RPI is officially here for high school football — with a new wrinkle
RPI is here for Kentucky high school football, and, this time, it’s the real deal — but maybe not the system observers had thought it was.
Over the weekend, the Kentucky High School Athletic Association’s online explainer for RPI, which it has had posted for several months, added a new wrinkle for football — wins against teams in different classifications will carry different values than wins against teams in one’s own classification.
Defeating bigger schools means more in the ratings. Defeating smaller schools carries less weight.
“The initial blog post in January may have omitted the class modifier in our haste to get information out, but it was always part of the plan and has been included in all rankings on the website to this point,” KHSAA communications director Joe Angolia told the Ashland Daily Independent’s Zack Klemme on Saturday when news of the difference began to get out.
In Monday’s announcement, the KHSAA released its football RPI page, as well as those for other sports. In football, users can filter the results by class.
Pikeville sits atop Class A while Lexington Christian (2A), Paducah Tilghman (3A), Franklin County (4A), Covington Catholic (5A) and Male (6A) lead their classifications.
Compared with the Herald-Leader’s ratings by Dave Cantrall, discrepancies include Lexington Christian (ranked third by Cantrall), Tilghman (fifth), Franklin County (fourth), and Covington Catholic (second).
Pikeville is the No. 1 team overall in the KHSAA RPI, attributable to the fact that any games it plays against bigger schools are worth more. The Panthers have defeated three Class 3A opponents. By comparison, Class 6A Male (also No. 1 in Cantrall’s ratings) has only played games against teams in its own class and an out-of-state team, which doesn’t get modified.
“While some coaches may argue about this piece or that piece, this poll or that poll, or whichever system would benefit them more than others, in the end, this project will result in people talking about high school sports which is always good. That discussion will lead to continual fine tuning to make the participation opportunities optimal for all involved,” KHSAA Commissioner Julian Tackett said in the announcement. “It is our hope that by laying out all the calculations and being completely transparent with the results, we provide a good publicity tool for our sports, a means for districts in all team sports to potentially break ties in standings, and of course, address the idea to revise the bracketing in football to peak interest in championship play during the playoffs, which was the original request.”
The RPI will update automatically each hour as results are posted for each week’s games.
“Our office is uniquely positioned to implement the RPI rankings in Kentucky high school sports because of our membership requirement to enter schedules and scores on the KHSAA/Riherds.com Scoreboard. The RPI is fully integrated with the Scoreboard, allowing for accurate and updated rankings on an hourly basis,” Angolia said. “As our programmer and database consultant, Frank Riherd has made literally hundreds of trials and calculations to ensure accuracy in this system and we appreciate his commitment to this endeavor.”
An RPI or “Ratings Percentage Index” began appearing on the KHSAA’s class standings for football and other sports two weeks ago, but all carried the disclaimer that the listed RPI was “experimental.”
“It has been an important period the last few weeks experimenting with output formats and screen layouts, while determining what documents were needed to ensure both transparency and replicability,” Tackett said. “We have continually stressed that as the only official results source, any other data outside of our website should be regarded with skepticism. We are very thankful to our friends with the Colorado High School Athletic Association who have been invaluable in troubleshooting our implementation questions. From the beginning, our goal was to internalize this project and not outsource to third parties in order to avoid creating duplicate reporting for our schools and opportunities for confusion.”
For football, KHSAA has said from the outset back in January that its RPI would not become official until about Week 6 of the football season. Week 6 has just passed. The KHSAA also said it would be based on the system used in Colorado and other states. Colorado has a similar “class modifier.”
How does the “class modifier” work? That takes a bit of an explanation, but here goes:
Teams in each class are assigned a different value as follows:
- 1A = 1.323
- 2A = 1.521
- 3A = 1.749
- 4A = 2.011
- 5A = 2.313
- 6A = 2.660
Each classification step roughly equals a 15 percent increase in value for that team.
If a 4A team beats a 2A team, the value of that win is modified by dividing its opponents class value by its own. So, its 2A opponent’s value (1.521) divided by its value (2.011) means its “winning percentage value” for that game equals 0.7563.
If the 4A team plays another 4A team, it’s 2.011/2.011 = 1.
If it plays up to beat a 6A opponent, the win becomes worth more — 2.660/2.011 = 1.3227.
That’s a convoluted way of saying beating a smaller school opponent is worth less than beating a larger school opponent. Losing to any opponent is always factored against zero, so losses to smaller opponents aren’t worth less, but ties against them would be (ties get multiplied by 0.50 instead of one for a win or zero for a loss).
The new “winning percentage value” then goes into another formula to result in the RPI. It goes like this: (0.35 × winning percentage) + (0.35 × opponents’ winning percentage) + (0.30 × opponents’ opponents’ winning percentage). That formula is applied to each game and then combined with every other game on the schedule and divided by the number of games to get the final ranking.
There is no “class modifier” in other sports, because, of course, other high school sports don’t have classes. And the RPI listed for those other sports remains experimental. There has been no proposal before the KHSAA Board of Control to use them to seed tournaments.
