|
Replies: 13
| visibility 51
|
CU Medallion [53754]
TigerPulse: 100%
Posts: 43551
Joined: 11/17/03
|
Boys, this RPI makes no sense to me...
3
May 26, 2023, 1:56 PM
|
|
Click on the top teams...and look at the break down of wins..Q1,Q2,Q3,Q4...W streak, Last 10 games etc. I mean, Kentucky ONLY has 36 wins and they are #3...The next team with that few wins is #16 Boston College...They are 5-5 in their last 10....Then look at the coots...2-8 in last 10.. 2 loss streak...3-11 in their last 14 ..AND STILL AT #7....
This is what happens when IN EVERY SPORT...half of your conference starts out in the TOP TEN....
|
|
|
|
 |
All-TigerNet [13838]
TigerPulse: 100%
Posts: 7829
Joined: 11/30/98
|
truth.... the initial ranking in the RPI
2
May 26, 2023, 2:04 PM
|
|
is totally subjective. I'd like to know 'who' decides the initial priming of the pump? LOL
|
|
|
|
|
|
110%er [5512]
TigerPulse: 100%
Posts: 4987
Joined: 10/21/15
|
Re: truth.... the initial ranking in the RPI
3
May 26, 2023, 2:15 PM
|
|
That's not true at all. There is no subjectivity or bias in RPI. It is just math.
|
|
|
|
|
|
110%er [6167]
TigerPulse: 100%
Posts: 5087
Joined: 8/30/17
|
Re: truth.... the initial ranking in the RPI
May 26, 2023, 3:10 PM
|
|
|
|
|
|
|
|
Legend [16902]
TigerPulse: 100%
Posts: 10768
Joined: 1/25/07
|
Re: truth.... the initial ranking in the RPI
1
May 26, 2023, 2:18 PM
[ in reply to truth.... the initial ranking in the RPI ] |
|
There is no “initial rating” in rpi. It is purely based on w/l records - yours, your opponents, and opponent’s opponents.
|
|
|
|
|
|
110%er [7470]
TigerPulse: 100%
Posts: 15404
Joined: 2/2/01
|
RPI doesn't care when you win or lose the games
May 26, 2023, 2:09 PM
|
|
NCAAT committee does
|
|
|
|
|
|
110%er [5512]
TigerPulse: 100%
Posts: 4987
Joined: 10/21/15
|
Re: Boys, this RPI makes no sense to me...
May 26, 2023, 2:20 PM
|
|
This article explains it well. There is no "initial top 10" in the RPI ratings. It is all solely based on math - there is no committee, no voters, not polls.
The rating is based on your record, your opponents' record, and your opponents' opponents' record.
There is no official RPI until after several games are played so that there is enough data to render a sensible ranking.
As this site points out, RPI has a major flaw - there is more focus on your opponents' records than there is on your own. I think that is largely why basketball has moved away from RPI and has instead moved to NET. It also explains why a team doesn't drop much this late in the season (see Coots) when they lose, because there is too much emphasis on their opponents.
https://thedatajocks.com/what-is-rpi-deficiency/
|
|
|
|
|
|
CU Medallion [53754]
TigerPulse: 100%
Posts: 43551
Joined: 11/17/03
|
YOur last paragraph explains it then...South Caroliana's
May 26, 2023, 2:36 PM
|
|
last 15 games has kept our RPI down, while our last 25 games has kept theirs UP!LOL!!
|
|
|
|
|
|
110%er [5512]
TigerPulse: 100%
Posts: 4987
Joined: 10/21/15
|
Re: YOur last paragraph explains it then...South Caroliana's
May 26, 2023, 2:38 PM
|
|
To a degree, that's absolutely correct and it's exactly how the RPI is designed...and why it's flawed.
|
|
|
|
|
|
Legend [16902]
TigerPulse: 100%
Posts: 10768
Joined: 1/25/07
|
Re: YOur last paragraph explains it then...South Caroliana's
May 26, 2023, 2:58 PM
|
|
If anyone comes up with a truly good system Vegas will have to close the spots books. AI hasn’t cracked it. It does better at stock picking than sports picking. The RPI is a better predictor of actual results than NET. It is inferior to ELO. None of which can consistently help you pick winners enough to cover the vig. Following ELO will usually put you in the upper quartile of any large season long pick ‘‘em though.
|
|
|
|
|
|
All-TigerNet [10489]
TigerPulse: 100%
Posts: 7757
Joined: 12/5/15
|
Re: Boys, this RPI makes no sense to me...
May 26, 2023, 2:25 PM
|
|
When winning & losing don’t matter, it’s no longer sport……it’s art.
The inventors of RPI would have us believe that winning and losing don’t matter for the SEC……but for everyone else it matters.
It’s brazen. They didn’t have the stat to support their BS rankings, so they invented it. BS stat supports BS ranking…..”hey everybody, look how amazing this one group is at everything”.
Uhhhhhh……BS!!!!!!!!
|
|
|
|
|
|
110%er [5512]
TigerPulse: 100%
Posts: 4987
Joined: 10/21/15
|
Re: Boys, this RPI makes no sense to me...
May 26, 2023, 2:26 PM
|
|
Again - there is no bias at all. There is nothing about conferences, there is nothing about perception, there is nothing about "where you start". It's a mathematical algorithm applied across the board to every team in an identical way.
|
|
|
|
|
|
Legend [16902]
TigerPulse: 100%
Posts: 10768
Joined: 1/25/07
|
Re: Boys, this RPI makes no sense to me...
May 26, 2023, 2:30 PM
[ in reply to Re: Boys, this RPI makes no sense to me... ] |
|
You’re simply incorrect. Please read any of the posts in this thread to understand how it works. Winning and losing are the only things that matter in rpi.
|
|
|
|
|
|
Legend [16902]
TigerPulse: 100%
Posts: 10768
Joined: 1/25/07
|
Re: Boys, this RPI makes no sense to me...
May 26, 2023, 2:29 PM
|
|
It’s a simple arithmetic formula. Quadrants records have no direct influence on your rpi ratings. That’s something looked at by the committee. SEC teams are able to stay so highly ranked because of their non conference records. Arkansas is now the top ranked because, even though they have 2 more NC losses than the next three SEC teams (rpi) they have two more conference wins. Your opponents combined record is the largest chunk of your rpi rating (50%) but each w/l is far more dilute. Kentucky is only there spots ahead of us and though they’ve lost one more game they’ve played the 1st versus 16th ranked NC schedule and the 1st versus 5th overall sos. You don’t have to like rpi as a metric but it contains zero bias or politics.
|
|
|
|
|
|
Replies: 13
| visibility 51
|
|
|
50
