Betting on sports events became increasingly popular with the advent of electronic betting, where bets can be placed over the internet or with the help of mobile apps. For bookmakers it is important to accurately predict the odds, not only before the event but also while it is running. Complex statistical models are used for odds estimation, typically using Monte-Carlo simulations. Being able to run these continuously during the event gives a competitive edge to bookmakers, e.g. by adjusting the odds every second.
This case study highlights how a major bookmaker used Xcelerit to implement a real-time sports betting simulation on GPUs. The bookmaker provides a mobile app to allow its clients to place live bets during the game. This app connects to their data center and updates the odds in real-time. The achieved performance level allows the Monte-Carlo simulation to be re-run every second, with the game’s current standings and statistics updated.