View previous topic :: View next topic |
Author |
Message |
pilight
Joined: 23 Sep 2004 Posts: 67487 Location: Where the action is
Back to top |
Posted: 09/22/19 3:25 pm ::: About old RPI data |
Reply |
|
The NCAA has an archive of RPI here: https://extra.ncaa.org/solutions/rpi/SitePages/Home.aspx
One thing you'll notice with a little searching is that they have men's basketball RPI going all the way back to the introduction of the formula in 1979. Women's basketball, OTOH, has nothing archived prior to 2002. The NCAA has used RPI for WCBB since the 1982-83 season. So where's the data?
Back in the spring I tweeted to @NCAAResearch. They directed me to David Worlock, the NCAA Director of Media Coordination/Statistics and Media Coordinator for March Madness. I've been emailing back and forth with him and Rick Nixon, NCAA Associate Director of Media Services for WCBB, trying to determine what the deal is. Turns out, they don't know where the data is. Nobody can find any official WCBB RPI prior to 2002.
After poking around online (and using the Internet Wayback Machine), I've been able to scrounge RPI lists from third party sites going back to the 1995-96 season. Now I've hit a wall.
So I'm asking my fellow RebKellians, where can I find WCBB RPI for seasons prior to 1996? It must have printed or published somewhere!
_________________ The truth is like poetry
Most people hate poetry
|
|
pilight
Joined: 23 Sep 2004 Posts: 67487 Location: Where the action is
Back to top |
Posted: 09/30/19 8:24 am ::: |
Reply |
|
Nothing, huh?
I've got some other feelers out, but if all else fails I'm working on a spreadsheet to calculate RPI. That will mean lots of data entry. My one good hand winces at the thought.
_________________ The truth is like poetry
Most people hate poetry
|
|
myrtle
Joined: 02 May 2008 Posts: 32341
Back to top |
Posted: 09/30/19 1:30 pm ::: |
Reply |
|
have you looked into a voice program for data entry?
_________________ For there is always light,
if only we’re brave enough to see it.
If only we’re brave enough to be it.
- Amanda Gorman
|
|
PUmatty
Joined: 10 Nov 2004 Posts: 16516 Location: Chicago
Back to top |
Posted: 09/30/19 5:44 pm ::: |
Reply |
|
I'd certainly be willing to help with data entry. I bet others here would be as well.
|
|
ClayK
Joined: 11 Oct 2005 Posts: 11402
Back to top |
Posted: 09/30/19 6:43 pm ::: |
Reply |
|
PUmatty wrote: |
I'd certainly be willing to help with data entry. I bet others here would be as well. |
I will too ...
_________________ Oṃ Tāre Tuttāre Ture Svāhā
|
|
pilight
Joined: 23 Sep 2004 Posts: 67487 Location: Where the action is
Back to top |
Posted: 09/30/19 6:59 pm ::: |
Reply |
|
These are so not the answers I was looking for...
_________________ The truth is like poetry
Most people hate poetry
|
|
Howee
Joined: 27 Nov 2009 Posts: 15837 Location: OREGON (in my heart)
Back to top |
|
pilight
Joined: 23 Sep 2004 Posts: 67487 Location: Where the action is
Back to top |
|
TechDawgMc
Joined: 12 Aug 2010 Posts: 405 Location: Temple, TX
Back to top |
Posted: 10/01/19 11:42 am ::: |
Reply |
|
Somebody who was on the old wbball-l discussion list used to run a website that included RPI and several other evaluations. I think he included the NCAA's numbers. I just can't remember who it was.
Maybe Clay or some of the other veterans from there can do so. I want to say it was one of the guys from La. Tech.
If you have any idea how to contact Rob Polinsky who ran that list, he might be able to tell you
|
|
pilight
Joined: 23 Sep 2004 Posts: 67487 Location: Where the action is
Back to top |
Posted: 10/02/19 9:31 am ::: |
Reply |
|
Putting together an RPI spreadsheet was not as tough as I thought.
Found a formula on an excel board to calculate opponent records which was easy enough to adapt to also do opponent's opponent's records. Tweaked it so it wouldn't include the wins & losses to the alpha team, like RPI, and the hardest part was complete.
