On average the expected value at the moment of commercialization of an R&D project should remain constant during the different stages of new product development. Contrary to this intuition however, a systematic, non-constant pattern is found in the average expected value of an R&D project. First, the value declines after the initial screening, then rises after the first market analysis, but on average does not reach the initial level at the final stage of development when resources for market introduction are approved. Moreover, uncertainty about the project value declines over time. The findings suggest a V-shaped value function of R&D projects. The study seems to be the first attempt to make direct measurements of valuing R&D projects through time in a real managerial setting.