After running a simulator to project the tournament, it’s time to review how this tool performed. To go in order of the previous post, the simulator thought the most likely scenario was that one #1 seed would make the Final Four.
The most likely scenario proved to be correct. I also made predictions of number of wins by conference. Here’s a summary of performance versus actual.
This was a surprisingly good approximation, with the SEC being the most off due to Kentucky getting 6 wins themselves. One other point to note is that the original post incorrectly assumed Cal beat South Florida, so that ended up changing the expected values a bit. Onto the prop bets:
Bet #1. The winner is not a 1 seed. (-110). This one flopped. Although only one #1 seed made the Final Four, Kentucky was the best team and won its games in fairly non-dramatic fashion. LOSS
Bet #2. At least one 15 seed advances. (+350). Whoa. Two advanced! Wonder what the odds on that were? I threw in a disclaimer about the simulator probably thinking a little too much of Lehigh’s chances against Duke (23%) to win, but they were clearly not overmatched. Norfolk State was a complete fluke. WIN
Bet #3. A Big 10 team wins the championship. (+220). Nope. Ohio State made the Final Four but flamed out. Michigan State probably got a bit overrated based on their run on the Big 10 Tournament. Wisconsin was in good position but some questionable late-game coaching against Syracuse did them in. LOSS
Bet #4. The combined seeds in the Final 4 is Over 10.5 (-115). This hinged on a low seed making it, but that didn’t happen. Generally speaking, the rankings thought the 1s and 2s were pretty close and then the field from 3s to 9s was fairly flat. If Florida had prevailed over Louisville this would’ve been a winner, but they didn’t and the number fell on 9. LOSS
Bet #5. The ACC stinks or (separately) North Carolina (+300) and Duke (+145) will not make the Sweet 16. UNC did advance and Duke didn’t. Loss, Win
Summary: I didn’t really scale anything, so there’s no way to objectively judge the outcome. If putting 100 on each bet, the net would’ve been +95, which isn’t bad, but that wouldn’t have been a logical way to scale things. If more lines were available, the results would’ve been different. Needless to say, the simulator did really well at predicting wins for each conference, which is important in understanding relative strengths of each conference and is very promising for the future.
Wins: +350, +145
Losses: -110, +220, -115, +300