replacement for hoop RPI
common sense replacement for RPI: Wins Above Bubble
I like it, rates us better.
But there is still no excuse for scheduling those bottom feeder non-conference teams
http://thebiglead.com/2016/03/09/michigan-is-getting-screwed-by-the-rpi/
March 12th, 2016 at 10:30 AM ^
You are correct, but I must stress that there is nothing wrong with scheduling a DII team. It doesn't count for you. It doesn't count against you. Scheduling D1 teams that end up with +200 RPI ratings is the problem, because those games are unfairly weighted in comparison to the more competitive portion of a team's schedule.
March 12th, 2016 at 12:33 PM ^
It matters because for two years in a row we have played less regular season games than (almost) everyone else... why on earth would this happen? Practice games are for exhibitions.
March 12th, 2016 at 10:25 PM ^
If the Div II team is replaced by a Div I tomato can, then it is better to play the Div II team. This is mentioned in the article linked by the OP. Michigan's win over Deleware State had a negative impact on Michigan's RPI, while its win over Northern Michigan did not affect it positively or negatively.
Switched out teams like Bryant, NKU and Houston Baptist with even some average Power 5 teams, the team probably wouldn't be in its current situation. Cheap wins early don't help at the end of the season. They need to really evaluate how they approach the non-conference in the future.
BTW, Michigan's RPI is at 62, up six spots, after the Indiana win.
Sent from MGoBlog HD for iPhone & iPad
March 12th, 2016 at 10:21 AM ^
Rivalry.
Sent from MGoBlog HD for iPhone & iPad
Sent from MGoBlog HD for iPhone & iPad
March 12th, 2016 at 11:43 AM ^
I won't exactly say we should have won, but I don't think either OSU or (especially) Iowa played well in those games. They were there for the taking if we could have shot our normal percentage. We went into a team-wide shooting slump at a very bad time.
When JB took over, Michigan had a 4-game deal with Oakland where we played 2 at home and 2 at the Palace. He ended it, saying it wasn't good for Michigan; Kampe added that Beilein told him that he didn't want to risk losing to OU.
It's too bad. Aside from the fact that playing local schools like Oakland and UDM is far more interesting than Bryant, it's much better for the RPI. But I guess he has reason to be worried, given the Eastern game last year.
Sent from MGoBlog HD for iPhone & iPad
Being inside the top 45 almost certainly means selection, particularly if in a power conference and if you have some quality wins. Being in the 46 to 55 range is the true bubble for power conference teams. Above that, it takes a lot of quality wins to overcome a bad RPI.
Michigan’s RPI is currently 66.
RPI is one of the reasons that, even if we talk our way into the finals of the BTT today, I still might be sitting on my hands to some extent because there are still teams that could easily fill the slots which Michigan could inhabit based on that number. It's part of the reason I've found myself watching other conference championships - either as they happen or following online - throughout the week. All this matters now.
If Michigan wins today, their RPI is going to shoot up into the 40s because not only would they get credit for the win over Purdue, but they would also add a W/L in the BTT Final against Maryland/MSU. There won't be any doubt if they can pull that off today.
too serioiusly by the committee.
Instead, I would hope that the committee uses the RPI wisely, as a first sort of your record in buckets like Top 25 RPI, 26-50, 50-100, etc. Then compare teams against each other mostly on human judgment, laying out one bubble team's Top 50 games against another bubble teams Top 50 games, (or 51-100, whatever), and then with human judgment deciding which team has the better resume.
If used in that manner, the known imprecision in the RPI might be a strength as it wouldn't be taken as being an ordering system when it comes down to who's in and who's out.
I think even a loss to Purdue on a neutral court will give our RPI a small boost, especially if they go on to win the BTT as that would probably put Purdue in the top 10 of the RPI or very close.
But a win today would definitely seal the deal for us. It may even earn Michigan a chance to avoid the First Four in Dayton.
