What’s up CougarBoard, some of you might remember I did a machine learning project to predict March Madness last year for one of my classes.  Quick update on me, I finish my undergrad in April and have been accepted into the Master of Math program here where I will focus my studies on applied and computational mathematics particularly for machine learning and data science purposes. I’ve been wanting to do this ever since the tournament got canceled earlier this month, but Kaggle, my data provider, didn’t release the data until early this week, and I had three exams and a final presentation this week, so this was the first time I could get at this. 

For this years March Sadness, I will go through some of the assumptions I had to make because of the incomplete nature of the data, i.e. the lack of conference tournament champions, which bracket we used and some slight modifications we implemented into the model, but before I get to that, I want to address something from my post a year ago. 

I said the following: “This year …  UVU was 108th in defensive field goal percentage, 285th in creating turnovers, and 69th in defensive rebounding. Honestly, none of these numbers from Coach Pope’s team stand out to me. His team was below average on defense in a below average conference. It is hard for me to get on the Pope train because I don’t see him as succeeding at BYU without an absolute emphasis on, and ability to implement a fundamentally successful defense which he did not come close to doing at UVU.”

I've watched more BYU basketball this season than I have since the Jimmer years and there were a couple of games where I sat very close to the floor; and let me tell you, this team can flat out defend. I don’t know what the difference is between this BYU team and Pope’s UVU teams, if it’s raw talent, desire, coachability, etc. - but these guys are awesome defensively. Kolby Lee and Yoli owned the paint defensively against the Zags in Provo, against USF and Santa Clara, the guards were awesome on the perimeter, and the same can be said for the USU game.  I have been very impressed with these guys and was flat wrong about Coach Pope, something I am very happy about. For reference, the defensive numbers for this season: Defensive rebounding 86th, 135th in creating turnovers, and 137th in defensive field goal percentage. — These numbers are slightly skewed because we were several games without our best rebounder, and there were a couple of blow outs in that stretch, Kansas and the Zags come to mind, so while these stats don’t jump off the page at me, they are respectable and in person, the cougs were nasty on defense. 

So Coach Pope, if you ever read this, 1 - I would love to be your stats/machine learning / waterboy during my time as a masters student, and think we could get some really awesome results, and 2- I am 100% on the Pope train! You turned a skeptic into a believer and it has been so fun to see how wrong I was. 

OK - so I wanted to create a way for there to be conference champions again in this go around other than Utah State and Gonzaga, so I went through and looked at every conference tournament for those who had already started and those who had yet started, I generated a weighted random sample to pick a champion. Basically took the Vegas odds, normalized the probability of each team in the conferences that had more than 1 bid, and took a weighted random champion from that group.  This was not a perfect assumption, for example, when I did the SEC, it gave me Mississippi St, which had something like a 3% chance of winning according to Vegas. Joe Lunardi picked them in the “Next Four Out”, so I had to run it again and got LSU, so like I said, not a perfect way, but because each conference has a different structure to their tournament, I did not want to have to program that for each conference and did this instead… call me lazy lol… my conference champions for reference were: {“Utah St”:“MWC” , “Gonzaga”:“WCC”, “Creighton”:“Big East”, “Duke”:“ACC”,  “LSU ”SEC”, “Penn St”:“Big 10”, “Kansas”:“Big 12”, “USC”:“Pac 12”, “Dayton”: “A-10”, “Houston”: “AAC”} 

We added page rank to the algorithm, which was the most predictive feature other than the winning team’s seed.  Basically ranks the qualities of wins doing so by saying, x Team beat y team by z differential, and y team beat w team by z differential, and so on. This provides a quality of win metric / strength of schedule.  The top teams were: ['Kansas', 'Baylor', 'Villanova', 'Duke' 'Oregon', 'Michigan', 'Colorado', 'BYU', ‘Creighton', 'Kentucky', 'Gonzaga', 'Virginia', 'West Virginia', 'Louisville', 'Wisconsin', 'Seton Hall', 'Maryland', 'Auburn', 'Butler', 'Michigan St.', 'Florida St.', 'Providence', 'Rutgers', 'Penn St.', 'Ohio St.', 'Iowa',  'San Diego St.', 'Utah St.'] That’s the top 27. - A quick caveat about Page Rank, Utah was ranked in the top 30 probably mainly because they beat us and Kentucky, so it’s not a perfect system, but it is a good indicator of the quality of opponent that a given team is capable of beating, because the sports adage is, “a win is a win”, but I think that we all know that not all wins are created equal. 

Ok So the bracket we used was the last bracket from Joe Lunardi. Some interesting takeaways, all three WCC teams win their first round games,  something that does not surprise me at all seeing as each BYU and Gonzaga had a PageRank in the top 10, and St Marys was in the top 50, but St Mary’s had a tough draw with Baylor and gets bounced second round.  I was glad to see BYU with the edge against San Diego State in a rematch with Yeoli this time, and then losing to the Zags in the Elite 8. It was good to see Utah State win their first round game, which I was kind of surprised at, but they did get a lot of love with their win over SDSU from Page Rank. I thought it was also funny how it was one of those, “Don’t pick against Butler” years, I swear they always make a run. 

I did think it was interesting how close my bracket was to the one on ESPN that has BYU in the National Championship game, we both had all three WCC in the second weekend, we both picked Providence in the first round and USC, that article had Maryland over Dayton, I had Butler, but we both had Liberty and Cincinnati in the first round  We had some differences, but seeing all of this, maybe that guy wasn’t as crazy as I thought when I first read it, and like I said before, I know who I would have picked if BYU had gotten to that point, I don’t care what the computer said lol 

So an Elite 8 finish is where I have the Cougs going, I honestly have to like our chances against the Zags . I know I have my blue goggles on, and I know Marc Few can flat out coach, but gosh, I like our chances, especially since BYU has a higher Page Rank than the Zags, even though the algorithm says that the Zage would be the runner up. Something that the BYU Sports Nation guys have mentioned is, given the way this season ended, it gives more significance to the truly iconic moments we saw. I have said my whole life, that the best sporting atmospheres I have ever been to - in my albeit limited experience- are #2, BYU v Utah football Max Hall’s senior season, and #1 the BYU vs SDSU game in Provo when Jimmer was a senior.  This years home game against the Zags is now my #2, and was special for some slightly different reasons. The added drama of senior night, the sheer noise of the crowd, what it meant to this team and program, it was just incredible! I was lucky enough to sit right in front of Jake Toolson's dad, and I don’t think I will ever forget the look on his face every time his son hit deep 3 after deep 3!! It was one of those sporting moments that is hard to put into words, it was almost as if Mr. Toolson was ready to get in the game himself. It really is a shame we couldn’t get to see how long the music would play for these guys and how many more magical moments we could have gotten.

So that is a slightly depressing, slightly comforting outcome for the model.  It truly is one of the highlights of my year to watch March Madness each season with my old man. I really feel for these seniors, for Teej, Jake,  Zach, Dalton, and for Yeoli. Selfishly, my uncles and I were planning on going to support the team in the first round, and it sucked that we couldn’t go, but even more, it sucks for the players who had dedicated so many years to a singular goal, and had achieved it, and then had it taken from them … In the unlikely event that the NCAA lets them come back for another season, I would hope a lot of them do! It would be fun to run it back, no guarantees, but put this in the back pocket and maybe next year we’ll be even better!!

Anyway, stay safe CougarBoard, much love.

#GoCougs!