|
|
Post by Judge Smails on Oct 24, 2019 16:47:32 GMT -8
Riiiiight. Pick a pacific-time-zone team at random and count how many central/eastern-time-zone teams it plays. The norm is three or four, out of 30+ game schedule. That isn't nearly enough to equalize the SoS component.
This really isn't something that you can argue, baseba1111. It isn't opinion, it's mathematics. Given the data and some programming skill, the precise amount of bias can be calculated. My back-of-the-envelope estimate is that it's between 0.0025 and 0.0075 points.
Oh 4 can argue it all day long. Don't underestimate him 50% of the time, he argues every time. It’s science. |
|
|
|
Post by shelby on Oct 24, 2019 18:31:13 GMT -8
But, he is never wrong because he loves to change the topic and berate the poster - a TGP ( Tough Guy Poser )
|
|
|
|
Post by baseba1111 on Oct 24, 2019 21:35:58 GMT -8
Riiiiight. Pick a pacific-time-zone team at random and count how many central/eastern-time-zone teams it plays. The norm is three or four, out of 30+ game schedule. That isn't nearly enough to equalize the SoS component.
This really isn't something that you can argue, baseba1111. It isn't opinion, it's mathematics. Given the data and some programming skill, the precise amount of bias can be calculated. My back-of-the-envelope estimate is that it's between 0.0025 and 0.0075 points.
LOL... yeah, I know it's a dead horse. I too have some Adv Math knowledge... and guess what I do not need it to explain/know there is no bias in a formula. The formula doesn't pick/sample who a team plays. The formula is completely subjective... plug in numbers from a teams schedule. Since you seem like a Math guy, you must know the statistical definition. Bias in the mathematical sense is the systematic versus the random distortion of a statistic as a result of sampling procedures. Doesn't exist here. No distortion... unless you pick crap teams (terrible SOS and record) to load your schedule with. But... being these are my last words on a very simple topic... back to simplicity. Play lousy teams that get pummeled by other teams it effects your RPI. The RPI formula did not select the teams, nor did it have anything to do with their W's or L's. Teams that play better teams have better RPIs assuming some wins are involved. Playing East coast teams was an example as you/others seem to think the RPI gods favor the east. There are plenty of Midwest/Western teams that OSU (or any Western university) could play versus this set of powerhouses from last season: (only D1 games count): Cal Poly (5-21) RPI #316 Ark PB (2-19) RPI #349 LaSalle (6-21) RPI #284 Santa Clara (13-17) RPI #163 E.Wash (12-20) RPI #182 CSU Bakersfield (12-17) RPI #276 Toss is WSU (RPI #158) and UW (RPI #136) twice and Colorado (RPI #178) whom you can't avoid playing... that's choices made not the formula. Luckily OSU also had ND (RPI#1), TexA&M (RPI#16), SCar (RPI #17), Oregon x2 (RPI #4), Furd (RPI #5)... and still OSU had the 3rd worse SOS ranking in the conference. BY choice...Basketball is not like football, schedules are not made years in advance. Schools/coaches make choices on who to play. It could be like (using last year's RPI): Iowa #7 Iowa St #11 Marquette #12 Zags #13 DePaul #18 Drake #20 BYU #26 SDakota St #27 Texas #28 Missouri #30 Rice #31 Kansas St #33 Boise St #39 SDakota #41 Closer to home but not in the 200-300's and all higher than UW, WSU Colorado:St Mary's #49 (did last year I believe?) New Mexico #55 TCU #58 Wyoming #64 UC Davis #71 Idaho #84 Pepperdine #99 Pacific #102 Loyola Mary. #104 PSU #106 Idaho St #109 Fresno St #120 New Mexico St #125 N.Colorado #129 ETC... There are tons of options to make choices to enhance a team's RPI here on the West Coast and not play so many teams in the 200-300's. That it what effects the RPI of a team. And, depending on where you want to draw your line in the sand... which conferences are "East"... top 64 RPI teams were pretty spread out and teams in the tournament from major conferences pretty much follows the number of teams with RPI #50-60 and above... |
|
|
|
Post by baseba1111 on Oct 24, 2019 21:37:17 GMT -8
But, he is never wrong because he loves to change the topic and berate the poster - a TGP ( Tough Guy Poser ) And... as always your cognitive skills and post leave nothing to the board's imagination! Actually read the thread? Yeah... thought so. But, way to chime in on a couple threads with such invigorating stuff! |
|
|
|
Post by nwhoopfan on Oct 24, 2019 23:31:25 GMT -8
The Beavs did indeed play St. Mary's last season, beat them 89-56, and it really wasn't even that close. They were up by 40 or 50 at one point. Oregon also drilled them 79-55. They did beat Wazzu and barely lost to Cal, were competitive in the WCC, but they looked completely hapless against OSU. Yeah, the Arkansas-Pine Bluff and CSU-Bakersfield type teams really kill the RPI. I seem to recall when Coach Neighbors was at UW he convinced everyone to do some smart scheduling that really helped the conference's RPI. Maybe when he left that all went out the window.
