Geoff,

The NSB's yearning for more high-risk/high-reward projects is laudable and laughable at the same time. Science (and science funding in particular) is an intensely social and socialized enterprise. As much as agencies try to promote high-risk projects the culture of science, and the process of peer review, are nearly antithetical to risk. In a world where 1 in 5 projects are funded, a single reviewer's negative comment is enough to kill a proposal. The simple process of building consensus almost by definition reduces the liklihood that a maverick idea gets support. And, if a maverick idea could get the support of an entire review group - how maverick is it anyway?

This is the same reason why committees are lousy at making works of art and why so many Hollywood movies (written, directed and produced by committees) are atrocious. The spark of genius is solitary.

My observation is that maverick ideas that end up succeeding are first nurtured in small protected enclaves - perhaps by a benevolent department chair who throws an additional $50K of discretionary money at something. Or some poorly supervised program manager in the depths of some department gets a bee in his bonnet about something and continues to sponsor work in that area long after the "smart money" has moved on. That, by the way, is how we have ended up with economical solar cells - a single maverick program manager at JPL continued to fund manufacturing technology for photovoltaics into the early 1980s - when the rest of the support evaporated for the subject (in the face of Reagan and $20/barrel oil...). This crazy dude - whoever he was (I'd love to hear the real story on this - the version I have is from the founder and CEO of SunPower) continued to fund research into how to economically slice up silicon boules - the group he supported ended up hitting on the wire saw - a totally catastrophic breakthrough. The rest, as they say, is history. NSF would have NEVER funded wire saw technology for silicon boules in 1981 - not because it wasn't maverick - but because the priorities had simply moved elsewhere.

So I think there's a myth and mystique to "high risk" research. People think high-risk means a radical approach that is a long shot. I think high risk is also investing in an area that appears to be dead.

If the NSB wanted to fund high-risk research they should recognize that they will likely do a lousy job of picking such winners and losers. They need to avoid the peer review process as it is and try something different. How about funding a contest for all crazy ideas anywhere - the winners get $2M over 3 years to make something happen. Or, talk to DARPA about how they select high-risk high reward ideas (though some complain that DARPA has gotten soft and risk-averse in its dotage...). Or look at how the Howard Hughes Medical Fellows are doing things - HHMI shovels big heaps of money on those guys and (few) gals and leaves them alone for 5 years.

Oh, and how about something else: if you want high-risk research, why not examine why we are producing so few high-risk researchers...