To increase the success rate you have to increase the failure rate

I was at the Berry Charity Chook Auction on Sunday and found myself sitting in the shade next to a chook enthusiast from Woodhill Mountain named Julia. We were talking about the value of letting kids injure themselves in order to learn, and she suddenly said:
"To increase the success rate you have to increase the failure rate."
I said: "Did you think of that yourself?"
She said: "Yes."
I said: "Just then?"
She said: "Yes."
I said: "Can I use it?"
She said: "Yes."
To increase the success rate you have to increase the failure rate. Something to carve on the heads of politicians and bureaucrats who are terrified of risk.
Failure is good. It's how you learn. The only real failure is a failure to learn from experience.
Which reminds me of my friend Geoff Brown's comment on a previous post which read: 
Lately I have been communicating the need for a mindset shift to tackling these complex problems. Like you say in this post, we need to be trying lots of different things and be aware that most of them will fail. Dave Snowden (Google Cognitive Edge) talks about a shift away from Fail-Safe strategies (where control of outcomes is assumed) to Safe-Fail strategies (where failure won't end in disaster, but we quickly learn from them). I like Clay Shirky's quote on this when he talks about the complexity of getting stuff to spread on the internet ... "We need to learn to try lots and lots of new things and fail informatively so that you and others can find a skull on a pikestaff somewhere".