Admittedly, I can't recreate BPI calcs, but it has to be about the worst

[mikeiastat]

ranking system possible. Can anyone explain how this piece of crap works. It doesn't seem to care much at all if you win or lose, or the quality of teams you play. Just your cumulative point spread. I can't respect a rating system that puts ISU behind teams with (by their OWN SOS STANDARD) worse records against astronomically worse SOS. Worthless if you ask me. Anyone who puts stock in the BPI either works for ESPN or is spoon fed participation ribbon BS. ISU according to their SOS calc has far and away the toughest schedule in the country yet Louville with a 117 SOS is ahead of thme with a similar record?

Pitt is an absolute joke in that rating. Wisconsin, Creighton, Kentucky, UCLA, all same or worse records agains weaker schedule. and Iowa plays a considerably Weaker schedule and lost more games yet almost 10 spots higher. Makes a ton of sense.

Bag on RPI all you want, but this is ridiculous. Sagarin at least makes sense. ESPN is Putting lipstick on a giant Pig in the BPI so it can have propritary control for insider access. I've never been more convinced of anything. Drop it already.

 

[hexodat64]

Lol BPI sucks. Right now it has Oregon at a BPI of 20 and a projected seed of 5. ESPN's own Joe Lunardi doesn't even have them in the tournament or even close

 

[HFCS]

In football I kind of like the rankings that use margin of victory, because there's only 12-13 games and 8-9 of them are conference. There's really only about 25 significant non conference games in the entire season to judge conferences against each other.

In basketball there are enough games and non-conference match ups that just winning against quality of schedule is generally enough to put together a solid ranking (taking home/road/neutral into account helps). Take tonight's game, when we were up 12 and the game was iced...kenpom, bpi and sagarin would have rewarded us for hitting a couple 3s or shutting out Texas down the stretch. The game was over, securing quality wins is how you advance in tournaments, it's all that should matter.

For gambling or statistical analysis purposes sure kenpom and sagarin are great for college hoops, but for a ranking teams according to merit, I'm not sold on them.

 

[atlantacyclone]

Well, it didn't seem so bad when ISU was ranked #1!

 

[mikeiastat]

Well, it didn't seem so bad when ISU was ranked #1!

Actually I sitll didn't like it. I though it was still unpredictable and only being pushed for propritary reasons. It has never mad sense to me to reward close losses more than 7 point losses and all that. A loss is a loss. The only think that intrigues me is the whole adjusting for missing key players, but even that seems arbitrary to me. I said back then it was garbage, didn't matter and noone would care if ESPN didn't spoon feed it too the uneducated masses. I believe this much more so now. Look at it even without ISU being snubbed it makes almost zero sense.

Its anomalies are not explainable and have no chance of being fixed and almost nothing of it seems right. It seems to completely contradict itself. If someone could explain how you mitigate the completely messed up results you are seeing right now in the top 20 I'd love to hear it, but It isn't even transparent. I am 100% certain its only an ESPN gimmick to get something that they have control over even it it's completely wrong. But hey whatever gets more people to be "insiders"

 

[TurbulentEddie]

The only (popular) rating systems that I know that have enough documentation/history to be reproducible are RPI (25% win pct + 50% opp win pct + 25% opp opp win pct), Sagarin ELO (same thing chess has used for a long time), and KenPom (off/def points per possession adjusted for opponent, Pythagorean expectation similar to sabermetrics). No idea about BPI other than that they claim this and that are taken in to account. I think that's part of why no one really takes it seriously outside of ESPN.

 

[HFCS]

The only (popular) rating systems that I know that have enough documentation/history to be reproducible are RPI (25% win pct + 50% opp win pct + 25% opp opp win pct), Sagarin ELO (same thing chess has used for a long time), and KenPom (off/def points per possession adjusted for opponent, Pythagorean expectation similar to sabermetrics). No idea about BPI other than that they claim this and that are taken in to account. I think that's part of why no one really takes it seriously outside of ESPN.

I'm not sure you can say KenPom is a valid ranking system, just because it's used as great statistical analysis by a lot of coaches and gamblers doesn't validate it as a ranking of merit. According to KenPom's own SOS ranking there are 11 teams ranked ahead of ISU who have lost more games against a schedule KenPom says is easier than ISU's. Often teams have 3-5 more losses than ISU against a schedule that is dozens of spots easier. That's a very worthless ranking. There's a little of it in Sagarin and BPI but not nearly as pronounced.

It's a mess, it's a mess every year. A few years ago Wisconsin was in the top 5 and weren't even ranked in the AP poll.

 

[Rhoadhoused]

Just so everyone knows this guy is the biggest RPI fan I've ever seen. BPI is vastly superior.

 

[CyTwins]

Just so everyone knows this guy is the biggest RPI fan I've ever seen. BPI is vastly superior.

And you hate RPI more than anyone I've seen. BPI has a lot of flaws too, everyone just has their own preference

 

[HFCS]

Just so everyone knows this guy is the biggest RPI fan I've ever seen. BPI is vastly superior.

