The NCAA men's basketball committee's conclusion on Friday to incorporate game location into how it determines the quality of a win or loss is a step toward increasing the use of analytics in its decision-making -- though it still has a ways to go.
The committee previously defined a quality win as any victory over a team with an RPI rank from 1 to 50. Now, it has changed the definition of quality win to mean one that comes at home against an opponent ranked 1 to 30, at a neutral site over a foe ranked 1 to 50 or on the road versus a team ranked 1 to 75.
That roughly jibes with BPI-based analysis indicating that in the 2016-17 season, a road victory by the 50th-ranked team -- your average bubble team -- over the 75th-ranked team would more or less be equivalent to that same team beating the 29th-best team at home.
That's obviously a fairly wide disparity, and one that makes judging home and road games a prudent move by the NCAA.
In other words, a win by BPI's 50th-ranked team, Vermont, at 75th-ranked Ole Miss would have been the same difficulty as a Catamounts victory at home against 29th-ranked Marquette last season. That same win in Oxford, Mississippi, would have been evaluated by BPI as slightly more impressive than a home victory against 33rd-ranked Xavier -- which went all the way to the Elite Eight in the 2017 NCAA tournament.
But it is important to recognize that the above analysis uses BPI -- a measure of a team's current strength and a predictive statistic -- to evaluate an opponent's strength. The committee, however, is still stuck using RPI, an arcane statistic that evaluates an opponent's strength by wins and losses and not some of the other factors -- including margin of victory, distance traveled and rest differential -- that BPI and other advanced metrics incorporate.
A step up from the committee's new approach to quality wins would be to use a predictive statistic such as BPI or a team's KenPom rating, which more fully captures an opponent's ability.
Even better, the committee could simply eliminate the need for an arbitrary cutoff point to define a "quality" win. And if the committee would like to judge a particular team's résumé, it instead could look at a statistic such as ESPN's strength of record, which paints a fuller picture of a team's accomplishments in a given season. Strength of record awards the proper amount of credit for a team's win-loss record based on the strength of its opponents, while incorporating factors such as distance traveled and rest differential.
Let's use North Carolina's 89-86 overtime win at Clemson last season as an example. Under the old system, that was deemed a second-tier victory, because Clemson was RPI's 75th-ranked team. But had the new system been in place, the Tar Heels would have been awarded a "quality" win for that same performance against the Tigers. But consider that if Clemson were ranked just one spot lower in RPI, the victory would no longer have been of the same "quality." Does one spot in RPI make that much of a difference in a team's strength? It likely does not. And that is one area where a strength of record statistic shines in comparison to the new RPI-based quality win analysis, which requires an arbitrary demarcation.
More tweaks are expected to be in the pipeline for the committee. The NCAA's press release indicated that there likely will be a new metric used to evaluate teams starting in the 2018-19 season.
That will be a welcome evolution, because while Friday's announcement was a step in the right direction of properly identifying and seeding 68 teams for the NCAA tournament, flaws in the system will remain for at least one more year.
Paul Sabin contributed to this report.
For more from ESPN Analytics, visit the ESPN Analytics Index.