Doing dangerous things well is a skill than can be learned and improved upon. Like many skills, doing dangerous things well can be broken down into smaller, more manageable chunks. With this in mind, we use a Risk Mitigation System (RMS) to help us do dangerous things well.
In Step 1, we state the dangerous thing we want to do well. In this case, we wanted to throw knives at a target.
In Step 2, we assess risks and problems , based on a. our personal experience, b. expert opinions, c. educated guesses. To assess the risks of throwing knives at a target, we recalled our own experiences with knives, and our experiences with ricocheting thrown objects; watched videos and viewed diagrams from professional knife-throwers; and guessed at problems specific to our space and goals.
In Step 3, we use iterative escalation to slowly build towards the goal activity with safety and control. To throw knifes, we added one element of danger at a time, simulating the throw with a sharpie, a block of wood, then a knife with the blade smothered in duct tape.
In Step 4, we do or do not do the proposed activity, based on the RMS so far. (It's important to note: if you're running the RMS correctly, lots of times you'll determine that an activity is just way too dangerous to do well!)
In the case of knife throwing, ricochets were controlled, there was no danger of passerby wandering into the danger zone, and we established a safe, controlled knife-handling procedure. So, we threw knives at the target!
Safe, fun, done well.