|
Written by: Matt Harris
Data/Numbers by: Dayton Robison
As we conclude the state basketball tournaments this weekend with the Boys Real Dairy Shootout and fans, coaches, and players alike begin to dig into the state tournament brackets, the inevitable discussion regarding bracketing procedures is raised. It's a question that comes with every state tournament season without fail.
That said, it doesn't mean that the state's current method of bracketing is wrong or anything like that - it's just a different system to bracket teams, one that the schools have agreed upon year after year. But that doesn’t mean we can’t throw some ideas around and experiment with a different system.
Earlier in 2017, I wrote about an experiment with a Ratings Percentage Index that was being conducted by my friend, Dayton Robison. You can read more about that by clicking HERE. He has dubbed his experiment the "RRPI", which stands for "Robison Ratings Percentage Index", and modeled it after the RPI formulas that nearby states use to help them seed their state tournaments, with minor modifications for customizable personal preferences. Robison took all of the football scores from Idaho (including out-of-state scores and out-of-state team records) and plugged them into his formula to produce the first complete edition of the RRPI, to which we seeded a hypothetical state football tournament. You can see those results by clicking HERE.
The goal of the RRPI was to be able to rank teams accurately based on their record and overall strength of schedule, then seed those teams into a hypothetical state tournament scenario based on those rankings. Robison took many elements out of the Washington Interscholastic Activities Association RPI and combined them with elements of his own preference. You can read more about his formula by clicking HERE.
For purposes of our experiment regarding the seeding of the state basketball tournaments, we decided to focus in on two formulas: the Washington base and the Colorado base. The Washington formula is the same formula used by the Washington Interscholastic Activities Association’s RPI to rank and seed teams. The same formula is used by the Colorado High School Activities Association to seed their state basketball tournaments, however for football they use a different formula – and for the purposes of this experiment and to show different examples, we decided to use their football formula to see how much would change.
The formulas each create composite scores for every team in the state using each team’s own win percentage, their opponents win percentage, and their opponents’ opponents win percentage. Each formula is as follows:
-Washington: 25% team’s winning percentage (known as WP), 50% team’s opponents win percentage (known as OWP), and 25% team’s opponents’ opponents win percentage (known as OOWP) for a total of 100%.
-Colorado: 37.5% team’s winning percentage, 37.5% team’s opponents win percentage, and 25% team’s opponents’ opponents win percentage for a total of 100%.
The Colorado formula also gives a ‘cross-classification weight’ for teams that beat an opponents in a higher or lower classification, which is set at 15%. For example, that would mean that if a 4A school beat a 5A school, the win would count for 15% more towards to their final RPI score. Conversely, if a 4A school beat a 3A school, their win over the smaller school would be worth 15% less towards their final RPI score. The Washington formula does not use a cross-classification weight.
When calculating out-of-state opponents in the formula, their direct winning percentage will count toward the score of the Idaho team they played, but the out-of-state team’s opponents each have their own win percentage set to 0.500. As the CHSAA states, “schools would be chasing tens of thousands of opponents and out-of-state opponents records over the course of the season, and there is no way to ensure the accuracy of that data.” Long story short, it’s easier and more efficient to do it this way.
Robison compiled the results of every single game in the state of Idaho plus the results of out-of-state opponents to be able to run the numbers and how an Idaho version of the RPI rankings would look. Below, you will see 12 links to see the results of how each state tournament would look if it were seeded using RPI, broken down into the six boys and six girls classifications. In each link, you will see six different brackets, three which use the Washington base formula and three which use the Colorado base formula:
-2018 Teams: This bracket is what the tournament would look like if we re-seeded the teams that have all qualified for this year’s state tournament. These teams would all be seeded using their RPI ranking.
-Champions + At-Large: This bracket is what the tournament would look like if each district champion automatically qualified and then the next three best teams were chosen for an at-large bid. These eight teams would then all be seeded according to their RPI ranking.
-Top 8: This bracket is what the tournament would look like if we took just the top 8 teams in each classification according to their RPI ranking and seeded them into the tournament.
Now, this isn't to say that this is the only way of operating a seeded bracket. There are many different ways in which it could be conducted, including different weight percentages, modifiers and more. We are just showing you a few different examples of what could happen if the tournament was seeded using RPI.
|