In football, RPI does not come into play until after district champions are determined after the second round of the playoffs, and it will be based on an RPI frozen at the end of the regular season (Week 11).
The football regional championships, two games in the east (Districts 1-4), two games in the west (Districts 5-8) in each class, will be seeded with highest seed in the east hosting the lowest seed in the east for a 1 seed vs. 4 seed matchup and a 2 seed vs. 3 seed matchup. The same goes for the west. Then the higher seeds would host each semifinal in each class. Breaking it down by eastern and western areas at the regional stages was done to alleviate travel issues.
RPI top 10s by class
Columns key, from left: Rank, team, class, district, record, winning percentage, opponent’s WP, Opponent’s opponents WP, RPI
Class A:
| 1 | Pikeville | 1A | 7 | 6 | - | 0 | - | 0 | 1.161 | 0.804 | 0.674 | 0.890 |
| 2 | Kentucky Country Day | 1A | 3 | 6 | - | 0 | - | 0 | 1.621 | 0.401 | 0.593 | 0.886 |
| 3 | Paintsville | 1A | 6 | 3 | - | 2 | - | 0 | 0.844 | 0.751 | 0.647 | 0.753 |
| 4 | Holy Cross (Louisville) | 1A | 2 | 5 | - | 1 | - | 0 | 1.118 | 0.423 | 0.632 | 0.729 |
| 5 | Raceland | 1A | 6 | 3 | - | 2 | - | 0 | 0.768 | 0.740 | 0.563 | 0.697 |
| 6 | Hazard | 1A | 7 | 5 | - | 1 | - | 0 | 0.937 | 0.474 | 0.619 | 0.680 |
| 7 | Newport Central Catholic | 1A | 4 | 4 | - | 2 | - | 0 | 0.910 | 0.515 | 0.565 | 0.668 |
| 8 | Williamsburg | 1A | 8 | 4 | - | 2 | - | 0 | 0.717 | 0.748 | 0.505 | 0.664 |
| 9 | Bethlehem | 1A | 2 | 4 | - | 1 | - | 0 | 0.903 | 0.415 | 0.624 | 0.648 |
| 10 | Berea | 1A | 3 | 5 | - | 0 | - | 0 | 1.038 | 0.272 | 0.573 | 0.630 |
Class 2A:
| 1 | Lexington Christian | 2A | 4 | 5 | - | 1 | - | 0 | 1.044 | 0.557 | 0.687 | 0.766 |
| 2 | Mayfield | 2A | 1 | 5 | - | 1 | - | 0 | 0.999 | 0.570 | 0.641 | 0.741 |
| 3 | Somerset | 2A | 4 | 5 | - | 0 | - | 0 | 1.082 | 0.458 | 0.656 | 0.736 |
| 4 | Owensboro Catholic | 2A | 2 | 5 | - | 1 | - | 0 | 1.099 | 0.512 | 0.508 | 0.716 |
| 5 | Breathitt County | 2A | 7 | 6 | - | 0 | - | 0 | 1.140 | 0.362 | 0.580 | 0.700 |
| 6 | Caldwell County | 2A | 1 | 5 | - | 1 | - | 0 | 1.048 | 0.407 | 0.616 | 0.694 |
| 7 | Murray | 2A | 1 | 4 | - | 1 | - | 0 | 1.074 | 0.383 | 0.611 | 0.693 |
| 8 | Washington County | 2A | 4 | 4 | - | 2 | - | 0 | 0.857 | 0.611 | 0.539 | 0.675 |
| 9 | Newport | 2A | 6 | 5 | - | 1 | - | 0 | 0.897 | 0.558 | 0.542 | 0.672 |
| 10 | Lloyd Memorial | 2A | 6 | 3 | - | 2 | - | 0 | 0.734 | 0.655 | 0.518 | 0.642 |
Class 3A:
| 1 | Paducah Tilghman | 3A | 1 | 5 | - | 1 | - | 0 | 0.958 | 0.709 | 0.578 | 0.757 |
| 2 | Mercer County | 3A | 4 | 5 | - | 1 | - | 0 | 0.862 | 0.772 | 0.549 | 0.737 |
| 3 | Bell County | 3A | 5 | 5 | - | 0 | - | 0 | 1.223 | 0.249 | 0.614 | 0.699 |
| 4 | Russell | 3A | 7 | 5 | - | 1 | - | 0 | 0.846 | 0.623 | 0.608 | 0.697 |
| 5 | Bardstown | 3A | 3 | 5 | - | 1 | - | 0 | 0.840 | 0.636 | 0.589 | 0.693 |
| 6 | Belfry | 3A | 8 | 3 | - | 2 | - | 0 | 0.638 | 0.765 | 0.589 | 0.668 |
