We find that for one-third of grant applications in 2009 to Australia's National Health and Medical Research Council, success is random owing to variability among peer reviewers. Increased competition for restricted research budgets means we must rectify this element of chance in selection.

A quota limiting the number of proposals per applicant would thwart researchers who have a high success rate, while improving the odds for others. Barring unsuccessful applicants for one 'cooling-off' round is another idea (Nature 464, 474–475; 2010).

Simplifying the application process would reduce costs for both applicants and peer reviewers (for example, some funding agencies request superfluous information). It would help in recruiting good peer reviewers and cut the likelihood of random decisions.

Funding for projects could be retrospective, as in some UK research institutions (D. F. Ball and J.Butler R&D Mgmt 34, 87–97; 2004). Researchers would complete their research before applying, and then use the award for their next project.