This paper describes the key characteristics of high-risk/high-reward research (HRHR), which has gained considerable interest from policy makers as a way to promote the development of new, ‘out-of-the-box’ ideas. It identifies three dimensions that are accentuated in HRHR research: higher levels of basicness, generality and novelty. These knowledge characteristics are commonly associated with market failure and research that requires public investment because it has large spill-overs, long time horizons and high levels of uncertainty. This is illustrated with examples of specific discoveries embedding each knowledge characteristic and the application of appropriate quantitative measures. The paper concludes with the computation and demonstration of an indicator of novelty that may be particularly well suited for the monitoring and evaluation of HRHR research policies.