2016-17 Non-Conference Schedule Ratings: Methodology

With just about everyone's non-conference schedule released, it's time to take a look at which teams built their (primarily) November and December slates with an at-large bid in mind and those who, bless 'em, elected to pre-emptively give the Selection Committee a reason to leave them out. But before I get to the schedules themselves, I must explain how I'm evaluating them.

For background on what I've tried to accomplish with my ranking formulas, check out my post on this subject from 2015. Most of the fundamentals haven't changed, particularly regarding conference strength and which teams you want to schedule and avoid. However, I've made some tweaks, specifically to the formula I use to rank schedules.

Opponent Scores

To calculate the relative strength of each of Division I's 351 men's college basketball teams, I used the same basic four-year formula from 2015-16, dropping 2011-12 and adding last season. As usual, more recent seasons are rated more highly, but last year is not given full weight due to roster changes.

The data consists of KenPom.com's Pre-NCAA Tournament Pythagorean Ratings (pwp) for 2012-13 through 2015-16 that are factored in the formula like so:

2013pwp*0.2 + 2014pwp*0.4 + 2015pwp*0.6 + 2016pwp*0.8 = Opponent Score

As an examination of the top and bottom 10s rankings based on this formula illustrate, teams typically fall where you'd expect them to.

Top 10: Virginia Cavaliers (1.878); Kansas Jayhawks (1.857); Villanova Wildcats (1.852); Arizona Wildcats (1.851); Kentucky Wildcats (1.829); Michigan State Spartans (1.828); Louisville Cardinals (1.821); Duke Blue Devils (1.815); North Carolina Tar Heels (1.811); Wichita State Shockers (1.807)

Bottom 10: UMBC Retrievers (0.262); Maine Black Bears (0.259); Central Arkansas Bears (0.256); Central Connecticut State Blue Devils (0.256); Presbyterian Blue Hose (0.251); UTRGV Vaqueros (0.247); Abilene Christian Wildcats (0.245); Mississippi Valley State Delta Devils (0.186); Florida A&M Rattlers (0.177); Grambling State Tigers (0.113)

You can access my full Excel spreadsheet here: Team Ratings 2013-16 .

Schedule Strength

To judge a team's non-conference schedule strength, I settled upon an improved formula from last season, one that reflects a game's location. For each team covered, you'll see their individual non-conference games appear under one of these columns.

True Home Games	Semi-Home Games	True Neutral-Site Games	Semi-Away Games	True Away Games
Opponent Score*1	Opponent Score *1.25	Opponent Score*1.5	Opponent Score *1.75	Opponent Score*2

Note that I'll be more specific about why I placed games in each category in each team's individual capsule.

To evaluate a single game, I multiplied each opponent's rating against the location factor listed above. For example, a true home game against Virginia is worth 1.878; a neutral-site game, with a factor of 1.5 included, worth 2.817; and a road game against the Cavaliers is worth double, 3.756. Therefore, a team gets credit for playing more difficult opponents away from home and is dinged for loading up on cupcakes, particularly at home.

To determine a team's non-conference schedule rating, I came up with this formula, which also accounts for the varying sizes of non-conference schedules.

Game score 1 + Game score 2 + Game score 3 + ... Game score N / N ...where N equals the total number of non-conference games a team scheduled. Note that individual Ns range from nine to 16 games this season.

This season, I've scored games scheduled against non-D1 opponents with a zero.

Teams participating in bracketed tournaments receive three scores.

• The average score accounts for the game score for the first opponent and the average game scores of all possible opponents for the second and third games.
• The max score accounts for the game score for the first opponent and the game score for the best possible opponent for the second and third games, i.e., this score projects a trip to the event's championship contest.
• The min score accounts for the game score for the first opponent and the game score for the worst possible opponent for the second and third games, i.e., this score projects a trip to the event's consolation game (in a four-team bracket) or seventh-/eighth-place game (in an eight-team bracket).