| 7 | Christian Academy-Louisville | 3A | 4 | 3 | - | 3 | - | 0 | 0.694 | 0.601 | 0.598 | 0.633 |
| 8 | DeSales | 3A | 4 | 3 | - | 2 | - | 0 | 0.769 | 0.508 | 0.607 | 0.629 |
| 9 | Taylor County | 3A | 2 | 5 | - | 1 | - | 0 | 0.771 | 0.453 | 0.551 | 0.594 |
| 10 | Ashland Blazer | 3A | 7 | 4 | - | 1 | - | 0 | 0.860 | 0.376 | 0.535 | 0.593 |
Class 4A:
| 1 | Franklin County | 4A | 4 | 6 | - | 0 | - | 0 | 1.053 | 0.622 | 0.568 | 0.757 |
| 2 | Boyle County | 4A | 5 | 6 | - | 0 | - | 0 | 0.991 | 0.544 | 0.646 | 0.731 |
| 3 | Johnson Central | 4A | 8 | 5 | - | 0 | - | 0 | 1.129 | 0.446 | 0.547 | 0.716 |
| 4 | Wayne County | 4A | 7 | 4 | - | 1 | - | 0 | 0.804 | 0.719 | 0.556 | 0.700 |
| 5 | Lexington Catholic | 4A | 5 | 4 | - | 2 | - | 0 | 0.724 | 0.804 | 0.531 | 0.694 |
| 6 | Anderson County | 4A | 5 | 4 | - | 2 | - | 0 | 0.774 | 0.673 | 0.552 | 0.672 |
| 7 | Harlan County | 4A | 8 | 5 | - | 1 | - | 0 | 0.679 | 0.650 | 0.600 | 0.645 |
| 8 | Logan County | 4A | 1 | 6 | - | 0 | - | 0 | 0.952 | 0.468 | 0.454 | 0.633 |
| 9 | Madisonville-North Hopkins | 4A | 1 | 5 | - | 1 | - | 0 | 0.904 | 0.356 | 0.612 | 0.625 |
| 10 | Waggener | 4A | 4 | 5 | - | 1 | - | 0 | 0.883 | 0.446 | 0.528 | 0.624 |
Class 5A:
| 1 | Covington Catholic | 5A | 5 | 6 | - | 0 | - | 0 | 0.996 | 0.466 | 0.692 | 0.719 |
| 2 | Frederick Douglass | 5A | 6 | 6 | - | 0 | - | 0 | 1.125 | 0.401 | 0.546 | 0.698 |
| 3 | Pulaski County | 5A | 8 | 5 | - | 1 | - | 0 | 0.815 | 0.565 | 0.593 | 0.661 |
| 4 | Scott County | 5A | 6 | 5 | - | 1 | - | 0 | 0.908 | 0.496 | 0.526 | 0.649 |
| 5 | Owensboro | 5A | 1 | 5 | - | 1 | - | 0 | 0.826 | 0.534 | 0.540 | 0.638 |
| 6 | South Warren | 5A | 2 | 6 | - | 0 | - | 0 | 0.897 | 0.469 | 0.501 | 0.628 |
| 7 | Highlands | 5A | 5 | 4 | - | 2 | - | 0 | 0.720 | 0.541 | 0.596 | 0.620 |
| 8 | Conner | 5A | 5 | 5 | - | 1 | - | 0 | 0.851 | 0.395 | 0.606 | 0.618 |
| 9 | South Oldham | 5A | 4 | 5 | - | 1 | - | 0 | 0.862 | 0.396 | 0.527 | 0.598 |
| 10 | Woodford County | 5A | 7 | 3 | - | 2 | - | 0 | 0.604 | 0.586 | 0.499 | 0.566 |
Class 6A:
| 1 | Male | 6A | 4 | 5 | - | 0 | - | 0 | 1.000 | 0.642 | 0.626 | 0.762 |
| 2 | North Hardin | 6A | 2 | 6 | - | 0 | - | 0 | 0.862 | 0.656 | 0.527 | 0.689 |
| 3 | Trinity (Louisville) | 6A | 5 | 5 | - | 1 | - | 0 | 0.833 | 0.609 | 0.553 | 0.671 |
| 4 | DuPont Manual | 6A | 3 | 4 | - | 1 | - | 0 | 0.725 | 0.662 | 0.478 | 0.629 |
| 5 | Central Hardin | 6A | 2 | 5 | - | 0 | - | 0 | 0.883 | 0.410 | 0.573 | 0.624 |
| 6 | Henderson County | 6A | 1 | 4 | - | 1 | - | 0 | 0.725 | 0.489 | 0.585 | 0.601 |
| 7 | St. Xavier | 6A | 3 | 4 | - | 2 | - | 0 | 0.547 | 0.675 | 0.567 | 0.598 |
| 8 | McCracken County | 6A | 1 | 4 | - | 2 | - | 0 | 0.574 | 0.597 | 0.579 | 0.583 |
| 9 | Fern Creek | 6A | 4 | 4 | - | 2 | - | 0 | 0.626 | 0.475 | 0.584 | 0.561 |
| 10 | Ballard | 6A | 5 | 3 | - | 3 | - | 0 | 0.421 | 0.715 | 0.507 | 0.550 |
This story was originally published September 30, 2019 at 9:45 AM.