Failure has always been a fundamental part of a market economy. When markets work, they do so because new ideas are constantly being tried out. Most fail. Those that succeed cause older ideas to fail instead. In the US, about 10 per cent of businesses disappears each year. This is an awkward insight – but trial and error could be starting to take its rightful place as a business technique, rather than the dirty little secret of capitalism.

There are some hopeful signs. Stefan Thomke of Harvard Business School has argued that advances in computation have made it possible to experiment on new products as a matter of course, trying many things and expecting many failures. It is now easy, for example, to experiment with changes in the layout of a website, showing different configurations to different visitors and tracking results in real time. Google, meanwhile, routinely launches new products with a “beta” label on them. And academic superstars such as Steven Levitt, co-author of Freakonomics, have been teaching executive courses in business experimentation.

We are also starting to learn more about the psychology of learning from mistakes. Richard Thaler, the behavioural economist behind Nudge, coined the phrase “hedonic editing” to describe our habit of lumping small losses together with larger gains in order to mask the pain of the loss. Sugar-coating is human, but it's also a recipe for failing to learn from failure. Thaler, with colleagues, even studied the behaviour of contestants on Deal or No Deal. He discovered that people who had made unlucky choices then started to take reckless risks, which often compounded the error.