The qualification process for the NCAA Wrestling Division I championships is, at once, both extremely complex and as simple as can be. On the one hand, if you win your conference championship, you’re in, no questions asked. In addition, once a wrestler gets to their conference tournament, they will know exactly what place they must finish to earn a trip to nationals. However, figuring out those numbers, the allocations for each weight class and each conference, is a process few understand. This year, that process encountered mistakes that caused the allocations to be revised after what should have been the final release not once, but twice.
The process is out there for anyone to see in the qualifier allocation criteria document and the pre-championships manuals produced by the NCAA. The basic premise is that any wrestler who represents their school in the post-season who finishes high enough in two of the three criteria, coaches ranking, winning percentage, and RPI, earns a bid for their conference to be disseminated to the highest finishers at the conference tournament. The devil is in the details, however, as the number of matches someone has to wrestle in order to qualify for each of the three criteria is different and some matches don’t count, such as matches wrestled against non-Division I opponents and teammates for the RPI. Add in that the RPI is a math problem that would take someone hours to complete for one competitor if they were doing it by hand and you’ve got a process that isn’t easy for the average fan to follow.
That doesn’t mean it is a bad process. Each of the three criteria is designed to complement the others in such a way that no deserving wrestler slips through the cracks. If you win a high percentage of your matches and are ranked well by the coaches, you’re in. If your schedule is tough and your winning percentage drops, but you’re still ranked well by the coaches, RPI will save you as it takes strength of schedule into account. If you’re a great wrestler, but the coaches somehow overlook you, you can earn a bid purely on winning percentage and RPI. No system is perfect, but this one is well designed overall.
That said, this season it had a couple of hiccups. As you might imagine, gathering the data for the calculation of RPIs and winning percentages is a labor intensive process. I traded emails with both Anthony Holman at the NCAA and Pat Tocci of the NWCA who are both familiar with what it takes to produce the allocation numbers. From them, I learned that, ultimately, it is the responsibility of each school to ensure their information is accurate. However, there are safeguards in place to minimize the chances of user error. The system of record is what we see looking up a wrestler’s results on trackwrestling so the general public can, in theory, spot an error. However, what transpired this season went beyond what could have been seen by someone on the outside.
When the initial allocation numbers came out, Penn State fans immediately noticed that Vincenzo Joseph was not in the RPI despite finishing the season with the required 17 applicable matches. The database had all of Joseph’s matches, but the query used to do the calculations had been run with an incorrect date range, omitting some matches, including one of Joseph’s. This error was found and fixed rather quickly, leading to a re-release of the allocation numbers the same day. It affected more than just 165 with the SoCon losing a spot at 125, the Big 10 and EIWA gaining a spot at 165, while the SoCon lost a spot to the ACC at 285.
The spot at 125 in the SoCon was, ultimately, restored before the conference tournaments began much to the consternation of the few who noticed. Holman informed me this was due to another mistake found after the re-release of the allocations. The wrestler who lost the spot when the RPI was figured correctly had lost a match to a teammate which should not have factored in, but it was found to have been mistakenly included. When that match was removed, the wrestler in question moved back into position to earn a bid for the SoCon so the bid was restored. No one gained a spot at 125 in the initial re-release and the weight had not given out the maximum number of automatic allocations allowed so the fix was simple.
That these errors were found and fixed is a good thing. However, the fact that they were made in the first place calls into question whether the system is as accurate as it should be. Having the match data on track allows anyone to go in and check for obvious errors such as having the wrong winner or opponent. However, I can tell you from experience that calculating the RPI takes a tremendous amount of effort which makes it next to impossible for a fan to check. What the public can check is if a wrestler should be listed in the RPI or not, but as we saw with the SoCon, catching a mistake in which matches are included or not can make the difference between getting an allocation and missing out. Only those on the inside can see which matches were counted as it stands now.
No system is perfect and Tocci assured me that the check to filter out matches between wrestlers from the same school has been implemented so that mistake will never happen again. That is a good thing, but the fact that it wasn’t implemented before makes it fair to wonder what other mistakes could be lurking. The date range mistake was a human error that was quickly caught, but there are ways to prevent that too.
The biggest concern in all this is that we have to take the word of the NCAA and NWCA because the data isn’t easily available for review. There is currently no way to know which matches are being included in the RPI calculation so we must assume the rules are being followed correctly. If this year’s date range problem had not happened to exclude a couple of wrestlers that had exactly 17 allowable matches, would it have been caught? If it wasn’t caught, the allocations would have been incorrect, but who would have noticed? There is no doubt in my mind that the people involved are doing their best to get it right every year and they are constantly trying to improve the process. However, as long as some of the data remains nearly impossible to verify, we’re reliant on a small number of dedicated people to make zero mistakes. That isn’t a recipe that inspires confidence.