If markets work, you don't need to incentivise - Risky investments pay out more when they do succeed.

If markets don't work yet, fix the failures with regulation until they do. No need to incentivise behaviour through taxation.

edit: my principal reason for believing it's better to do things this way round is cost: re-writing the rules only costs you (us, the government) money once, and it works forever. Writing an exception in costs you money once to write the exception, plus every year in lost revenue.

TL;DR - laws = code, exceptions = CPU-hours.

Oh dear yet another junior coder blindly trying to apply simplistic solutions from his experience in coding to the real world.

Answer this do you really think that investing in government bonds (say 3% consols) should be taxed the same as taking a risk in a start up?

NO because no one would invest in any company of the economy woudl stagnate.

Actually I was trying to use a coding analogy to explain to a technical audience. I apologise, should all future contributions be in the form of a mathematical proof? perhaps latin?

Seriously, you have no idea what an economy would do given different changes, as no one has tested it - we don't have any good[1] evidence for anything at all in the sphere of economics, except perhaps that people don't behave as economists model them.

As I said in my previous answer, different levels of risk are given different levels of reward already; changing the tax level of each won't make any difference, relatively speaking. (I won't bother explaining how VC funds spread risk to make it acceptable to lay-investors, because you doubtless already know.)

If markets are to be trusted for anything, putting correct prices on different levels of risky investment should be it. If not, as you say, no one would ever invest, and we would need to ask the government to run all our industries for us.

[1] from the point of view of physicists who actually care about uncertainty values.