#wichitast

By Dan Loman

As everyone knows, the NCAA Tournament is crazy. Every year the tourney can seem like an outlier, and every year narratives arise about why things happened the way they did. Here’s 3 myths that I’ve seen arise entering the Final Four of the NCAA Tournament in 2016 that can be debunked using analytics:

1. There’s more parity in college basketball this year and going forward:

After one of the craziest first weekends of all time, which included half court buzzer-beaters, a 12 point comeback with under a minute remaining and the first day in NCAA Tournament history in which a 13, 14, and 15 seed all won on the same day, one of the big narratives that has been the amount of parity that supposedly exists in college basketball this year and moving forward. After all, we’re talking about a year where three 11 seeds, two 12 seeds, a 13, 14 and 15 seed advanced to the second round, off the heels of a regular season with arguably the highest turnover rate at the top of the AP poll in recent memory.

I used kenpom.com’s pythagorean rating - a team’s estimated winning percentage against an average team - since 2003 to see if this year looked any different than previous years. Higher parity would imply less spread among the quality of teams, with the best and worst teams closer to the middle of the pack than usual. To test this, I looked at the standard deviation of all pythagorean ratings across years. Theoretically, a smaller spread in team quality would equate to more parity.

As it turns out, there’s no significant difference in this year or any year since 2003. In fact, even with the top-heaviness of last season, there’s been even less “parity” (i.e. a higher standard deviation among all pythagorean ratings) in 2016:

But maybe by “parity”, we really mean that there are fewer dominant teams. After all, only Kansas entered the tournament with 4 or fewer losses and the top of the AP Poll has been a revolving door all year. Compared to last year this is certainly the case, but last year was a bit of an outlier when we were spoiled with a handful of high-quality top teams. If we look at the top 10 teams across the past 13 seasons, 2016 is generally right in line:

While 2016 is certainly on the low-end for top heaviness, it’s well within the range of most seasons and we’re certainly not trending towards any more parity in college hoops. As an aside, note the quality of the 2015 draft vs the 2016 prospects. College basketball lost a ton of talent to the NBA last season and replaced it with a less heralded 2016 class (and a class where the top prospect, Ben Simmons, didn’t even play for a blue-blood program). With the highly regarded 2017 class entering college next year, expect the quality of top teams to return closer to 2015 than 2016 next year.

2. Mid-major conferences are catching up to power conferences

An extension of the first myth, but after the plethora of mid-major schools that pulled upsets in the first round of the 2016 NCAA Tournament, many people started to wonder if mid-major talent simply isn’t as far off from major talent as it used to be. A look at overall team trends since 2003 suggests that this isn’t true:

Overall, the quality of mid-major teams has stayed the same over the past 13 seasons. It’s worth noting that the overall rating of power conference teams dipped slightly over the past 5 seasons, which likely correlates with the creation of the American Athletic Conference (Since the AAC was formed with many teams from the old Big East, and can claim two of the past four NCAA Champions, I counted it as a power conference).

But maybe we’re only concerned with tournament teams. After all, there will always be bottom-feeding teams in the NCAA, but if mid-major talent is catching up to high-majors, it will show in their best teams:

It looks like if anything, mid-major talent was among the weakest in recent history this year. This can probably be explained by a relatively down year from more traditional mid-major powers like Gonzaga and Wichita State, who have garnered a top 2 seeds in recent years but were both 11 seeds in 2016.

3. The seeding was wacky:

If there’s no discernable increase in parity this year, maybe the upsets can be explained by some questionable seeding decisions by the selection committee. After all, entering the tournament who pegged Oregon as a 1 seed, Utah as a 3 seed or Syracuse as a 10 (although Syracuse is making that claim look silly in hindsight)? On the flipside, Kentucky, Wichita St. and Stephen F Austin all seemed drastically underseeded as 4, 11 and 14 seeds, respectively.

I looked at all NCAA teams since 2003 and for each team, I found their “seed bias” by calculating the difference between their actual pythagorean rating and the average pythagorean rating for all teams with their seed. I omitted 15 and 16 seeds, since they can get a little tricky (an exceptionally poor team that snuck into the tournament can skew this analysis).

As you can see, there’s a higher seeding bias than usual this year, but it’s not statistically different from any other year. In fact, the seeding bias this year is right in line with 2015, where Texas, Wichita St. and Utah were all substantially underseeded according to their pythagorean ratings.

Although the correlation is relatively weak, underseeded teams do tend to outperform expectations. This helps explain why Stephen F Austin, Wichita St and Gonzaga, three of the most underseeded teams in the 2016 tournament, all pulled off upsets. It also means that tournament underseeding, not parity, is a better indicator of first round upsets: