This year I finally did try my hand at an analytical approach, and for the moment, those brackets are in much better shape than my personal modifications (which we'll get to later). I ran three different analysis using a regression formula: Round by Round score, Best "Champion" score, and the average of the Round by Round (this one is sucking so I won't go into too much detail, but it is self explanatory)
The Round by Round score was the result of 6 regression formulas, one for how likely the team was to win that corresponding round, meaning the first formula was how likely a team was going to win their first round game, the second formula was how likely they would win their second round game (provided they won their first round) and so on and so forth all the way up to 6, which is the number of wins you need to be the champion. For example; Kansas vs. Penn and Rhode Island vs. Oklahoma: Kansas came up with a first round score of 1.12 while Penn had -3.11. Rhode Island had a -1.42 while Oklahoma had a staggering -3.84 *cough cough*. Naturally, both Kansas and RI were selected as winners (and they won yay). But, lets say it was Kansas vs. Rhode Island and Oklahoma vs. Penn in the first round; Kansas still has the advantage over Rhode Island's score, but here we would be favoring Penn over Oklahoma (although not by much). For the second round though, we would then look at their second round scores. So using this example again (since it is what happened); Kansas vs. Seton Hall and Duke vs. Rhode Island: Kansas's second round score was a -.49, compared to Seton Hall's -6.799. For Duke and RI; Duke had a second round score of -2.696, while RI had a -7.54. Again, had we had the matchups swapped, Kansas would still be winning, whereas Seton Hall would get the advantage on Rhode Island in that instance.
Does this lead to some skewed results? Sure. It hella depends on who you draw as a matchup. Personally I would think Oklahoma wouldn't be bad enough to lose to Penn, but the world will never know. To rectify any potential skewness of the "depends on the matchup" theory, this lead me to create the Best "Champion" score bracket, which was derived by solely taking the best score from Formula 6 and working my way through the bracket in reverse, starting with who was most likely to win it all. So here, spoiler alert, Villanova was the overall favorite, so I dragged them to the championship immediately, regardless of who they played. Then working down the list, we get North Carolina as the second favorite (RIP). Since they do not play Villanova until the Finals, UNC got a free ride there (whoops). Michigan ended up being 3rd, but what stunk for them was that they would've had to face UNC in the Sweet 16. And since UNC already had scored higher, we would chop Michigan's trip up early. This hurts the bracket, but we might have a new finals favorite behind Villanova now.
And to those Nebrasketballers that are curious: In the Round by Round, Nebraska comes in as the 57th most likely team to win their first round with a -2.299. And they came 104th most likely to win it all. Both scores were still behind Mississippi State. Just sayin.
Anyway let us dive into how these brackets are actually doing compared to mine and let's see if we can predict the Sweet 16 onward!
To start with, we'll look at my bracket, which is a humanly modified version of the Round by Round bracket. Aka, I just couldn't stand taking Tennessee into the Sweet 16. And I was right!
Current Score: 420. Remaining: 720
Teams picked in the Sweet 16: 9/16 (Missed on: Virginia, Arizona, Cincinnati, Xavier, North Carolina, Michigan State and New Mexico State)
Elite 8 Teams left: 4/8 (Villanova, Purdue, Gonzaga, Kansas)
Final Four Teams left: 1 (Villanova)
Champion alive?: Yes, Villanova.
I'd like to think my bracket was on the traditional chalky path. The only bold picks I had was Loyola and New Mexico State going into the Sweet 16 and we at least split that difference. The East is still near perfect, but it's a dumpster fire outside of that. And even then, Villanova has a tough task ahead to get to the Final Four by having to get through a rowdy West Virginia team as well as *cough Purdue cough* albeit they're injured. The fact that they're my last hope is sad though, as I don't really like Villanova.
