Abstract: A computer-implemented method enables the selection of improved algorithms that solve technical problems, such as data processing problems, computational science problems, inverse problems, and optimisation problems, using a data network of connected nodes. The technical problem is an optimisable proof of work problem (OPOW) deployed at a network node as part of a protocol requiring OPOW to be performed. An algorithm implementation is run on the node to generate a solution to the OPOW problem, and the solution is sent for verification. Reward tokens are sent to the ‘innovator’ entity that creates or publishes the algorithm that solves the instance of the problem, as an incentive to create and publish improved algorithms. Reward tokens are also sent to the ‘benchmarking’ entity that runs the algorithm implementation to generate a solution to the instance of the problem, as an incentive to identify or select an improved algorithm.