Who gets the ball?
Estimating which team will control the opening possession.
FIRST BASKET LAB · ABOUT US
A small moment.
A big research question.
Can basketball's first point be predicted more reliably using historical data? First Basket Lab is an independent project exploring that question through software, statistics, and LLM-assisted development.
Who scores first?
Estimating which player scores the game's first points, including made free throws.
How might they score?
Estimating the selected player's scoring method if that player scores first.
What we're building
First Basket Lab brings previous-season basketball data, statistical modeling, and large language models (LLMs) into one research and entertainment project. We study patterns in opening possessions, player scoring history, lineups, and matchups, then compare our forecasts with what actually happened.
The aim is to learn whether those patterns can improve first-point predictions over time. A probability is an estimate, not a promise: basketball is unpredictable, and even the highest-ranked player will often not score first.
The developer behind the lab
The project is led by a software developer with a degree in computer science and more than 20 years of experience in software and web development. First Basket Lab combines that experience with LLM-assisted coding, research, and deployment to build and refine a basketball prediction platform.
LLMs help us develop and investigate the system. The probabilities displayed on the website are calculated by our statistical models using available basketball data; they are not simply answers generated by a chatbot.
Research we can check
Our How it works guide explains the three forecasts, their inputs and how we grade them, including why first points and first field goals can have different winners.
We test ideas chronologically, using earlier games to inform later predictions. Historical replays are reconstructed after the games and are labeled separately from predictions saved before tip-off. Keeping that distinction lets us evaluate the system without presenting hindsight as a live forecast.
We compare model versions, report results and data limitations, and preserve saved pregame predictions so they can be checked against later outcomes. Research findings do not automatically replace the active models.
For research and entertainment
This is an experimental, independent project, not a sportsbook or a betting advisory service. We do not accept wagers, guarantee outcomes, or promise financial returns. First Basket Lab is not affiliated with the NBA, its teams, or any sportsbook.
Explore the Model Research page for more detail, review past predictions and results, or send questions and feedback through our Contact page.