Skip to main content

This story features in Unearthed: The Age Of Data

It’s worth remembering that whatever errors and false conclusions inadequate data might have led you to in the past, there’s always someone who’s managed worse.

A textbook mistake?
In September 1999, the Mars Climate Orbiter entered the planet’s atmosphere full of potential and scientific endeavor – and plunged to a fiery and $125 million dollar doom. The reason? Almost unbelievably, the power the thrusters needed to deliver for a successful maneuver had been calculated in pounds of force, but a separate piece of software had taken the data in as Newtons. Basically, they forgot to convert to metric. NASA engineer Roger Cook later said, “everyone was amazed we didn’t catch it.” Which is perhaps putting it mildly.

The Mars Climate Orbiter

You never know what old data could be worth
Pity James Howell from Newport, England. In 2009 he obtained 7,500 Bitcoins for virtually nothing. He stored the data on a hard drive, popped it in a drawer and forgot about it. By 2013 they were worth £4.6m (around $7.5m at the time). However, he’d thrown the hard drive away. He even went to the landfill site in South Wales where the drive had ended up, but confessed to the BBC “It was about the size of a football field. I took one look and thought ‘no chance’.”

$23million dollars for a single book
In 2011, Peter Lawrence’s biology textbook “The Making of a Fly” appeared on Amazon on sale for the modest sum of $23,698,655.93. (Plus shipping, of course.) Two automated bidding programmes from competing sellers had run constantly and uncorrected for 10 days, each nudging up the price. Amazon reviewers demonstrated their usual talent for a put down with one reporting: “I was fortunate enough to buy this at the bargain price of $19,087,354. Must have been a sale on. The next day it was $23m!”

“The Making of a Fly” by Peter Lawrence

Accurate entry is everything
One Thursday, late in 2005, a Mizuho Securities trader on the Tokyo Stock Exchange set about selling 610,000 shares in recruitment firm J-Com Co at 1 yen each. Unfortunately, the data had been typed in incorrectly. What he’d wanted to do was sell every 1 share at 610,000 yen each… By the end of the day, Mizuho Securities had lost $225 million.

Better poor than perished
In 2016 a data glitch informed some senior citizens in Michigan that their benefits would be cut. 12,000 of them. They each received a letter erroneously announcing that their income would be reduced. The Michigan Department of Health & Human Services had been trying to bring its database in line with federal statutes and one small error had triggered an avalanche of false results. Unpleasant, but perhaps not as bad as when the St Mary’s Medical Centre (also in Michigan) mistakenly informed 8,500 healthy patients and their insurance companies that they were, well, dead. They did at least refer to the demises as ‘unfortunate’ when they sent the notifications out.

There should definitely be a Stanislav Petrov Day
On September 26, 1983, Petrov was on duty at a Russian command centre when its computer told him (incorrectly) that America had launched a missile attack on his country. It was Petrov’s responsibility to inform his superiors and so initiate a counter-attack. The fact that we are all still here and there wasn’t a World War lll is down to Petrov having “a funny feeling in my gut” (seriously) that it was a false alarm, and disobeying his orders. He was of course right. Later in the 1990s, when the incident came to light, Petrov largely played it down saying that the famous red button launcher that would unleash nuclear Armageddon had never worked properly anyway.

And finally
It’s 1979. You are Ross Perot and you own Perot’s Electronic Data Systems. You’re looking to invest in a small computer company up for sale. You study the market, you analyse the data, but decide the $50m the firm is asking is too much. So you turn down Bill Gates and Microsoft.