2024 March Madness

The 2024 model had a new format where the model went through thousands of matchups and selected viable brackets based on criteria. - an iterative model. The model also consolidated the viable brackets into one singular composite bracket. 

The composite bracket had Houston as the champion. I offered free picks, and then Bronze, Silver and Gold level picks. The last three picks were for-pay picks. The free picks had Duke winning the tournament (if someone in the same region beat Houston). This time, 51 people received the free picks from the website and 4 people bought the for-pay picks.

The 2024 model did not fare well. I went all in with Houston, and that was not a good choice in hindsight (which is 20/20). In the Sweet Sixteen game against Duke, Houston's point guard, Jamal Shead, went down with an injury. They lost that game, and my model took a serious hit. The composite model (Gold), scored 640/1920 ESPN for 47%. The free pick version score a 790/1920 for 62%. I refunded the money to those who paid for picks. 

Would Houston have gone on to the championship game if Jamal Shead did not sustain the injury? That is hard to tell. Jon Scheyer, the Duke coach, said this about Shead: "Look, Shead is the ultimate competitor and winner, and I saw it firsthand. We scrimmaged Houston last year, and just seeing what he does for their team and watching them on film, there's no replacement for him." His own coach, Kelvin Sampson, said of Shead: "I doubt any team in America has - maybe Edey from Purdue - (a guy) that means as much to their team as Jamal means to this team. There's just not another Jamal. He was the best player on the floor tonight. He's been the best player on the floor every game we've played this year except a few."

The model did not do as well as anticipated, but circumstances did impact the model adversely. Hopefully, in 2025, a better model will overcome that and more. On to 2025!