Distributed system for evaluation of strategies in the stock market and the soccer betting industry

Abstract: Development of software and strategies for prices, odds calculation, and risk management in the stock market and the soccer betting industry involves testing functionality using recorded real-world data. This process is called backtesting. If a system for backtesting is represented by software that is run on a single workstation then due to the complexity of mathematical computations, amount of recorded data and variety of strategies — backtesting can require significant amount of wall time. This is a problem because the development cycle increases in time. This paper proposes a design of a distributed system for backtesting that distributes the workload among available workstations considering limitations on the network usage.