Go Back   Fighting Illini Forums > Sports > Fighting Illini Basketball

KenPom preseason rankings are up - #57 Illinois projected to go 15-15 (7-11)

Reply
 
Thread Tools
Old Nov 13, 2011, 11:14 AM   #101
Illest
Posts: 6,844
Quote:
Originally Posted by DaytonIllini View Post
If you look at the histogram the chance of winning 15 is signficantly less than winning 17. The chance to win 14 is significantly less than winning 18. I think 16 is their modal prediction. I don't know enough about statistics to properly define what I am trying to say but the data looks skewed toward the pessimistic case.
You get skewness if the expected win total doesn't land squarely on a whole number. My bigger issue with the distributions of projected records is that I don't know how much they account for uncertainty about the quality of a team in addition to uncertainty about how many games a team of a certain quality will win.
Illest is offline Reply With Quote
Old Nov 13, 2011, 11:51 AM   #102
danielb927
Orange Krush Class of 2013
danielb927's Avatar
Location: Stanford, CA
Posts: 4,377
Quote:
Originally Posted by Illest View Post
You get skewness if the expected win total doesn't land squarely on a whole number. My bigger issue with the distributions of projected records is that I don't know how much they account for uncertainty about the quality of a team in addition to uncertainty about how many games a team of a certain quality will win.
Probably not much, I would think.

__________________
Former Illini Pride President (2012-2013)
Orange Krush Class of 2013
danielb927 is offline Reply With Quote
Old Nov 13, 2011, 03:01 PM   #103
DaytonIllini
DaytonIllini's Avatar
Posts: 20,417
Quote:
Originally Posted by Illini70math View Post
The more important point to keep in mind is that this is an extremely marginal use of statistics for prediction purposes. The results cannot be more accurate than the input data, and the input data consists largely of wild guesses.

Statistical prediction works very well with certain types of data. You can predict mechanical failure rates very accurately over large data sets because the historical data you use to make your predictions is based on stable and measurable tendencies. If production standards are carefully monitored, one rivet is a lot like another. That is what makes predictive statistics work in that type of situation.

Plainly how we do this season will have a lot to do with individual performances by various players, and there is no basis for predictions of the type I mentioned above. Neither you nor I nor Weber nor Kenpom knows how Meyers Leonard will do this season, for example. But it's not just Meyers. How will Tracy Abrams do? How will Brandon Paul do? Without sufficient relevant and valid input data, predictive statistics amounts to pure guesswork in college basketball. That's why the correlations between predictions and what actually happens is not very high. Remember the predictions for how well we would do last season? Someone forgot to tell UIC...

GIGO

__________________
No nation can preserve its freedom in the midst of continual warfare.
-- James Madison
DaytonIllini is offline Reply With Quote
Old Nov 13, 2011, 07:19 PM   #104
Illini70math
Posts: 2,099
Quote:
Originally Posted by DaytonIllini View Post
GIGO
Exactly.

Funny stuff: there was a Peegs poster today pointing out that IU was ranked #24 in the RPI to start the day. Of course so were about 100 other teams (including UI), as he admitted!

Best rankings: Duquesne, Belmont, and South Alabama were ranking #7, #10, and #21 with matching 0-1 records. Stranger still, Army was ranked #22 with a record of 0-2. I think the RPI hasn't quite settled in yet!

Illini70math is offline Reply With Quote
Old Nov 15, 2011, 12:11 PM   #105
TEYPAY
Location: Springfield
Posts: 1,252
2-0 and we go from 54 to 56??????? Still early.
TEYPAY is offline Reply With Quote
Old Nov 15, 2011, 12:33 PM   #106
danielb927
Orange Krush Class of 2013
danielb927's Avatar
Location: Stanford, CA
Posts: 4,377
Quote:
Originally Posted by TEYPAY View Post
2-0 and we go from 54 to 56??????? Still early.
When you play teams ranked in the 330s out of 345, you have to really blow them out to see any move upwards in the rankings. Just playing the game alone is gonna drop you 9 times out of 10.

__________________
Former Illini Pride President (2012-2013)
Orange Krush Class of 2013
danielb927 is offline Reply With Quote
Old Nov 17, 2011, 10:21 AM   #107
TEYPAY
Location: Springfield
Posts: 1,252
Maybe it's only me (and I know it's early) but I can't help but check this and see - haven't played since Monday and down to #58! Iona - who hasn't played a game - is ahead of us......
TEYPAY is offline Reply With Quote
Old Nov 17, 2011, 10:23 AM   #108
Illini70math
Posts: 2,099
Quote:
Originally Posted by TEYPAY View Post
Maybe it's only me (and I know it's early) but I can't help but check this and see - haven't played since Monday and down to #58! Iona - who hasn't played a game - is ahead of us......
Sure, they haven't played, but they looked really good doing it!

