Your winning percentage is equally important as your opponents opponents winning percentage and half as important as your opponents winning percentage.
That's enough for me.
ISU #1 in BPI
- Thread starter Althetuna
- Start date
Your winning percentage is equally important as your opponents opponents winning percentage and half as important as your opponents winning percentage.
That's enough for me.
They still are instructed to employ the philosophy that MOV doesn't matter. Go by MOV and then watch teams playing their best players for 40 min. shooting 3's with 4 min left up by 40. MOV shouldn't matter to a large extent. Say every game for a data set every game is won by 1 pt then some game gets won by 2 pts. Is that twice as good of a win?
How do you evaluate the margin of victory, when some teams have some class and don't dunk the ball with less than a minute left up by 6, while KU thugs don't mind doing so?
That's why my favorite is kenpom. Doesn't matter how much you win by, just how good you play. How efficiently you play.
Of course RPI is what is used, but IMO that is a terrible predictor of a team's worth.
Also, KenPom and BPI both diminish the importance of large MOVs, so running up the site against weaker teams doesn't help much.
Yep. They also factor in how much better you played vs that team compared to how other teams play. Not just that you beat up a crappy team.
I see nothing wrong with this. This is how many SOS calculations work. All that data and the "75% that is controlled by other teams" at the end of the year Really only varies from best to worst about 20% of your score. It would be Impossible to have a .0000 Sos and it would be impossible to have a 1.0000 SOS. the range is typically in the area of better than .4000 and around .6000. for a range of just .2000 so it needs to be weighted. Your own winning percentage very realistically could be either .0000 or 1.0000. Technically you could even have a better than 1.0000 winning percentage in most SOS calcs when you add in the road premium. The other two components that is completely impossible. In fact over the entire year it is nearly impossible for the range to go below .4000
Its weighted high because the range is very low.
Another way to put it is that every one of your wins are worth roughly 15 or more wins by your opponent and more than 150 wins by your "opponents' opponents" Most people who make the 75% arguement don't understand it fully. While the 75 percent weighting seems high without understanding it, it needs to be put into perspective that the range that is ever possible for those 2 components seldom varies more that 10 percent from a .500 record.
Last edited:
I'm actually happy, for some reason, to see someone defend RPI as vehemently and cogently as mikeiastat. I still disagree with him, because I think RPI is pretty clearly an inferior metric for basically any use. But he's doing a good job here.
edit: I take back that first and last sentence. The cogency waned the more I read.
edit: I take back that first and last sentence. The cogency waned the more I read.
Last edited:
Doesn't kenpom only measure efficiency? Wins and losses are not a factor at all?
While I see the merit in kenpom, I don't get how a metric that doesn't factor in the very reason anyone plays/watches the games can be valid.
I appreciate the comment, but Bolded I completely disagree. I'm not vehemently defending it as a predictor. I just don't think its as bad as many who listen to the justifiably biased ESPN talkinghead, think it is.
As a predictor I like Sagarin as he was a pioneer and BPI is clearly doing a better job right now, but could be gamed if given enough data, time and use. I think that it over weighs efficiency numbers, because you can be really really efficient against really crappy teams. Like Iowa was last year, then just bog down the really good teams that will pick them apart, but because you bog down a KU or an OSU last year doesn't mean you have the closers to finish the game when it matters.
RPI is inferior when trying to predict one specific game, but that's not what its meant to do. It builds in other rewards for Scheduling and overall body of work that speak to overall quality of team. Just because you can beat one team because of matchups doesn't mean that you are a better team overall than that team. You can just match up well against one team and get your *** kicked you 10 others because you don't match up well anywhere else.
I see a lot of merit in valuing a win by ISU against KU when ISU plays (Or Iowa playing Mich St, Michgan and OSU) 9 tough games during that month. I think that win should be worth more when you combine that with a tougher schedule.
Then consider a Memphis team that beats KU in the past where the old Memphis teams could dedicate time in practices to an upcoming KU game for up to 3 weeks because they didn't have to worry about Playing Iowa the tueday before and OSU or KState the week after.
You can't tell me that when hypothetical Memphis is practicing before the Ark State game that they don't get to devote some extra time to what they are going to do when they Play KU 2 weeks later.
Last edited:
Depends on if you are measuring how good a team is or will play against another vs how good their resume is.
Look at two teams that have very similar resumes and you need to rank them. Just look at their Kenpom ratings and see if there is a large gap.
So for what purpose do you think RPI is the best metric? There are other, better systems that also don't use margin of victory.
But it's acceptable to game the RPI (for example, by playing DII and DIII teams OOC). Ok. At least for the other systems, it's not mostly who you play, but mostly how you play.
http://m.espn.go.com/general/blogs/blogpost?blogname=collegebasketballnation&id=64576&src=desktop
Ugh, mobile not working well.
Last edited:
This ^ but as such there won't be much of a gap between them in the RPI either and RPI follows the mandates put out by the NCAA guidelines for ranking systems.
It may well be the better metric for judging quality of team as beating 10 of the same kind of team based on how you match up well with them doesn't do you much good when you get exposed by the rest of the field. But IMO RPI IS a better metric for comparing overall team resumes because it puts a premium on wins in the middle of a meat grinder as opposed to the same win when you can focus on the beast on your schedule for 2 or 3 weeks.
No its not acceptable to game the RPI, and in the past the committee has adjusted the RPI for teams that are clearly gaming the system. They are known and well documented flaws that can be mitigated by making adjustments. I was only saying that all systems will be gamed in time.
As far as mostly who you play, I still say that you don't understand RPI. That is not true at all. I've already explained this if you don't understand my explanation I don't know how to help you but do some research and take a statistics class.
Last edited:
So you think RPI is the best metric for comparing overall team resumes? Ergo, you think KU has the best resume in college basketball right now (or at least as of this morning)?
Last edited:
So if you want to play that game, you really think that ISU has the best shot at being the National Champ? Outliers are fun.
We'll see if it becomes clear that KU has too many gamey teams on their resume at the end of the year, then we have an outlier.
Will the ncaa bounce them down like they have MWC teams in the past that have been able/lucky enough to run the same outliers into their schedule? No because they are a blue blood. But don't play the outlier game this early in the season because I think only the biggest ISU homers believe we are the odds on favorite for the NC.
When did I ever say anything that would lead you to this strawman? I think ISU has, maybe, the 10th or 12th best shot at being the National Champ.
Last edited:
The amaising BPI has ISU at #1, so compare apples to apples. YOu never said that just like I never said that KU has the best resume. It was a shortcut Ergo Rebuttal to your Ergo.
What's your metric of choice, Don't even come back with "use the voters" as that is even dumber than switching to using the BPI.
Last edited:
I like BPI better than RPI as a resume tool, but I don't like either as a predictor. And I'm going to take back that nice post I made about you earlier, because you are getting far less cogent.
Resume does not equal national championship odds. ISU has a #1 seed resume as it stands.
