INNOVATIONS IN SPORTS ANALYTICS
Bringing new models to the professional levels of the game.
Our philosophy
DSA Labs has long recognized that the most valuable insights often require working with limited data. Studying “Big Data” sets from across multiple leagues and teams is of little value in learning about your particular group of players!
We pride ourselves in answering the toughest questions in sports with any amount of available data. We make “Small Data” work! Check out some of our favorite innovations in data science below.
OUR EDGE
Customized A.I.
Working with our colleagues at the University of Virginia, we have developed algorithms that can automatically learn the most effective player progressions directly from game video. Most importantly, the method requires minimal training data and computational resources. The result is “practical A.I.” that can work for any team at any level of the sport.
In the example shown, player positions were extracted from game video and used to learn the most effective approach to “building out of the back”.
Player compatibility
Some players just work well together. Our “Player Compatibility Index” (PCI) was designed to find those complementary pairings for you. Based on a probabilistic analysis of possession sequences, the PCI tells you which players are combining to give you the highest production on the field. Use it to determine the right formation, set a starting lineup, or find the right substitution patterns. The PCI is a remarkably effective tool for game day management.
Player impact
Arguably the most important goal of an analytics program is to identify each players’ contribution to a team win. This is particularly challenging in sports where scoring is limited. DSA’s “player impact” index overcomes this difficulty - providing a direct measure of each players’ contribution to victory! Battle tested in both college (like the example to the right) and pro games, the impact index is one of our most innovative tools.
Performance MOdeling, Tracking, and Prediction
Every player, coach, and parent wants to know if that hard work is paying off. Whether it’s fitness in returning from injury (like the example to the left), passing efficiency, or 1v1 success rates, understanding player trends over time is a critical piece of any analytics program. StatLink software allows you to track any player metric over a month, a season, or a career.
REGRESSION TOOLS
How does practice translate to game? Whether you coach at the recreational or professional level, the answer matters! We have developed a number of modeling tools that clarify the linkage between practice & game. Use the results to understand which players are over/under-performing on match day, or simply see if your training is positively impacting game performance.
ongoing research & development
The DSA Analytics Team continues to develop practical answers to the toughest questions in sports. The results of our R&D are piped directly into StatLink to provide our users immediate access to advanced analytics.
Read the latest from DSA Scientists James and Jonathan Nichols on the key factors influencing player improvement during training.
Customized A.I. that learns the most effective progressions directly from raw video data.