Illini70math is offline Reply With Quote
Old Nov 17, 2011, 10:26 AM   #109
TEYPAY
Location: Springfield
Posts: 1,252
Purdue - remember them - they beat "High Point" by 2 - they are 2-0 and ranked #12???
TEYPAY is offline Reply With Quote
Old Nov 17, 2011, 12:45 PM   #110
WesterveltVictoryCigar
Location: People's Republic of Massachusetts
Posts: 8,563
Quote:
Originally Posted by TEYPAY View Post
Maybe it's only me (and I know it's early) but I can't help but check this and see - haven't played since Monday and down to #58! Iona - who hasn't played a game - is ahead of us......
They are beating Purdue late in the first half down in San Juan.
WesterveltVictoryCigar is offline Reply With Quote
Old Nov 17, 2011, 04:35 PM   #111
zoggle101
zoggle101's Avatar
Posts: 5,408
How'd Lamont "momo" Jones play for Iona? He was a big loss albeit not huge for Shaun Millers Wildcats. The only other things I know about Iona is they make the tournament most years and they share the same nickname as St. Mary's who Klee loves. Rob Jones is a killer and not like his father who really was a killer. I mentioned Jones transfer a couple of times last year iirc.
zoggle101 is offline Reply With Quote
Old Nov 18, 2011, 08:33 AM   #112
TEYPAY
Location: Springfield
Posts: 1,252
Go 3-0 and we are now down to #60??? Hopefully at this pace we can finish dead last (and go undefeated!):laugh:
TEYPAY is offline Reply With Quote
Old Nov 18, 2011, 11:12 AM   #113
Illest
Posts: 6,844
Quote:
Originally Posted by TEYPAY View Post
Go 3-0 and we are now down to #60??? Hopefully at this pace we can finish dead last (and go undefeated!):laugh:
The system bases rankings on the final score, not wins and losses. If you don't think margin of victory is an important signal about a team's quality, I direct your attention to the football team's 6-0 start.
Illest is offline Reply With Quote
Old Nov 18, 2011, 11:15 AM   #114
4LaCosaNostra
4LaCosaNostra's Avatar
Location: schaumburg
Posts: 3,442
Quote:
Originally Posted by Illest View Post
The system bases rankings on the final score, not wins and losses. If you don't think margin of victory is an important signal about a team's quality, I direct your attention to the football team's 6-0 start.
then a lot of bcs schools should be slipping-otoh- duke 1 point over belmont- purdue back to back close wins- pitt losing to LBS and michigan 5 point win over WIU.. a lot of close / almost upsets so far this year
4LaCosaNostra is offline Reply With Quote
Old Nov 18, 2011, 11:30 AM   #115
TEYPAY
Location: Springfield
Posts: 1,252
Purdue is #14!!!! Michigan is #48. Pitt is #16. I know there is a lot that goes into it. I just think it shows that early rankings (with whatever system you use) are not that meaningful. I would rather win 10 games by 3 points than win 2 by 30 and lose the other 8 (again - I know more goes into than that). In all honesty, probably early in the season, the coaches / writers polls are probably more accuate. In football, the BCS doesn't even come out until 1/3 or so into the sesaon?
TEYPAY is offline Reply With Quote
Old Nov 21, 2011, 01:11 AM   #116
Future Walk-On
Future Walk-On's Avatar
Location: Peoria, IL
Posts: 1,721
Quote:
Originally Posted by TEYPAY View Post
Purdue is #14!!!! Michigan is #48. Pitt is #16. I know there is a lot that goes into it. I just think it shows that early rankings (with whatever system you use) are not that meaningful. I would rather win 10 games by 3 points than win 2 by 30 and lose the other 8 (again - I know more goes into than that). In all honesty, probably early in the season, the coaches / writers polls are probably more accuate. In football, the BCS doesn't even come out until 1/3 or so into the sesaon?
More goes into it.....

JK. I don't look at it unless Illinois is high....then I brag to my friend from Columbia (Mizzou fan) and can't explain why we are high.
Future Walk-On is offline Reply With Quote
Old Nov 21, 2011, 08:34 AM   #117
ill07
ill07's Avatar
Location: Chicago
Posts: 3,068
Quote:
Originally Posted by danielb927 View Post
Another bit of info from Pomeroy's blog:

Quote:
So beginning this season, you’ll need to pay $20 per year to access the data on the site. What does your subscription get? Well, for starters, you should know that the main ratings page and the blog will still be freely available to the public. However, if you want to drill down into team details or see any of the other information that brings insight to your college hoops experience, you’ll need to subscribe. And while the blog will still be on the free side, much of its content will refer to stuff behind the pay wall.
Yuck. KenPom's stuff is fun to look at for sh!ts-n-giggles, but I could think of better ways to spend $20. More power to him though, if he can get people to pay... He'd be better off with ads IMO though.

