You'd also have to define what constitutes "clean" (or what negates it).
The definition of what "a clean problem" is can actually vary depending on the desired effect. The bar can be put very high, so that bonuses are rarely awarded. Or, it can be comparatively low, i.e., just positive GOE for all technical elements and no deductions (minor mistakes don't count).
Also, judges can be involved into the bonus-awarding process or, contrary, not involved at all - calculating bonuses automatically when all conditions are met.
Honestly, I don't think that this is the most interesting part of discussion. The most interesting part is...
It would surely encourage skaters to leave out their most risky elements, to intentionally double jumps later in the freeskate when they're tired, etc.
Especially if the clean program bonus is worth more than the second-half bonus.
...this. Could it influence the program-building strategy?
Currently, the only way how to scoop the most points is to put in the program all "pointiest" elements up to the degree that it becomes physically impossible to perform everything at the right quality.
This is because we have scoring tools that promote difficulty/quality of separately-taken elements and components. So, the question: are there any tools, already existing or to be invented, that could promote a qualitatively delivered whole program?
And, of course - do we need any?
?
