For a few years now, I have observed differences that exist in D3 MVB between mathematical models like Massey, Inside Hitter, KPI, and one I author, Frog-Jump’s T100. They are compared 129 deep this year, but the AVCA Coaches poll only goes about a couple dozen. (Actually, top 20, but the “Others receiving votes” section regularly lends another 4 or 5 more.) As one who often builds models with a “Wisdom of the Crowd” approach, I created a weekly post at Frog Jump called “Tuesdays with MORE.” (Median Order Ranks by the Experts) – Probably the ultimate expression for a rank order of Men’s teams playing D3 Volleyball closest to the truth. It’s determined every week on Tuesday evening after the onslaught of data from hundreds of matches are incorporated into every model’s latest iteration. All are produced for entertainment purposes. They have no skin in the game – No ability to influence a team’s journey across the season, other than informing the threat level posed by a future opponent, I suppose.
Coming in about a month will be output from a new mathematical model adopted by the NCAA which does have the ability to influence some team’s journeys. Whether producing a verdict of “In” or “Out” for the best not to have won conference titles, or to seed those “In” together with teams who have won them. The NPI has skin in the game. The most leverage it exerts will be on programs up to and including the 10 best out of 129, i.e. Above the 92nd percentile. I am both thankful and troubled by this realization. Thankful because even a poor model should still be good enough to capture the best 8%. In all Fall-Season D3-Sports from 2024 the last at-large team’s metric was always at or below the 92nd percentile, which is to say Men’s Volleyball, together with its stringent dial settings, is the most likely D3 Sport to not reveal any cracks. This is because cracks, should they even exist, often wouldn’t turn up before encroaching the 90th percentile. Not dissimilar to the cracks in “O-Rings” in ’86 Challenger’s booster rockets which only appeared after temperatures dropped under 40oF. Admittedly a metaphor whose stakes are worlds apart. The 92nd percentile’s “last at-large” is equally troubling because this means the new selection criteria on the whole is no more forgiving than what it’s been in the past. Neither able to guarantee all among the best 10% in the landscape will have the privilege to compete for a Natty. All this while the total team participation increased by 3 this year, no less. That is no easy task! It would be on par with making $20 more per hour and still saving less than you did before. Wait! Maybe it’s not that hard, after all. LOL All of this not losing sight for the insanely illogical Pool B criteria being banished this year -That’s a win for any who seeks rational consistency! Seeding teams for byes and related to conditional constraints regarding logistics and travel is the least leverage the NPI will apply. Though more teams will be influenced by this aspect, it will be in less significant ways. i.e. Not being invited is a big deal, but who I play after I get invited? Not as much. Earlier, I wrote that even a poor model should be able to “capture the best 8% of teams,” but now with the added burden to get them in the right order, too? C’mon Man! That’s a lot to ask of a poor model!
The phrase “poor model” and “NPI” being mentioned multiple times in the same paragraph was purposeful with the hope to make a point later. Many readers may believe I’ve already judged the NPI as poor because of it. Not at all! It is an elegant piece of mathematics I have yet to see in action, and if you were thinking that I believe it is poor, then read again to see how relieved I am to know the best teams probably will be awarded at-large bids. Simply because chances are so good this model can’t be that bad. It isn’t an endorsement nor is it an indictment, either. It is saying the bar is set so low, this model for MVB shouldn’t fail! Now, I wouldn’t be saying that if this were most any other D3 sports. There is some real consternation regarding the NPI out there. That might be an indictment of the NPI on the whole, the dials its committees set, or even how deep it’s required to go to capture the last at-large bid. These all determine how good a math model is required in order to get the job done well. I also implied I don’t trust a poor model to both, “Get the right teams and put them in the correct order,” but thinking those dots are connected to the NPI is on the reader. Granted, with just a little subliminal encouragement. LOL
Right now, the T100 & IH agree on the top 10 (Not on the order, though! LOL) and the Massey and AVCA Coach’s Poll have 9 of them in their top 10, too. This is just 4 weeks in and with less than a third of all games played. Imagine how in lockstep they might be regarding the truth of such things with twice as much data still to come? All are very good! Will their truth confirm the NPI when it gets published? Probably. Might it show a chink in the NPI armor as seen through the eyes of the other models? Probably as well, but likely deeper than what most will be interested to find out when they read it -The next best five 2nd & 3rd ranked teams from any conference! Remember as you read it, though: 1] As long as champions of the UVC, CVC, MAC, CCIW, and GNAC come from the top 10, the next best 5 will too, regardless of whom among them have conference titles. 2] As soon as one of those 5 doesn’t, there will be a top 10 team on the outside looking in, and this might put a burden on any system, human or otherwise. In that moment, certainly destined to be considered a failure from the perspective of the #10 and its fans. 3] All I know for sure is that I hope to see the NACC & MCVL Champ in the top 10, too. This makes the next 5 become no less than the 12 best in the land to arrive with the chance to win a Natty, if the NPI can gives us the best 12, that is. Then we’d be pretty darn close to that best 10% with a chance I referred to earlier. That would be just!
I attempted to make the point earlier that we sometimes gloss over the surface of things and make assumptions about what we think we saw, heard, or even read. The NPI for MVB might seem to be on par with committee decisions in their 3-year look back, and it might even look pretty good this year and next. That doesn’t mean we should just wait until there’s a head-scratching moment to then look for reasons why it went wrong. MVB is in a unique position right now with so many exemplars of credible ranking systems already in place with a valid track record. They can and should be utilized as a measuring stick to determine what chinks in the NPI armor might lie beneath the surface. Maybe there will be some cracks that go unnoticed or maybe there won’t. Either way, should something be lurking beneath what seems to be another good valuation sports model and a nifty expression of mathematics, I would want to know what it is now more than ever before. Now it really matters.
Over the next month before the NPI is published I will put on my thinking cap to suggest where I would look for such chinks, should they exist.