Beat Purdue and we'll take our chances with the committee
March 12th, 2016 at 10:27 AM ^
March 12th, 2016 at 11:08 AM ^
Most places have Butler as a lock and Michigan either barely in or out of the field. Why, exactly?
Michigan vs Butler
RPI: 55 vs 56
Record: 21-11 vs 21-10
SOS: 54 vs 83
NCSOS: 209 vs 259
Record vs Top 25: 3-6 vs 3-5
Record vs Top 50: 4-10 vs 4-8
Record vs Top 100: 4-11 vs 5-8
Tell me exactly how Butler's resume is better than Michigan's.
March 12th, 2016 at 11:38 AM ^
Record vs Top 100: 4-11 vs 5-8
38% > 26% ???
Isn't Michigan now #55 RPI, so this article is now irrelevant.
I say just win today and set aside all doubt.
March 12th, 2016 at 11:56 AM ^
Yeah, the article is three days old. Came out before NW and IN wins.
March 13th, 2016 at 12:52 AM ^
Butler's #39 at kenpom, Michigan's #55.
That's not an enormous difference, but when you're picking out the top 50 teams it's not insignificant.
And I suspect the reason for the difference is the relative heaviness of the losses. Before Thursday, the only teams that had beaten Butler by 10+ were Xavier, Villanova, and Miami, all three-seeds or better.
Michigan's lost by 10+ to Iowa, Wisconsin, Ohio State, Michigan State, Indiana, SMU, Purdue, Connecticut and Xavier. Three or four of those will be three-seeds or better; one of them isn't even under consideration (and I'm not talking about SMU, who would probably be about a six if they were eligible).
March 12th, 2016 at 11:34 AM ^
I heard the comittee is using other metrics other than the RPI. I hope this is true as the other rankings have us ranked higher than RPI. It is pretty stupid the diffference between 300 and 200 when in reality there is little difference.
March 13th, 2016 at 12:36 AM ^
I don't think any sane person would rely entirely on a metric that would look at that Cincinnati/UConn game from the other day and treat it exactly the same as a 30-point bludgeoning. Whoever wound up winning, those were two virtually equal teams on the day.
The committee talks about the RPI because it doesn't weight margin of victory and they don't want to seem to be encouraging teams to run up the score, and because it's a way of trying to leverage teams into scheduling fewer cupcakes. But there's no evidence they make much use of it as a metric once they get into the room, except as a means of sorting opponents (record vs. RPI top 50 or 100, for example).
If you want to test this, try an experiment tomorrow. Set up a bracket using kenpom or massey--just s-curve your way down the rankings, paying attention to avoiding repeat matchups etc. and making sure you get the autobids in--then set up a second bracket the same way but using RPI. One of them will be a lot closer to the final product, and it won't be the one that uses RPI. (In the past, for what it's worth, the headline Massey and Sagarin rankings have been closer than the more predictively accurate power ranking, or kenpom.)
I think the real problem for Michigan isn't the soft nonconference schedule, it's too many bad losses...not the kind where you lose to a bad team, but the kind where you get beaten badly by a pretty good team and observers are left in no doubt about who was better. I don't think there's another team left on the bubble, or getting an at-large, with as many double-digit losses as Michigan's had.
---
Looking that up now:
- VCU 1
- South Carolina 3
- Florida 5
- Georgia 5
- Temple 5
- Syracuse 6
- Florida State 7
- Pittsburgh 7
- USC 7
- Oregon State 8
- Michigan 10
I know there's another side to this, but that's the case for the prosecution. Ten losses, period, has you close to the cut line at best. Ten losses by double digits? It doesn't show up in the RPI but you can bet somebody in the committee room is pointing it out.
March 12th, 2016 at 11:45 AM ^
There needs to be some kind of cutoff for low-RPI teams. Pretending that the #275 team is markedly better than the #300 team is dumb. Just cut it off after like 200 or 225 and treat all the rest equal.