When executives tell me that they learned the most in their careers from mistakes, I always wonder why they don’t make a few more.

On the one hand, we all hate mistakes. On the other hand, they are often the source of deep personal learning, breakthrough ideas in business, and game-changing advances. In my book, Brilliant Mistakes: Finding Success on the Far Side of Failure, I suggest that we should actually make more mistakes (of a certain kind) given how valuable they often are. The more your business strategy is one of innovation and challenging conventional wisdom, the more you should view and utilize mistakes as portals of discovery.

Mistakes are like financial investments: They entail inherent risks while offering the potential payoff of new insights and opportunities. Just as with financial options, it’s wise to establish a portfolio of potential failures, so that a few may lead to brilliant insights.

Eight years ago, Congress passed a mandate requiring that at least a third of the Defense Advanced Research Projects Agency's military ground vehicles be unmanned by 2015. Ordinarily, Darpa would seek out the most experienced experts and companies to help design and develop prototypes. This time they made the “mistake” of turning to amateurs. They invited inventors and entrepreneurial engineers of all stripes to come up with designs for unmanned vehicles. Their products would be tested during a race across the desert in Nevada. The winner would receive a $1 million prize.

Darpa was well aware that their process would result in many failures. The first time around, in 2004, the outcomes were indeed dismal. The competitors barely made it off the starting line. It would have been easy to drop the entire project and turn to acknowledged experts, but Darpa pressed on. By explicitly encouraging a high failure rate through the “mistake” of using amateurs, Darpa was able to build a portfolio that contained many mistakes and a few successes. They decided to repeat the contest the following year, and even increased the prize money to $2 million—essentially, in gambler’s terms, going “double or nothing” on their original idea.

In the 2005 race, five vehicles finished the 142-mile course. A modified Volkswagen Touareg created by a team from Stanford University took the prize. This was real progress. It put Darpa on a path to meet Congress’ mandate. The lesson is clear: A portfolio of mistakes, if wisely managed and understood as being interdependent, offers greater opportunity for a positive payoff than a mistake made in isolation.

Nearly a century ago, biologist Max Delbruck recognized this exact notion and advocated “the principle of limited sloppiness.” He advised his students to be sloppy enough in their lab experiments to allow for the unexpected, but not so sloppy that they could not identify the reasons for their anomalous results. The trick is to make “wise mistakes,” ones that produce the most learning and manageable pain. If you play it very safe, you will make few errors and learn little.

Some key takeaways to help you navigate your own brilliant mistakes are:

Portfolio Approach

Since deliberate mistakes are low-probability, high-payoff bets, a portfolio approach can help improve the overall risk-return profile of this counterintuitive strategy.

Counterbalancing

Deliberate mistakes can serve as a hedge against received wisdom. Their negative correlation with conventional wisdom would be reason alone to invest some resources away from the crowd.

Testing Beyond

The deeper argument for deliberate mistakes is the wisdom of testing beyond the boundaries of your deepest assumptions.