Basketball Colorado – The Answer to RPI
10/26/2016
Face it, Colorado high school sports are stuck with computer rankings and the RPI. Valid reasons were given as to the switch to RPI but the formula needs to change or a whole new system needs to be considered.
In late August of 2015, CHSAA commissioner Paul Angelico told CHSAANow.com that “The membership, for years, has been requesting a more transparent system that would provide more consistency among all sports and state qualifiers,” as to why there was a move to the RPI system. Last spring CHSAA implemented RPI to determine who would make the state playoff series in all sports. One of the first sports that got to use the new RPI format was baseball and let’s just say that the RPI system was not warmly received. Fast forward to the fall sports and again there have been some head-scratching moments when the RPI rankings were released. CHSAA has preached patience. That it is a work in progress and it will work out in the end.
Ryan Casey made some good points on how the formula could change in his article on ChsaaNow.com. I can back either formula change but what I can not and will not accept is a system or formula that ranks undefeated teams or possibly one loss teams below the teams that they have defeated soundly on the field or on the court. A system I would consider to go along with a human component is Professor Erik Packard’s Rankings. Packard is an associate math professor at Colorado Mesa University and has been running computer rankings for a number of years and the prediction machine is at times eerily accurate. If last year’s 5A tournament had been seeded by Professor Packard’s Rankings these are the teams that would have qualified.
1 OVERLAND (1) 13 CHERRY CREEK (4) 25 GREELEY WEST (7) 37 GRAND JUNC. (10)
2 RANGEVIEW (1) 14 BOULDER (4) 26 AURORA CENTRAL (7) 38 R. VALLEY (10)
3 EAGLECREST (1) 15 DAKOTA RIDGE (4) 27 LAKEWOOD (7) 39 SMOKY HILL (10)
4 THUNDERRIDGE (1) 16 GRANDVIEW (4) 28 PALMER (7) 40 FRUITA (10)
5 CHEROKEE TRAIL (2) 17 ARAPAHOE (5) 29 LIBERTY (8) 41 GJ CENTRAl (11)
6 REGIS (2) 18 ARVADA WEST (5) 30 HI. RANCH (8) 42 DOUG. COUNTY (11)
7 CHATFIELD (2) 19 Monbello (FNE) (5) 31 MONARCH (8) 43 WESTMINSTER (11)
8 ROCK CANYON(2) 20 RAMPART (5) 32 BROOMFIELD (8) 44 MTN RANGE (11)
9 DOHERTY (3) 21 LINCOLN (6) 33 CHAPARRAL (9) 45 MULLEN (12)
10 GEORGE WASH. (3) 22 MTN. VISTA (6) 34 HERITAGE (9) 46 F.Ft.CARSON (12)
11 DENVER EAST (3) 23 FORT COLLINS (6) 35 ROCKY MOUNTAIN (9) 47 FAIRVIEW (12)
12 LEGEND (3) 24 FOSSIL RIDGE (6) 36 LEGACY (9) 48 LOVELAND (12)
Teams In That Were Left Out
Douglas County, Fountain – Ft. Carson, Mullen, Mountain Range
Teams Left Out That Were In
Castle View, Brighton, Hinkley, and Bear Creek
Here is what the bracket would have looked like if strictly seeded based on the rankings with no consideration given to league opponent match-ups or if teams played each other in nonconference games. As it should be. If you are going to use a system like this anyways then what’s the point if humans are going to come in and shake things up anyways.
How the bracket would have played out using only Packard Predictions where teams didn’t face each other.
Only Cherry Creek and Doherty won games where the Packard Predictions had them losing.
The Packard System is more complex than the current RPI. “My system is more complicated and hard to describe,” said Professor Packard. ” Whenever two teams play their power ratings are compared and if one team does better than it is supposed to it goes up. The season is run over and over up to the current games. So, if you want a team to rise in the power rankings you want the teams that they played to do better than they were supposed to and also teams that played teams you played etc. There are diminishing returns for blowouts. And most importantly it isn’t all that simple because there is a reward for winning.”
However while the Packard Rankings seem to be the best choice when it comes to computer rankings, the computer only knows what you tell it. That is why there must be a human component to the rankings. Perhaps working in a coaches poll or a combination of media and coaches polls could work.
Like Casey mentioned in his article today about Rangeview, the RPI would have absolutely hurt Rangeview’s ranking last year because of the league that they are in. The RPI didn’t see how the Raiders came back from Christmas to defeated what would turn out to be 6 of the sweet 16 teams including two league champs. The computer rankings also do not take into effect styles of play or injuries. That is why there must be a human factor in the rankings to make the state tournament. Now, how the state tournaments are seeded, is another post altogether.
