UDPride Discussion Forums    
     

Go Back   UDPride Discussion Forums

» Log in
User Name:

Password:

Not a member yet?
Register Now!
» Advertisement
UDPride Discussion Forums

UDPride Discussion Forums (http://www.udpride.com/forums/index.php)
-   UDPride Articles (http://www.udpride.com/forums/forumdisplay.php?f=2)
-   -   FLIGHT SCHEDULE w/AD Neil Sullivan: Part 1: The Rise of Arithmetic (http://www.udpride.com/forums/showthread.php?t=33184)

Chris R 08-12-2019 04:46 AM

FLIGHT SCHEDULE w/AD Neil Sullivan: Part 1: The Rise of Arithmetic
 
https://live.staticflickr.com/65535/...16cc143d_c.jpg


Athletic directors share a lot in common with coaches and CEOs: long hours on the job, many nights away from home, and one phone call away from the next big crisis. For most of us, job performance is a no-fly zone of the public’s right-to-know, but that luxury doesn’t exist in college athletics as success and failure are an open book for intense scrutiny. Steve Jobs built the iPhone with unparalleled entrepreneurial talent, design skills, and leadership qualities, but that never stopped millions of others from telling Jobs how to build a better product and run a better company. The CEO of UD Athletics is like being the CEO of Apple: in charge of a brand with fierce loyalty but also impassioned customers unaware of the organized chaos in offices, meeting rooms, and airport lounges required to bring the product to market. If college basketball fans do one thing extremely well it’s holding coaches and ADs accountable, but the messy details in the pursuit of excellence aren’t always at the top of their minds.

UD Vice President/Director of Athletics Neil Sullivan is uniquely self-aware of those pressure points and believes, like the fans do, that keeping score is the ultimate report card. He arrived at Dayton in 2006 and over the last 14 years moved up the chain of command until reaching the top office in 2015. A lot has changed in that timeframe including the landscape of college basketball, but not Sullivan’s expectations or his insatiable appetite for learning something new. He’s driven by it.

The fascination with intel began shortly after his arrival at UD, tutoring under prior ADs Ted Kissell and Tim Wabler as the Flyers developed into a full-fledged A10 Championship-level athletic department. While holding many positions, Sullivan’s ascension to Associate Director of Athletics afforded greater responsibility and input on the direction of the mothership. Kissell and Wabler were exceptional at ‘getting out of the way’ of talented people and letting them thrive. Sullivan’s skills were leveraged everywhere but perhaps most notable were his insights and observations on men’s basketball scheduling. The process used to be straightforward: work the speed dial buttons for traditional matchups with rivals like Cincinnati, Miami, DePaul, Marquette, Western Kentucky, and Louisville, pitch a tent in a holiday exempt tournament, and spend money on a few buy games to round out the schedule.

The exercise is more challenging since Sullivan arrived. Conference re-alignment changed the landscape. Historic opponents stopped returning phone calls. The NCAA Tournament continued to grow as a measuring stick, publicity machine, and cash cow for TV networks and universities. College basketball also developed more parity. George Mason, VCU, Butler, Wichita State, Gonzaga, and most recently Loyola-Chicago punched their tickets to the Final Four. The NCAA Selection Committee leaned harder on performance data metrics to sort the contenders from pretenders and while the Ratings Percentage Index (RPI) had been around since the early 1980’s, Sullivan kept watchful eye on the arithmetic and noticed some interesting trends.

RULES OF THE GAME

The RPI formula was open-source material, allowing those willing to study it some tactical advantages such as leveraging those realities in scheduling and performance evaluation. Scouring over data input and output, Sullivan identified certain biases within those metrics for and against certain outcomes. For North Carolina or Duke, the weeds and grass clippings of NCAA metrics were a trivial matter when wrapped around 30-win seasons and the enormous zeppelin of the ACC. The margins at the NCAA bubble were a different reality. He digested the RPI but also ratings systems from Jeff Sagarin, Ken Pomeroy, Kevin Pauga, and eventually the NET by cross-referencing those data outcomes with NCAA tournament brackets to see if they were inter-related in any way.

