I've made some changes for this year. For instance, last year RPI carried heavy weight in my projections. That will be scaled back. This is partially in preparation for the selection committee evaluation changes coming in 2019, but also because I over-valued it in 2016-17.
I've also experimented a bit with pre-season predictions, using the past five years of game results, while also sprinkling in some recruiting and coaching changes. From a macro view, I more pleased than displeased with the results. However, it's not without flaws (Duke as a #5, Stony Brook as a First Four out). Regardless, I will be phasing out the predictions by December.
Happy Basketball New Year, and RIP 24 hour Marathon.