Peer reviewers for biology grant proposals submitted to the US National Science Foundation (NSF) do not accurately predict the apparent success of projects, according to a study (S. M. Scheiner and L. M. Bouchie Front. Ecol. Environ. 11, 406–407; 2013). Reviewers' scores and rankings for 41 environmental-biology projects funded by the NSF in 2002 did not correlate well with productivity measures, including the number of publications produced by 2012 and the mean number of citations per year. But reviewers do provide value by weeding out flawed proposals and suggesting improvements, says co-author Samuel Scheiner, programme director for the environmental-biology division of the NSF in Arlington, Virginia.