The average score is the most meaningful one, in my opinion, simply because it's impossible for all evaluated teams to experience either the best or worst case scenario. Indeed, the overall min and max rankings will wildly vary from the average ranking, as you'll see in future posts. All teams, even those whose schedules are set in stone, are placed within all three rankings.

Naturally, the higher the score the better, and a team can have a score higher than two if they scheduled many road games against quality opponents (Long Beach State, not to give anything away). Likewise, if a team plays a lot of cupcakes at home, their score can fall below one (looking at you, USF).

At the start of each conference's entry, I've posted the league's rankings for all three categories.

Exempt Tournament Location Ratings

I also applied location ratings for exempt events, so if a team could potentially play a true or virtual home or road contest in one of these tournaments, my calculations account for this. As an example, let's take a look at the Gonzaga Bulldogs' schedule:

	Gonzaga Bulldogs
Average Opponent Score	1.584 (avg.); 1.643 (max); 1.486 (min)
At-Large Pool Ranking	30th (avg.); 30th (max.); 37th (min.)
Mid-Major Ranking	9th of 24 (avg.); 8th of 24 (max.); 11th of 24 (min.)
Exempt Event*	AdvoCare Invitational (three at Lake Buena Vista, Fla.)

True Home Games (7)	Semi-Home Games (0)	True Neutral-Site Games (2-4)	Semi-Away Games (1-3)	True Away Games (0)
San Diego State Washington Akron South Dakota Bryant Utah Valley Mississippi Valley State		Quinnipiac* Florida/Seton Hall* Iowa State/Miami/ Stanford/ Indiana State* Arizona (Los Angeles)	Florida/ Seton Hall* Iowa State/Miami/ Stanford/ Indiana State* Tennessee (Nashville)

This season, the Zags will look to win an unprecedented third AdvoCare Invitational title in Central Florida. This time around, however, not one, but two Sunshine State teams will share the eight-team bracket with Mark Few's squad. Therefore, Gonzaga's three games at ESPN's Wide World Of Sports look like this in the formula.

• Game one against Quinnipiac counts as a true neutral-site game and the Bobcats' opponent score is multiplied by 1.5 (0.752*1.5 = 1.128).
• Game two could be a neutral-site contest against Seton Hall, multiplied by 1.5 (1.471*1.5 = 2.207), or a semi-road game against Florida. In the latter case, the value jumps up to 1.75 (1.696*1.75 = 2.968). Keep in mind that, average score of the two options (2.589) also accounts for the location, not just Seton Hall and Florida's raw opponent scores.
• Game three could be a neutral-site contest against Indiana State (1.027*1.5 = 1.541), Iowa State (1.775*1.5 = 2.662), or Stanford (1.495*1.5 = 2.243), multiplied by 1.5. But it could also be a semi-road game against Miami, multiplied by 1.75 (1.657*1.75 = 2.900). Again, the average game score for the third game (2.337) accounts for the four location factor options.

You'll note that games for a bracketed tournament appear twice if there's a location-rating difference, with the names of the teams that don't fall under the category crossed out. In our example above, you'll note that Seton Hall is crossed out of the semi-away game column for Gonzaga's second AdvoCare game, while Florida is struck through for the neutral-site game column.

If a game is part of an exempt event, I've denoted it with an asterisk. If a game is part of another off-campus event, I've placed the location in parentheses below.

Teams Covered

This year, I've expanded my analysis from 116 teams to 135. Each team from the American, Atlantic 10, ACC, Big 12, Big East, Big Ten, Mountain West, Pac-12, and SEC is covered in that conference's specific post, while I have also included 24 mid-majors that might have a shot at an at-large bid in a tenth post.

Therefore, for overall ratings, teams are classed as part of the "at-large pool" and ranked from 1st to 135th. The American Athletic Conference is first up.

Be sure to follow @ChrisDobbertean on Twitter and to like Blogging the Bracket on Facebook and Google Plus.