We’re now less than two months away from Selection Sunday. It’s officially bracketology season, and Go Joe Bruin is getting in on the action.
The NCAA Division I Men’s Basketball Tournament is impossible to predict. There are too many intangibles, too many upsets, too many factors that cannot be accounted for on paper. You can’t predict what the bracket will like after all the madness has taken place.
I was thoroughly convinced of this after 2012, when I picked Missouri to cut down the nets, and then in 2013, when I picked Georgetown to go all the way. Tired of losing bracket pools with family and friends despite my quasi-obsessive knowledge of college basketball, I needed a pick-me-up. I needed to be right about something when it came to college basketball.
That’s how I became a bracketologist. Exasperated with forecasting the storm that ravages brackets across the country every March, I became determined to correctly forecast what the calm before the storm would look like.
It turns out I wasn’t alone. Evidently, bracketology isn’t limited just to ESPN’s Joe Lunardi; in 2015, 136 bracketology entries were submitted to the Bracket Matrix by Selection Sunday. The Matrix ranks every bracketologist based on their accuracy throughout the years.
I couldn’t beat my friends at a game I know much more about. So I have decided instead to refocus my efforts at beating Joe Lunardi at a game which happens to be his profession. What could possibly go wrong?
Next: Cracking The Code
Cracking The Code
Ever since I started my own bracketology in 2014, I’ve done things a little differently; I have a bigger goal in mind than just accurate predictions. The ultimate goal is to find a formula that can project what the bracket will look like once Selection Sunday is over, without any subjective human judgement.
UCLA Bruins
The heart and soul of these projections lives in a colorful, headache-inducing Excel spreadsheet, where I input a couple thousand numbers as data. 2014 went okay, 2015 was below average. In 2016, I’m back with a vengance, and more numbers than ever before. I will be submitting these projections to the Bracket Matrix to see how I stack up for the second consecutive year.
This year’s formula draws inspiration from the BCS, which gave teams a score using computer rankings and data from multiple polls. For the 3rd installment of my bracketology, I am incorporating RPI, strength of schedule, win-loss record, good wins, bad losses, the AP poll, and rankings and strength of schedule evaluations from KenPom, Sagarin ratings, and ESPN’s BPI.
Using these polls will hopefully serve as a substitute for the eye test. I tweak how much weight I put into each of these categories from week to week if something looks off (for example, if the Bruins had a .500 record and my formula had them in as a 9 seed while experts and common sense alike said that UCLA had no shot, I would change the formula so that a team’s win-loss record was a more significant factor). I will never, however, change a team’s score arbitrarily and by standards that differ to every other team I evaluate.
Before you go to the next slide to see the actual projections, keep in mind that I may not necessarily agree with what the formula tells me, and that projections are made as if today is the last day of the season. All data is accounted for through Monday’s games (so Xavier‘s loss on Tuesday was not taken into account, for example).
Alright, after 6 hours of data collecting and spreadsheet combing, I present to you the first bracketology of 2016.
Next: The Bracket
Click to enlarge
The Bracket
Here I have assembled a possible outcome for the NCAA Tournament. Starting with the four #1 seeds, I estimated who are the top four teams in the nation and what region they would conceivably end up in.
As of now, the four #1 are Oklahoma, Villanova, Kansas and Xavier. This is the easy part. From there is where I started inserting the lesser ranked teams. Though it is conceivable to have a four #2 seeds line up, the closer I get to the lower seeds, the more strenuous it became to place each team in a appropriate seeding and region.
The formula used will be explained in the next slide, but from this, you can see how difficult it can become for the Tournament Selection Committee to complete their task.
Not to mention factoring in “bubble teams” and the “last four out”.
Next: The Breakdown
The Breakdown
The advantage of using a formula is that it quantifies just how close a team is from being in the tournament or from being out of it. My formula always tends to center around 0.00 as the cutoff mark for an at-large bid. A team with a negative score is a below average bubble team.
There are 351 teams that play in division I. Inputting data for 351 teams would take an eternity, so I only evaluated teams that meets at least one of 3 qualifications:
- Currently lead their conference (RPI used as tiebreaker)
- Are in the top 125 in RPI
- Are in the top 100 in KenPom, Sagarin, and BPI (because RPI can be unreasonable at times, see: Houston, Nebraska, Oklahoma State)
The number of teams I evaluate grows smaller as the season progresses. But for now, things are still wide open, so more teams deserve a close look.
More from Go Joe Bruin
- UCLA Football: It’s time for the nation to meet Dante Moore
- UCLA Football: Where are they ranked heading into week 4
- UCLA Football: Position battle breakdown for Utah showdown
- UCLA vs. Utah: Location, time, prediction, and more
- UCLA Football: Highlights from Chip Kelly’s appearance on the Jim Rome Show
One more thing: you may notice that Cal has higher score than Florida State, but is projected as an 11 seed while the Seminoles are projected as a 10 seed. I fill out the bracket using the official “2015-16 NCAA Division I Men’s Basketball Championship Principles and Procedures for Establishing the Bracket,” which prohibits any teams from the same conference that played each other twice during the regular season and conference tournament until the Sweet 16.
Because we don’t know if a given pair of teams from the same conference will meet a second or third time in the conference tournament, I set up the bracket so that no two teams from the same conference can meet until the Sweet 16, regardless of how many games they played against one another. Because the ACC already had a 3 seed in the South, a 6 seed in both the Midwest and the West, any ACC team that was an 11 seed had to play in the East.
Because of the number of conference winners that would not have received an at-large bid otherwise, each of the last 4 teams in, including Syracuse (ACC) had to be an 11 seed. Thus, I had no choice but to switch Florida State, the best 11 seed, and Cal, the worst 10 seed, to avoid the inter-conference conflict.
Next: The UCLA Outlook
The UCLA Outlook
There are three big reasons as to why the Bruins currently sit on the wrong side of the bubble. The first is their win-loss record. An 11-7 record will impress no one on the selection committee, and won’t help them when being compared with other bubble teams. Here’s a deeper look at the bubble as it stands:
The Bruins’ second big issue is the steady deterioration of their strongest wins. Kentucky has fallen all the way to 18th in RPI, Arizona dropped out of the top 25 (28th), and Gonzaga is out of the top 50 (55th).
The third and biggest contributor to the Bruins’ downfall in these projections is their computer rankings. On the surface, the biggest flaw of these week’s formula appears to be that it places too much weight into the computer rankings (sorry, Vanderbilt), but that doesn’t change the fact that these metrics exist and are used by the committee. They are worse than every team projected to make the field in terms of KenPom, Sagarin, and BPI rankings.
Next: The UCLA Basketball All-Time Team
There is one surefire way to reverse all of these issues, though: win more games.