The value of experimentation rises when downside cost is minimized.

Experimentation allows you to find new viable ideas, even if you don't understand why they work.

If an experiment might kill you, you won't do experiments.

But if you reduce the downside cost of a failed experiments, then the cost becomes primarily only opportunity cost...

...but the upside remains!

You can reduce the downside cost of experimentation by for example introducing fire breaks between components (so an explosion in one part won't affect others), or in organizations by having the ludic mindset.