Projecting player longevity and progression is extremely hard. Often, people follow conventional time frames. Here, we present a new metric that uses a deep learning sequence model to project players’ future All-NBA probabilities.
By training a machine learning model on historical tweets from Adrian Wojnarowski and Shams Charania, we can generate fake NBA news tweets.
Last year, we created machine learning models that nailed the All-NBA teams. This year, we’re using even better methods to predict the All-NBA teams.
Last season, we correctly predicted the MVP with four machine learning models. This year, we’ll take a better approach to predict the 2020 MVP.
Let’s evaluate contracts in the NBA by applying machine learning to compare how much a player earns versus how much the models expect them to earn.
In evaluating players, teams look for who will give them the highest chance to win. Let’s create a metric based on the win probability given a player’s stat line.
Following the end of the 2018-19 NBA season, many released their All-Decade teams for the 2010s. Different authors weight longevity and peak differently, creating some subjectivity. Let’s use machine learning to approach this objectively.
Publicly available defensive metrics either fail to quantify defense or are hard to understand. Let’s create a better, more transparent public defensive metric, DAVIS.
Each year, more and more “unicorns” enter the league. Let’s define the uniqueness of these unicorns in a comprehensive metric.
Let’s use college stats to predict All-Star probabilities for the top-10 picks in the 2019 draft.