Roster Metrics

Machine Learning Powered Fantasy Sports Projections


Roster Metrics is a platform using machine learning and advanced statistical algorithms to create projections for fantasy sports. The core part of the platform is making projections for NFL, MLB, and NBA as well as project fantasy points for the DraftKings, FanDuel, Yahoo, and FantasyDraft scoring systems. It’s algorithms seek to minimize mean squared error and maximize the Kendall’s Tau rank correlation coefficient for future games, doing all the math for the users where they can focus on managing their fantasy teams and building lineups.

David Israel

Preezma completed the project on time and within budget. Managing the project efficiently, the team has an effective workflow and good communication skills. They exceeded the my expectations.

Founder & CEO at
Rochester, New York

Challenges and involvement

CEO David Israel of Roster Metrics was looking to collaborate with a development team who has experience in developing high performing APIs, a team that’s able to build and maintain a scalable SaaS platform and one with experience of implementing AI within a platform.

As David is an experienced Data Scientist, he knew exactly the qualifications he needed within the team to develop the platform’s requirements. One of the challenges we faced was implementing 3rd party integrations (services) to keep the platform running as close as possible to real time big data. With this, we also developed an AI based sports projection analytics platform based on sports lineup optimizers.

Like many CEOs, David is also a very busy person and he put the trust in us in leading the developmental process. Within a 4 month period we developed Roster Metrics MVP followed by fully developing the platform and successfully deploying it on the market.

Technology stack

R lang




Amazon Web Services