Next I needed a list of teams. Started with the men's 1995 RPI list, which had 303 teams. I knew not all of those had WCBB, so I went to http://web1.ncaa.org/stats/StatsSrv/careersearch . On the right under School/Sport Search I left School as all, set Year 1994-95, Sport Women's Basketball, Division I. Then I identified the schools that were on the men's list but didn't come up. A few have never had WCBB, a couple had their WCBB programs at lower divisions, and some had discontinued their programs after the AIAW folded and didn't pick the back up until the late 90's. I removed those that didn't have D-I WCBB programs. Lots of name changes. I left schools as they were called in 1995.
I tested with some simulated game data and it appeared to work properly. All it needs now is some actual data. Here's what it looks like without any:
The next question is how many games should I plan for. I guessed 5000, but further research suggests that's high. The NCAA reported 5520 games last season, but that number includes postseason tournaments. With 50+ fewer teams and the shorter schedules of 25 years ago I'd guess the number will be closer to 4000.
The other issue to address before uploading to Google Docs, so that y'all can help, is how to avoid games being double entered. If someone inputting Akron games and someone inputting Dayton games both add a game in which they played each other it would throw the numbers off.
_________________ The truth is like poetry
Most people hate poetry
|
|
Coyotes
Joined: 28 Jan 2018 Posts: 1468
Back to top |
Posted: 10/02/19 5:41 pm ::: |
Reply |
|
What about a rule where you only enter the game if that team is first in the alphabet? So for example, for Dayton vs. Akron, you would only enter the result when you were submitting Akron’s results.
It’d also be good to have a master list of teams that people can shade green once said team is done.
|
|
pilight
Joined: 23 Sep 2004 Posts: 67487 Location: Where the action is
Back to top |
Posted: 10/02/19 5:58 pm ::: |
Reply |
|
Coyotes wrote: |
What about a rule where you only enter the game if that team is first in the alphabet? So for example, for Dayton vs. Akron, you would only enter the result when you were submitting Akron’s results.
It’d also be good to have a master list of teams that people can shade green once said team is done. |
I don't know that everyone will enter them by team. It seems the most logical to me, but someone else might think some other way is better.
My thought was to only enter a team's losses when doing a school. That would have the extra benefit of screening out nearly all games against non D-I opponents.
_________________ The truth is like poetry
Most people hate poetry
|
|
pilight
Joined: 23 Sep 2004 Posts: 67487 Location: Where the action is
Back to top |
Posted: 10/03/19 7:56 am ::: |
Reply |
|
The 1994-95 sheet is live HERE
Rules are simple
Enter games from the 1994-95 regular season and conference tournaments. Do not enter any games from post season tournaments (NCAA, WNIT, WBI, AIAW, etc). The winning team goes in Column N. labelled "Team Win". The losing team goes in Column O, labelled "Team Lose". Each game goes on one row.
When entering games, choose a school and enter ONLY THE GAMES THAT TEAM LOST. This will avoid double entry of games. When all a teams losses are entered, go to page three of the sheet and mark the column next to that school with an x.
When typing teams names, they must match the way they are listed in Column A. Capitalization does not matter, but spelling and punctuation do. If columns Q&R both come back blank or with all 0's there's an issue with the winning team name, if Columns S&T do so there's an issue with the losing team name. The second page of the sheet lists some of the common errors, with the correct version first.
If you've got questions or problems contact me.
I've gotten the ball rolling by typing in all the losses for Pan American and Connecticut
_________________ The truth is like poetry
Most people hate poetry
|
|
pilight
Joined: 23 Sep 2004 Posts: 67487 Location: Where the action is
Back to top |
Posted: 10/09/19 7:36 pm ::: |
Reply |
|
If you have a keen eye, you may have noticed that my formulas for opponent's record and opponent's opponents record are not producing the expected results on the uploaded spreadsheet. Not to worry, with some help from the good people at the excel help forum I've got the correct equations. When all the game data is input, I'll reupload with the corrections.
Thanks to everyone who's helping. I've been mostly concentrating on the teams that lost a lot. Still got 180-ish schools to do!
