How can R&D managers pick the best projects? In particular, how should they compare projects with different distributions of possible outcomes and different amounts of risk? Our approach to this question begins by addressing a special class of projects -- those that have equal expected values, but varying levels of uncertainty and risk. We use this intentionally simplified framework to highlight some general insights about why R&D projects may be less valuable to the firm than R&D projects. The intent is not to provide new modeling methods, but rather to shed light on underlying principles that apply, no matter how complicated the analysis. We show that if two projects have the same expected payoffs and the same costs, but different risks, and different ranges of possible outcomes, a conscientious R&D manager should bet on the long shot and choose the riskier project (if he or she must choose only one). We further show that this result often holds even for managers, and that if a manager can choose several projects, there are benefits to having a portfolio of risky projects. This initially surprising answer that riskier is better hinges on the fact that if the R&D phase of either project fails, the large follow-on investment for commercialization need not be made. In other words, if the research fails, then only the initial R&D investment is lost. Thus, even though the so-called project has less chance of failure, the riskier project cannot lose more money than the safer project. On the other hand, more can be gained with the riskier project because of its chance of exceptionally high values in its commercial stages. Hence, the riskier project has a higher expected payoff if the R&D is successful. Related Literature Our results extend ideas presented in two Harvard Business Review articles. Hodder and Riggs argue that correct valuation of a long-term investment, such as R&D, requires explicit consideration of the risks of three distinct project phases: research, product development and sales (2). Our insights also hinge on separate consideration of the research and commercialization phases of a project. We show analytically that a project with more risk has a higher expected value in the commercial phase and that this higher value outweighs the lower chance of success in the research phase. Kester argues that investments that create growth possibilities for a company are analogous to call options on securities (3). (This idea is also mentioned by Brealey and Myers in 4, Chapter 7 and by Mitchell and Hamilton in 13.) As a result, Kester finds that risk is a positive factor in valuing these investments. Our logic and results for R&D projects are consistent with Kester's view. We also rely on option valuation for our basic intuition, and illustrate and expand the insight using decision analysis. R&D is analogous to a call option in the sense that it creates future opportunities (such as development of new product lines or efficiency improvements) without committing the company to the full investment. We go beyond the arguments of the previous articles to show that even risk-averse managers will often prefer R&D projects if they consider the separate phases of the project and account for the option value of future opportunities. We also note that a company can employ this risky strategy for individual projects while reducing the risk of the company's R&D portfolio through diversification. What Is Risk? Many alternative definitions of risk have been proposed, ranging from simple statistics of a probability distribution (e.g., range and variance) to more general measures that include the decision maker's aversion to potential bad outcomes (e.g., utility functions). For example, Rothschild and Stiglitz compare four specific risk measures (5,6). Here, we use an intuitive definition of called downside risk, which is the expected loss if a project is not successful. …