Round by Round Bracket:
Current Score: 400. Remaining: 1120
Teams picked in the Sweet 16: 8/16 (Missed on: Virginia, Arizona, Tennessee, Cincinnati, Xavier, North Carolina, Michigan State and New Mexico State)
Elite 8 Teams left: 6/8 (Villanova, Purdue, Gonzaga, Michigan, Kansas, Duke)
Final Four Teams left: 3 (Villanova, Michigan, Kansas)
Champion alive?: Yes, Michigan
As mentioned above, my actual bracket was pretty much a copy from this one with the only change being Loyola over Tennessee before the Sweet 16. Where my computer decided to slap me in the face was what happens beyond the Sweet 16. Much like the rest of the world, the South Region is completely fucked over. The biggest difference that this bracket decided to take on that I might lose sleep over because I struggled with this decision personally, was the predicted Michigan vs. UNC matchup in the Sweet 16. The computer picked Michigan as the slight favorite, and while I cannot deny that Michigan is always hot late, I also cannot deny past results, which had UNC slaughtering Michigan earlier in the year. The world will never know due to the upset of North Carolina, and now Michigan has a fairly safe ride into the Elite 8, where they could match up better against a weaker-than-usual Gonzaga team (most likely), or get lucky and pull Florida State.
Best "Champion" Bracket:
Current Score: 440. Remaining: 960
Teams picked in the Sweet 16: 10/16 (Missed on: Virginia, Tennessee, Xavier, South Dakota State (lol), North Carolina and Auburn)
Elite 8 Teams left: 6/8 (Kentucky, Nevada, Villanova, Purdue, Kansas, Duke)
Final Four Teams left: 3 (Kentucky, Villanova, Kansas)
Champion alive?: Yes, Villanova
Right now in the best position potentially, and probably one of the only brackets that lucked out on what happened in the South. Call it crazy stupid luck or call it science, but it actually has Kentucky and Nevada meeting in the Elite 8 (Kentucky over Virginia and Nevada both on Cincy and Tennessee). We can easily argue that both Kentucky and Nevada has significantly easier games to reach the Elite 8. Unfortunately, where there are wins, there are losses, and the loss came in the mighty West, because for whatever reason the model loved South Dakota State to go on a run. Additionally with this bracket, the East region is perfect. The Syracuse over Michigan State upset was also predicted in this bracket, and as such, we have an early favorite when it comes down to predicting March Madness.
But with the new results, let's see how each would have done if we reset everything to the current Sweet 16, eh?
Round by Round (so Sweet 16 would be Round 3):
Kentucky over Kansas State
Nevada over Loyola
Villanova over West Virginia
Purdue over Texas Tech
Gonzaga over Florida State
Michigan over Texas A&M
Duke over Syracuse
Kansas over Clemson
Elite 8 (Round 4):
Kentucky over Nevada
Villanova over Purdue
Michigan over Gonzaga
Kansas over Duke
Final Four (Round 5): Actually after re-reading my sheet I realized I made a mistake in my bracket, I should have had Kansas beating Villanova anyway, whoops.
Michigan over Kentucky
Kansas over Villanova
Finals: Michigan over Kansas
Best "Champion" Bracket:
Sweet 16:
Kentucky over Kansas State
Nevada over Loyola
Villanova over West Virginia
Purdue over Texas Tech
Gonzaga over Florida State
Michigan over Texas A&M
Duke over Syracuse (barely)
Kansas over Clemson
Elite 8:
Kentucky over Nevada
Villanova over Purdue
Michigan over Gonzaga
Kansas over Duke
Final Four:
Michigan over Kentucky
Villanova over Kansas
Finals: Villanova over Michigan
So it is looking like both models are going about a similar path, with the only difference being Villanova beating Kansas and winning it all, instead of Michigan beating Kansas. Either way, it is looking like Michigan is in good favor after North Carolina's early exit. Only time will tell as we still have two more fantastic weekends of this tournament! (And more time for my actual bracket to burn in flames and hope that I score the "lucky" 50th place prize in my work pool. Which is not last!)