If I had to just use one computer ranking and nothing else I'd use BPI. What I really think is the best indicator personally is an average of half AP Poll, and half computer average of ratings like Sagarin ELO that don't use margin of victory...same thing the BCS ranking does. I think the BCS ranking is the best ranking in sports because of the 1/3 computers correcting brand bias in the voter polls.

Prior to Texas win we were #17 AP and #16 Sagarin ELO. It makes sense to me that we're 16.5...also matches our BPI (but 6 spots lower than our RPI). It doesn't make sense to me at all that we were #27 as Kenpom had us before Texas. RPI is a DRASTICALLY better tool to seed the NCAA tournament than Kenpom is, Kenpom is a joke in determining who has won the most quality games against a quality schedule.

You advance in the NCAA tournament by beating good teams by 1, not with any sort of style points. That's why RPI is used and not polls that would have bumped us up from turning our 12 point lead with the game over last night into a 20 point lead...or take away value of the win if Texas had sank a few garbage 3s and we almost lost.

 

[Rhoadhoused]

And you hate RPI more than anyone I've seen. BPI has a lot of flaws too, everyone just has their own preference

I just acknowledge it has few inputs and many flaws. It does a worse job than both BPI and Sagarin at ranking teams. It is far far too simple and doesn't take many things into account. Everyone agrees on this.

However, RPI is what the committee uses so we have to live with it. Just don't try to tell me it's a great system because it isn't. I could find a ton of outliers with RPI if I wanted. You can with any system.

 

[mikeiastat]

Just so everyone knows this guy is the biggest RPI fan I've ever seen. BPI is vastly superior.

Actually not that big of an RPI fan. Sagarin is my choice for actually making sense, but It will never be able to be used by the NCAA because of MOV. I specifically state that there are glaring flaws in the RPI. But they are well documented and easily mitigated. I just explain to people that cut on it how they are completely wrong. There's a difference. How can BPI be superior when by their own SOS calcs ISU has the toughest schedule by a pretty good margin after KU, yet there are what 7 of the 16 ahead of us, that are teams with the same or worse record that are above ISU with often a very very much weaker schedule? Those aren't anomalies, thats about a 44% fail rate. I'm mostly just looking for a solid explanation, but

Yup sounds vastly superior to me.

 

[Rhoadhoused]

Actually not that big of an RPI fan. Sagarin is my choice for actually making sense, but It will never be able to be used by the NCAA because of MOV. I specifically state that there are glaring flaws in the RPI. But they are well documented and easily mitigated. I just explain to people that cut on it how they are completely wrong. There's a difference. How can BPI be superior when by their own SOS calcs ISU has the toughest schedule by a pretty good margin after KU, yet there are what 6 teams with the same or worse record that are above ISU with often a very very much weaker schedule?

Yup sounds vastly superior to me.

I just see you going on long long rants all the time about how RPI is this great system. That post sounds reasonable though. And, again, if you want to point out outliers of each system trying to invalidate it we will be here all day.

RPI simply doesn't consider enough variables to be a great ranking tool. The NCAA loves it because you can't use MOV and teams won't run up the score for a better ranking. That's about it.

 

[HFCS]

Actually not that big of an RPI fan. Sagarin is my choice for actually making sense, but It will never be able to be used by the NCAA because of MOV. I specifically state that there are glaring flaws in the RPI. But they are well documented and easily mitigated. I just explain to people that cut on it how they are completely wrong. There's a difference. How can BPI be superior when by their own SOS calcs ISU has the toughest schedule by a pretty good margin after KU, yet there are what 6 teams with the same or worse record that are above ISU with often a very very much weaker schedule?

Yup sounds vastly superior to me.

Is there a reason to use RPI over Sagarin ELO? Wichita State is #1 and Kansas is #6, but outside of that its top 25 is more logical than RPI, BPI, Kenpom, and about any other popular computer ranking out there. Even Wisc and Wich St you can argue why they'd be that high. There's no way to argue ISU is #26 as of today which is what the favorite worthless computer ranking of Joe Fan has us.

 

[mikeiastat]

I just acknowledge it has few inputs and many flaws. It does a worse job than both BPI and Sagarin at ranking teams. It is far far too simple and doesn't take many things into account. Everyone agrees on this.

However, RPI is what the committee uses so we have to live with it. Just don't try to tell me it's a great system because it isn't. I could find a ton of outliers with RPI if I wanted. You can with any system.

Expain to me the Inputs and I may be able to agree with you, but Just

MORE INPUTS = EQUALS BETTER SYSTEM

has to be about the worst an intelligent person could put out there. I hope you have something better than that. MORE arbitrary crap does not make the system better. I can make a system that tells me how many banana's the team eats on average add it to BPI and bingo call it a better system? The results of May cumulatively put ISU at its right spot, but relative to its own data, it doesn't make sense. Its ESPN propaganda pure and simple. They are looking to make a revenue stream out of people dumb enough not to see it. But hey feel free to follow that bag of garbage right up through when they make it an "insider" segment. Congrats on being duped. Name one source national source that watches BPI outside of ESPN?

 

[bawbie]

These threads are always important to remind me that most people are really bad at math, statistics and logic.

