Chapter 3 Traditional Season Simulations

With a pre-season game win-prob model in place, we now theoretically know the probability of the home team winning the game in any matchup between two teams. For my simulations, I used a “fake” schedule by using the 2013-14 NBA season (not included in modeling process). Each game has a win probability associated with it. I simulated this “fake” season 10,000 times to obtain both win total average predictions and average playoff probability for each team based on all of the simulation results.

---

3.1 Simulation Results

First, I needed to gather the variable data for this past season (2019-20 Season) in order to make score differential & win probability predictions. Next, I simulate 10,000 seasons in the traditional format (No Play-in). The code to this and the results from the simulations is shown below.