 2, 2004 Also the teams strength should be frozen when you play them ... in other words we beat team A and they are ranked #10. The next game they lose their star due to injury ... or just don't ever get back on track ... as they lose the value of our win goes down. Official Billikens.com sponsor of H. Waldman Quote Link to comment Share on other sites More sharing options...
kshoe Posted January 2, 2004 Author Share Posted January 2, 2004 Early season ranking are as subjective as anything and aren't fair indicators of how a team plays throughout the whole season. Someday take a preseason poll and compare it to a poll right before the tourney, you'll see what I mean. In a world like that teams would be lining up outside Columbia to play a team that is consistently ranked in the top 10 to start the season but every year drops out of the top 25 by the time conference season begins. I agree it is unfair when injurioes happen but those are more infrequent than teams simply being over/under ranked which is commonplace. Plus, think if the ranking was done by computers. Some of those RPI calcs super early in the season are just crazy and teams like Wofford can be #1. Surely a win over them shouldn't count as a win over #1. Quote Link to comment Share on other sites More sharing options...
cheeseman Posted January 2, 2004 Share Posted January 2, 2004 Gee so Brad was right - the 4 buy teams did not impact our RPI in a bad way. By the way isn't the Sagrin the one the NCAA uses for picking tourney teams? Quote Link to comment Share on other sites More sharing options...
billiken_roy Posted January 2, 2004 Share Posted January 2, 2004 no the ncaa uses their own formula, but experts are pretty much in agreement that the rpi is probably identical to the ncaa formula. Quote Link to comment Share on other sites More sharing options...
Basketbill Posted January 3, 2004 Share Posted January 3, 2004 but I do not see the Big 12 or Big ten if they ended up below CUSA losing out. The MVC or other 2 or three spot conference would lose out imho. But Hey if we are in who cares!! Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.