edit--unfortunately it appears there are some real bottom feeders again this year; CSUB again, I saw Northern Arizona against UW early last season and they were awful, I'm guessing Southern Utah isn't very good.
|
|
|
|
Post by beaverstever on Oct 25, 2019 8:00:19 GMT -8
The Beavs did indeed play St. Mary's last season, beat them 89-56, and it really wasn't even that close. They were up by 40 or 50 at one point. Oregon also drilled them 79-55. They did beat Wazzu and barely lost to Cal, were competitive in the WCC, but they looked completely hapless against OSU. Yeah, the Arkansas-Pine Bluff and CSU-Bakersfield type teams really kill the RPI. I seem to recall when Coach Neighbors was at UW he convinced everyone to do some smart scheduling that really helped the conference's RPI. Maybe when he left that all went out the window.
edit--unfortunately it appears there are some real bottom feeders again this year; CSUB again, I saw Northern Arizona against UW early last season and they were awful, I'm guessing Southern Utah isn't very good.
The Pac-12 is to WBB that the SEC is to football - they can schedule the sisters if the poor in preseason and all that matters is what happens in conference. They will be tested plenty in league. This conference could produce the situation like UConn men a few years back where they finished in the middle of their conference and won the title. I too would prefer they be tested a bit more to prep for conference, but what is going to really matter is bringing along they young talent and staying healthy going into the postseason. |
|
|
|
Post by willtalk on Oct 25, 2019 11:21:03 GMT -8
You could still schedule preseason cupcakes. Just better ones. What is the difference if you beat someone by 40 or 20 points, but it makes a huge difference in the RPI. Just schedule teams in the 150 to 200 range instead of 250 to 300. I suppose the lower-ranked teams are easier to get scheduled without having to do a home and home.
Oregon St is at the level where no team lower than 50 on the RPI has any chance to beat them. Anyone lower than 150 would be considered a cupcake team. That is the problem with the RPI. If they really want to have a realistic value on a team's strength of schedule, they should only count games against teams that have a realistic chance to beat them in the formula. The won-loss records still remain, but the strength of the schedule should be figured using only teams within a certain numerical range. That range would have to be figured based on probable upset parameters. If a team loses then that team should be figured into the equation no matter their numerical standing.
Basically the RPI as a basis for rating strength of schedule has too many built-in flaws to be much value in that respect. It was a simple system worked out to give some sort of value to a teams won-loss record to people who knew little about the teams they were evaluating. While it is not entirely useless, it has too many flaws for it to still be used as a standard by the selection committee.
I remember a few years ago when the second-place team in the AAC had an RPI of 15 without beating a single team in the top 50. They owed their inflated RPI rating to the fact that they lost to top-rated Uconn 3 times that year. Playing and losing to a top 3 team, no matter by how much, will do more than beating a top 25 team. Every team in the AAC also had their RPI inflated by losing to Uconn twice. This also inflated the entire leagues' score which was again magnified when they played each other. Conversely, a team that plays no cupcakes can get their RPI damaged by playing even a very strong team who did. Anyone but a numerical methodologist can recognize the inherent problem in this.
|
|