They were. The challenge was leveraging the metrics to produce desired outcomes the bean-counters on Selection Sunday preferred. Dayton was already feeling the pinch so even the slightest edge mattered.

UD spent the last decade pouring over that data and Sullivan was (and remains) the tip of the spear. The first major revelation? Recognizing that for all the merits of the RPI, it had two fatal flaws: the computation rewarded playing strong opponents twice as much as winning, and defined strong opponents by their W/L record only.

https://live.staticflickr.com/65535/...44fb39c2_c.jpg

Beating 14-1 Kentucky carried the same basic mathematical weight as beating 14-1 Murray State. The Flyers were rarely on the scheduling radar of Kentucky, but the Murray States of the world were still willing to play. It wouldn’t replace the obvious prestige and traction of beating the Wildcats at Rupp Arena, but whatever credibility gap was lost by the national pollsters and sportswriters could be partially mitigated in the computer rankings. Thus began an additional deep study into the RPI’s Strength of Schedule (SOS) rankings.

CAUSE AND EFFECT

With metrics in hand, Sullivan compiled NCAA at-large résumés and tournament seeding over the last 20 seasons to spot areas where Dayton men’s basketball could exert some push-back. The metrics and the bracket were linked — far tighter than first imagined. The next decision was how to respond. Dayton had no direct control over conference opponents, but non-conference scheduling still had some (albeit dwindling) flexibility. He made it his annual pet project and Wabler often trusted Sullivan’s conclusions and recommendations.

“I’m smart enough to realize that I don’t know everything and that’s what keeps me on the prowl,” Sullivan explained about the data. “The goal is to provide the best possible competitive advantage to our student athletes. Understanding and leveraging performance evaluation metrics is just one cog in the wheel alongside coaching, training, academic support, nutrition, or facilities. No stone will go unturned and that includes the impact of non-conference scheduling on NCAA at-large opportunities.”

Fans have taken notice too. The hot-button topic all summer on UDPride was men’s basketball non-con scheduling for the 2019-20 season. The conversation has gained traction in recent years as the Flyer Faithful enroll in their own self-education bootcamp to better understand the method to the mathematical madness. They follow third-party web sites used as sounding boards for college basketball programs looking for opponents, run their own modest computations, and even lobby for their own Flyer foes. It’s well understood that schools outside the power conferences feel the squeeze with fewer NCAA at-large bids. The Selection Committee often dismisses the criticism by citing a lack of quality wins.

THE GLASS MENAGERIE

There’s also a shortcoming of opportunities.

“The challenges we face in building a quality non-conference schedule have as much or more to do with mathematic realities and narrowing windows of opportunity as it does past performance,” Sullivan said.

The NCAA Evaluation Tool (NET) was introduced last season to address some of the shortcomings of the RPI. The NET’s full-fledged arithmetic remains a tightly-guarded secret, but the NCAA has released bullet points on some of the basic ingredients in the witch’s brew.

https://live.staticflickr.com/65535/...6b049cb9_c.jpg


Unlike the RPI, the NET produces no public rating; it merely publishes the rankings based on a rating that remains confidential. That might sound like a distinction without a difference, but there’s no way to know if the NET’s 20th ranked team is statistically closer to the 19th ranked team or the 50th ranked team. The NET remains an imperfect replacement without open auditing and accountability, but last year’s rankings indicate some promise despite the darkness in the details. The NET also carried over the Quad system from the outgoing RPI to better aggregate wins and losses against stronger and weaker opponents, but that has its own shortcoming such as Sullivan’s narrowing window observation. Power-conference teams benefit from a system where two-thirds of their schedule is pre-cooked with filet mignon.

“Most power conferences have expanded to 20-game league schedules in the last five seasons, eliminating two available non-con scheduling dates among each of those schools,” Sullivan explained. “Dozens of potential scheduling opportunities are no longer available.”

