The problem with fellowships is that the longest I know of only go for 5 years and most of the junior ones are only for 3 years. This is just not long enough to take the chance on failure.
If you want to encourage young talented people to tackle really hard and high risk problems you need to provide funding for at least 10 years.
I would go as far as adding a restriction that recipients of these high risk fellowships can’t publish in anything but the absolutely top tier journals (Nature, Science, Cell, etc) for the first 7 years so they don’t get distracted from the high risk work that other scientists can’t afford to do.
We will move to a system where we only ask 'can this question be answered by the scientific method?' If we use the scientific method to answer these questions we will uncover secrets of the universe that would never survive in the current system humans have designed.
The best scientists do not design their experiments with a goal of publishing in top tier journals. And as a community, we should not be encouraging this behavior.
The best scientists investigate questions out of pure curiosity and as a side effect publish the most impactful research.
Young talented people are already tackling really hard and high risk problems. If this is truly what young scientists want to do, they will find a way. 10 years of guaranteed salary is not the solution, but giving everyone the opportunity to be a scientist may be.
I love your idealism, but having climbed the greasy pole to academic tenure (I have since left science) I can say that you really can’t afford to work on the high risk projects. When you are a junior scientist the more senior scientists control what you can work on and they want you generating publications. When you are a post doc you can’t afford to not get publications out and hence you have to work in areas which will produce results within 2 years. Once you are a senior scientist you are forced to work on projects that will generate x number of papers within the 3 year grant period or else you wont get any further funding. At every stage of your career you are effectively forced to work on projects that are guaranteed to generate publishable results within the next 2 years.
The reason I suggested that the scientists awarded one of these 10 year fellowships are not allowed to publish in anything but the top journals is to remove the pressure of publishing off the fellows and allow them to concentrate on solving the difficult problems. The only reason I suggested an exception for the top tier journals is in rare case the fellow made a lucky breakthrough - we don’t want them sitting on some important result for years just because they not allowed to publish.
I have many high risk/high reward projects that I would have loved to have worked on when I was a professional scientist, but I could not afford to do so. I knew if I failed to produce consistent publishable results I would be out of a job. Even once I had tenure I could not in all honesty ask my students or post docs to work on projects with a significant risk of failure. We have created a system where we have a high probability of making incremental progress, but almost no chance of big breakthroughs.
The problem you face running such a trial is those scientists that control the current system are not too interested in trying anything that might take away their control. They love to mis-quote Churchill and claim their is no better way than the current system, but they are none too keen to try any alternatives.
Fellowships typically fill the roll of paying for long-term, high risk research.
If it needs to be weighted away from established researchers, make it a lottery.
Current funding models are optimised for convergence on leading hypotheses. A lottery would increase coverage.