__________________
And unto thee we pledge our heart and hand, Dear Alma Mater
ill07 is offline Reply With Quote
Old Nov 21, 2011, 09:06 AM   #118
illineye2
Banned
Location: Lockport il.
Posts: 263
I take about as much stock in this rating system,as any other one.SCRAP!
illineye2 is offline Reply With Quote
Old Nov 21, 2011, 09:42 AM   #119
ChicagoJoe
Banned
Posts: 635
Quote:
The system bases rankings on the final score, not wins and losses.
It doesn't. It bases the rankings on offensive and defensive efficiency, with a strength of schedule component thrown in.

Quote:
In all honesty, probably early in the season, the coaches / writers polls are probably more accurate.
Pomeroy admits as much, though by the end of the season, he's more accurate.
ChicagoJoe is offline Reply With Quote
Old Nov 21, 2011, 10:23 AM   #120
Illest
Posts: 6,844
Quote:
Originally Posted by ChicagoJoe View Post
It doesn't. It bases the rankings on offensive and defensive efficiency, with a strength of schedule component thrown in.
Ok, technically I was imprecise. Yes, there's a strength of schedule adjustment, and adjustments for where games were played, when they were played, etc. Yes, he uses defensive and offensive efficiencies as an intermediate calculation, which requires information about pace.

However, at its essence, it's a Pythagorean rating system, and the fundamental statistic for a pure Pythagorean rating system is the ratio of points scored to points allowed. That's what I was getting at.
Illest is offline Reply With Quote
Old Nov 21, 2011, 11:04 AM   #121
NV Illini 74
NV Illini 74's Avatar
Posts: 5,279
Just to show how messed up RPI ratings can be in the early going, CBS Sportsline hase the Illini at RPI #251 and a SOS of 303 (out of 343) after a 3-0 start.

http://www.cbssports.com/collegebask...ngs/rpi/index2

Pretty funny stuff. Hope it PO's our guys and they take care of business vs Richmond.
NV Illini 74 is offline Reply With Quote
Old Nov 21, 2011, 11:08 AM   #122
illynifan34
That's a winner!!
Location: Cincinnati, OH
Posts: 8,109
Quote:
Originally Posted by Illest View Post
Ok, technically I was imprecise. Yes, there's a strength of schedule adjustment, and adjustments for where games were played, when they were played, etc. Yes, he uses defensive and offensive efficiencies as an intermediate calculation, which requires information about pace.

However, at its essence, it's a Pythagorean rating system, and the fundamental statistic for a pure Pythagorean rating system is the ratio of points scored to points allowed. That's what I was getting at.
So what you're saying is that anyone running the "triangle" offense or a triangle and 2 defense will do best in his ratings?

__________________
www.afsp.org

In Mike I Trust

Aggressive

illynifan34 is offline Reply With Quote
Old Nov 21, 2011, 11:08 AM   #123
Illini70math
Posts: 2,099
Quote:
Originally Posted by Illest View Post
Ok, technically I was imprecise. Yes, there's a strength of schedule adjustment, and adjustments for where games were played, when they were played, etc. Yes, he uses defensive and offensive efficiencies as an intermediate calculation, which requires information about pace.

However, at its essence, it's a Pythagorean rating system, and the fundamental statistic for a pure Pythagorean rating system is the ratio of points scored to points allowed. That's what I was getting at.
I never trusted that Pythagorean thing. I know it's got to be wrong. We just haven't found the right triangle yet.

Illini70math is offline Reply With Quote
Old Nov 21, 2011, 11:10 AM   #124
orangekrush117
orangekrush117's Avatar
Location: Chicago, IL
Posts: 1,564
There's a team who is 0-4 ranked ahead of us. I'm pretty sure it means next to nothing at this point in the season.
orangekrush117 is offline Reply With Quote
Old Nov 21, 2011, 11:20 AM   #125
TEYPAY
Location: Springfield
Posts: 1,252
However, the 3 wins we have and the 4 losses they have do count!
TEYPAY is offline Reply With Quote
Reply


« Previous Thread | Fighting Illini Basketball | Next Thread »
Thread Tools

Forum Jump