_________________ The truth is like poetry
Most people hate poetry
|
|
pilight
Joined: 23 Sep 2004 Posts: 67487 Location: Where the action is
Back to top |
Posted: 10/10/19 2:04 pm ::: |
Reply |
|
I had always thought a teams' record wasn't counted in opponents opponent's record for RPI. Going 30-0 is reward enough without counting as a 30 30-0s in your OOWP. Today's research indicates I may have been mistaken. It's funny, because finding a way to filter a team's own record from OOWP was the hardest part of developing an RPI spreadsheet and now it appears I needn't have bothered.
_________________ The truth is like poetry
Most people hate poetry
|
|
pilight
Joined: 23 Sep 2004 Posts: 67487 Location: Where the action is
Back to top |
Posted: 10/11/19 5:07 pm ::: |
Reply |
|
We're over halfway, in terms of teams. Probably more in terms of games.
_________________ The truth is like poetry
Most people hate poetry
|
|
pilight
Joined: 23 Sep 2004 Posts: 67487 Location: Where the action is
Back to top |
Posted: 10/12/19 9:11 am ::: |
Reply |
|
pilight wrote: |
I had always thought a teams' record wasn't counted in opponents opponent's record for RPI. Going 30-0 is reward enough without counting as a 30 30-0s in your OOWP. Today's research indicates I may have been mistaken. It's funny, because finding a way to filter a team's own record from OOWP was the hardest part of developing an RPI spreadsheet and now it appears I needn't have bothered. |
I'm learning so much about RPI during this project!
We all know the basic formula. To find Team A's RPI you add their winning percentage plus 2x their opponents winning percentage plus their opponents opponents winning percentage then divide by four. Seems straightforward enough.
Winning percentage is easy enough. Wins/(Wins + Losses). Basketball doesn't even have ties to complicate matters.
Opponents winning percentage seems easy enough. Same formula except you take out their results against Team A, right?
WRONG! Teams play different numbers of games. The NCAA, in an effort to make each opponent count equally despite different length of schedule, doesn't figure OWP that way. Instead they average the winning percentages of each of Team A's opponents. They do OOWP the same way.
I also found the NCAA news bulletin that announced the introduction of RPI:
https://web.archive.org/web/20120324221807/http://fs.ncaa.org/Docs/NCAANewsArchive/1981/19810215.pdf
It also includes a FAQ section about governance of women's programs during the transition from the AIAW, a letter from then Loyola assistant "Don" Bruno (better known as Doug Bruno) suggesting the rim be lowered for WCBB, and an article by Mel Greenberg suggesting the AIAW might still have a role to play even after the NCAA's move to sponsor women's championships.
_________________ The truth is like poetry
Most people hate poetry
|
|
pilight
Joined: 23 Sep 2004 Posts: 67487 Location: Where the action is
Back to top |
Posted: 10/15/19 2:28 pm ::: |
Reply |
|
pilight wrote: |
We're over halfway, in terms of teams. Probably more in terms of games. |
Down to about 75 teams left
_________________ The truth is like poetry
Most people hate poetry
|
|
pilight
Joined: 23 Sep 2004 Posts: 67487 Location: Where the action is
Back to top |
|
pilight
Joined: 23 Sep 2004 Posts: 67487 Location: Where the action is
Back to top |
|
myrtle
Joined: 02 May 2008 Posts: 32341
Back to top |
Posted: 11/20/19 3:04 pm ::: |
Reply |
|
thanks for all this work. some interesting stuff here.
a lot of teams in that top 20 who have gone into irrelevancy.
times change.
_________________ For there is always light,
if only we’re brave enough to see it.
If only we’re brave enough to be it.
- Amanda Gorman
|
|
pilight
Joined: 23 Sep 2004 Posts: 67487 Location: Where the action is
Back to top |
|
pilight
Joined: 23 Sep 2004 Posts: 67487 Location: Where the action is
Back to top |
|
pilight
Joined: 23 Sep 2004 Posts: 67487 Location: Where the action is
Back to top |
|
pilight
Joined: 23 Sep 2004 Posts: 67487 Location: Where the action is
Back to top |
|
|
|