It’s been a bad few months for British-owned yachts. There’s Ineos Britannia’s loss to Team New Zealand in the America’s Cup. That was disappointing but, let’s face it, predictable. But something no one saw coming was the sinking of the Bayesian off the north coast of Sicily in August.
It was owned by Mike Lynch, a tech billionaire I met almost 25 years ago, not long after he’d set up a software company called Autonomy. He was a charmingly nerdy kind of guy, bald and grinning, with no pretensions to coolness.
He told me back then he liked to breed koi carp and work on his model railway. From a humble background, he made his way to Cambridge University to read natural science, and then went into business after a chance meeting with a financial backer in a pub.
What really caught his imagination was Bayesian statistics, which is a system of evaluating uncertainty that combines shifting beliefs and mathematical probability.
Lynch beamed when he spoke of Thomas Bayes, an 18th-century Presbyterian minister whose groundbreaking theorem was published posthumously and wasn’t fully grasped for a further two centuries after his death. Lynch loved Bayes (hence the name of his yacht), and the statistical model he introduced inspired much of Autonomy’s software. So successful was the company that it was sold in 2011 to Hewlett-Packard for US$11.7 billion. And that’s when the trouble started. The American company had a bad case of buyer’s remorse and, after it took a financial hit, decided a fraud had been committed.
For the next decade and more, Lynch was involved in a number of legal suits and was the subject of criminal charges in the US and extradition proceedings. Lynch fought and lost several battles, including the extradition – the treaty between the UK and US is heavily weighted in the US’s favour.
I kept meaning to catch up with him, but I thought I’d wait until it was all over, when he’d be able to speak more freely.
Eventually, he was put on trial in San Francisco, where he faced 20 years in prison. That was a very serious prospect when you consider that only 0.4% of defendants in US federal criminal cases are acquitted. I was relieved when in June he was found not guilty because the case had always seemed like sour grapes to me. Lynch’s partner, Stephen Chamberlain – who’d been vice president of finance at Autonomy – also walked free.
Two months later, Chamberlain was dead, after being hit by a car while out running in Cambridgeshire. Two days after that accident, Lynch and his daughter drowned when the Bayesian sank in a freak summer storm – he had taken family and friends on a holiday to thank them for their support.
No amount of Bayesian statistical analysis would have led anyone to predict that Chamberlain and Lynch would escape prison only to die weeks later in separate accidents within two days of each other. Probability, I remember Lynch saying, is a concept that humans struggle with. We overestimate certain threats and underestimate others on largely irrational grounds.
If the odds of walking out of that San Francisco courthouse were very slim, the odds of drowning in an “unsinkable” yacht were vanishingly small. Lynch’s great wealth doubtless helped in the first instance, but perhaps it worked against him in the second. The Bayesian had one of the world’s tallest masts, a conspicuous symbol of success that may have helped drag the boat down.
It feels like some old Greek myth updated to our tech-driven times, except Lynch never acted like a demigod. He was always more a model railway guy than superyacht big shot. At least that’s how I’ll remember him.
