# High-risk research is a malnomer The idea of `high risk` research implies a known distribution of outcomes and (rationally) leads to risk management and optimizing for expected value. However, truly new research is *high uncertainty* but not necessarily high risk. [[Donald Braben]] uses the example of a skydiver. While skydiving is considered a ‘high risk’ activity, few skydivers think that their particular skydive has a high chance of failure. (Otherwise they probably wouldn’t do it.) They know their own skills and have checked their own equipment. Yes, you might say, but people have done tens or hundreds of millions of skydives. We actually have enough samples that we can model the underlying distribution. We know that skydiver’s risk better than she does. However, the idea that *every* skydive is drawn from the same underlying distribution is patently false. In this situation, you and the skydiver *disagree* about the underlying outcome distribution. That is, it’s an *uncertain* situation. However (and this is going beyond Braben) skydiving is a sufficiently constrained domain that perhaps a master skydiver could know an underlying possibility distribution that both you and the skydiver would agree to. In this case skydiving becomes a knowably uncertain situation. Research (which here I’m using as shorthand for never-been-done-before work) on the other hand is *unknowably* unknowable - it’s impossible for anybody to say a priori what the distribution of outcomes from the work looks like. As Braben points out, the honest researcher probably has both the best sense of the underlying distribution and has the most [[Skin in the game (SITG)]] (they’re the one wasting years of their life if the project goes nowhere.) `On the other hand, if it were decreed that the risks of skydiving had to be managed by those paying to enjoy the spectacle — that is, the spectators had to take full responsibility for anything that might happen — it is most unlikely that any jumps would take place` `Why should academic researchers embark on major quests *if their funding agency expects most of them to fail?* That is what sponsoring high-risk research implies` `In research, the adoption of “best value for money” policies means effectively, that funding agencies give the highest priorities to the projects deemed to have the lowest risk of failure.`([[brabenScientificFreedomElixir2008]]) As Braben points out, the language of risk applied to research incentivizes either to suppression of high-uncertainty (ie. Actually novel) work or (in the case of programs that are supposed to support “high-risk” research) encourages researchers to stop managing their own risk, which *actually* increases the chance of failure, further poisoning the perception of high uncertainty work that gets *called* high-risk. ### Related * [[Uncertainty always involves risk but risk does not always involve uncertainty]] * [[Malnomers are terms for things that actually create actively worse outcomes]] * [[People giving out grants try to derisk them as much as possible]] * [[If you need to have the possibility of near infinite returns on VC investments how is it possible to invest in higher risk things and expect a return?]] * [[Evaluating transformative research programmes: A case study of the NSF Small Grants for Exploratory Research programme]] * [[Knightian Uncertainty occurs when you hit the limits of theory]]