With so many Quad 1 and 2 opponents baked in, there’s little incentive for power-conference schools to challenge themselves with the few remaining open dates. A power-conference team can lose 12 of 13 games in the middle of the season and still be in the at-large discussion because of the strength of their league. While the non-conference schedule is critically important at Dayton, it’s becoming an afterthought for power-conference schools and more like preseason training. They know it and they know Dayton knows they know it. Everyone is focused on self-preservation and if filling out the remainder of a non-conference schedule with Quad 4 buy games makes no abstract or objective difference to their NCAA at-large chances, power conferences are more inclined to do so as long as the NCAA Selection Committee rewards volume of opportunity within their league schedules.

“We are in constant communication with the NCAA and members of the NCAA Selection Committee about all of our concerns, along with my athletic director colleagues,” Sullivan said. “I share data, we hire consultants, we talk with every scheduling guru in the country, including yourself.”


ON TO PART 2: Feeling the Squeeze

TXFlyerFan 08-12-2019 08:55 AM

We've all said and it will say it again. The NCAA could fix this if it wanted to. For example, we've all stated no team with a sub.500 record in conference should be eligible for the NCAA tournament, regardless of their overall W/L record. But here's what I would say.

No sub .500 conference record team should be eligible if their non-conference record is built on home wins over cupcake teams and neutral game sites. IOW, P5 teams should be required to play x number of games, home and away, against non P5 schools that finished in the top 4 of their conference (or say average of top 4 over the past 3 years).

The Fly 08-12-2019 08:57 AM

Great setup piece, Chris. Provides the background necessary to understand the changing landscape of college basketball. Looking forward to the rest of the series!

N2663R 08-12-2019 10:51 AM

Great insight Chris. It's a shame though, since all of this is a data/numbers/AI play and it takes the fun out of college basketball. I want to go back to the Fieldhouse days. :(

rollo 08-12-2019 10:54 AM

I no longer want Neil’s job.

Or Chris R’s.

;)
Posted via Mobile Device

UDEE79 08-12-2019 01:07 PM

Chris isn’t offensive or defensive efficiency measured in points per possession? Your graphic give a formula that shows possessions squared divided by points
Posted via Mobile Device

Wallage 08-12-2019 01:26 PM

Great start. Thanks Chris!

Th NCAA could alter the metrics in a way that make the non-conference schedule non-trivial. Weight it in a way where it can't be overlooked. Something like making non-con count as a stand alone 1/3 of a team's NET ranking. Make it so that no matter how good the conference slate/score is, there's still a full 1/3 of a team's ranking up for grabs.

This probably doesn't affect a P5 team that goes 16-4 in conference. They've done so well its likely still in the noise for them. But for a P5 thats hovering around .500 in conference it could certainly push them into making better use of non-con opportunities.

Special-K 08-12-2019 02:44 PM

My mouth is watering for the next installments. This is going to be good, although is does make my head hurt a little! :)

Chris R 08-12-2019 04:08 PM

Quote:

Originally Posted by UDEE79 (Post 590547)
Chris isn’t offensive or defensive efficiency measured in points per possession? Your graphic give a formula that shows possessions squared divided by points
Posted via Mobile Device


Yeah I inadvertently transposed some division. I fixed. Been staring at this stuff WAY too long the last few days.

springborofan 08-12-2019 08:34 PM

Great stuff but I feel like I just watched an episode of Game Of Thrones and I have to wait a week for the next episode!


All times are GMT -4. The time now is 02:43 PM.

Powered by vBulletin® Version 3.8.6
Copyright ©2000 - 2019, Jelsoft Enterprises Ltd.

Article powered by GARS 2.1.8m ©2005-2006

Powered by vBadvanced CMPS v3.2.1

All times are GMT -4. The time now is 02:43 PM.


Powered by vBulletin® Version 3.8.6
Copyright ©2000 - 2019, Jelsoft Enterprises Ltd.
Advertisement System V2.6 By   Branden

Article powered by GARS 2.1.8m ©2005-2006

     
 
Copyright 1996-2012 UDPride.com. All Rights Reserved.