Thanks.

 

[Rhoadhoused]

Expain to me the Inputs and I may be able to agree with you, but Just

MORE INPUTS = EQUALS BETTER SYSTEM

has to be about the worst an intelligent person could put out there. I hope you have something better than that. MORE arbitrary crap does not make the system better. I can make a system that tells me how many banana's the team eats on average add it to BPI and bingo call it a better system? The results of May cumulatively put ISU at its right spot, but relative to its own data, it doesn't make sense. Its ESPN propaganda pure and simple. They are looking to make a revenue stream out of people dumb enough not to see it. But hey feel free to follow that bag of garbage right up through when they make it an "insider" segment. Congrats on being duped. Name one source national source that watches BPI outside of ESPN?

[h=4]System comparison[/h]How is the College Basketball Power Index by the ESPN Stats & Information group) different than RPI or other advanced rating systems like Kenpom.com and Sagarin? Here is how the included elements compare to other systems.

Includes	RPI	BPI	Sagarin	Kenpom
Scoring margin	No	Yes	Yes	Yes
Diminishing returns for blowouts	No	Yes	Yes	No
Pace of game matters	No	Yes	No	Yes
Home/Neutral/Road	Yes	Yes	Yes	Yes
SOS beyond Opponent's opponents' W-L	No	Yes	Yes	Yes
All wins are better than losses (before Opp Adj)	Yes	Yes	No	No
De-weighting games with missing key players	No	Yes	No	No

 

[mikeiastat]

I just see you going on long long rants all the time about how RPI is this great system. That post sounds reasonable though. And, again, if you want to point out outliers of each system trying to invalidate it we will be here all day.

RPI simply doesn't consider enough variables to be a great ranking tool. The NCAA loves it because you can't use MOV and teams won't run up the score for a better ranking. That's about it.

Reading comprehension is your friend. I've never said RPI is a great system. It works better than almost any current system for many reasons. admittedly some political and monetary, but I am constantly pointing out its flaws. In fact am glad to explain its flaws to anyone who doesn't understand it. I'll also explain to stupid people that are up in arms that schedule strength can actually sway the RPI by 75% that they are dumb. And don't understand statistics and Math enough to intelligently discuss the topic. The other problem is that there isn't another system that works better for what it needs to do.

I do believe it is better than BPI, because BPI is adding in too much that has less and less to do with actual wins and losses. Who gives a **** if you you were 3 pts or 6 points worse than MSU, you were worse. Too big of a reward for mucking up games and just because MSU arbitrarily didn't step on the throat of some losers and did so with others. BPI rankings are showing it by not even matching it's own data.

Do you really believe that Pitt should be 13 in the country? or Iowa should be above MSU. This thing is supposed to take into account and adjust games where a team is missing key players.

 

[mikeiastat]

Is there a reason to use RPI over Sagarin ELO? Wichita State is #1 and Kansas is #6, but outside of that its top 25 is more logical than RPI, BPI, Kenpom, and about any other popular computer ranking out there. Even Wisc and Wich St you can argue why they'd be that high. There's no way to argue ISU is #26 as of today which is what the favorite worthless computer ranking of Joe Fan has us.

This I think I can get on board with, I think the NCAA would decide but not "say outloud that RPI serve an additional purpose in greasing the wheels of Scheduling in the NCAA. But yes I think Sagarin IS the best system out there, and don't understand why they wouldn't use ELO Chess. In fact in MOV calcs his others are very good too.

I also have Mixed feelings on the whole Wichita State thing. They are vastly different resume's and I can't begrudge any system that weights the intangible of actually winning all those games differently. I personally think KU would beat WSU 7/10, but who's to say comparing those two is a study in contrasts really.

What I don't get and I'm honetstly looking for discussion and explanation, how one can respect a system that doesn't even match its own primary data. I don't see the love for BPI outside of the benefactors.

I agree that you can't just point to OUtliers. But as I explained. Its not even matching its own data at about a 44% clip in its own top 20. That isn't about OUtliers its about a big failure rate.

 

[Rhoadhoused]

Reading comprehension is your friend. I've never said its a great system. I am constantly pointing out its flaws. In fact am glad to explain its flaws to anyone who doesn't understand it. I'll also explain to stupid people that are up in arms that schedule strength can actually sway the RPI by 75% that they are dumb. And don't understand statistics and Math enough to intelligently discuss the topic.

I do believe it is better than BPI, because BPI is adding in too much that has less and less to do with actual wins and losses. Who gives a **** if you you were 3 pts or 6 points worse than MSU, you were worse. Too big of a reward for mucking up games and just because MSU arbitrarily didn't step on the throat of some losers and did so with others. and their rankings are showing it by not even match it's own data.

Do you really believe that Pitt should be 13 in the country? or Iowa should be above MSU. This thing is supposed to take into account and adjust games where a team is missing key players.

That's a fundamental difference between me and you. I think MOV matters. How you actually played matters. Not sure what the argument is for ignoring how a team actually place, its pace, its efficiency, ect and just looking exclusively at Ws and Ls.