So says Richard Pike, a former oil-industry adviser and chief executive of the UK Royal Society of Chemistry, who blames flawed statistical calculations.

Oil companies produce a bell-shaped probability distribution for how much each oil reservoir might hold, and then quote as an indicator of the reservoir's capacity a figure they are 90 per cent certain they can exceed. When publishing a result for multiple reservoirs, they simply add up the figures for each one. And this is where the problem lies.

"They should be combining the bell curves for each reservoir," says Pike. Adding the numbers for each reservoir ignores statistical information about the extremes of the distribution, giving a result which underestimates the true total figure for all the reservoirs.

According to published estimates, there are 1200 billion barrels still to be extracted, but Pike says there could in fact be twice as much. "The figures are almost meaningless and just provide a conservative estimate for shareholders."

Pike claims that most oil companies do calculate statistically accurate estimates of the combined capacity of their oil reserves, but no one can access this information to work out how much oil there really is in total. "All companies keep their internal probabilistic estimates quiet," he says.

This story contains 255 words.

If you are a paid subscriber, check to make sure you have logged in. Otherwise our system cannot recognize you as having full free access to our site.

If you are a paid print subscriber and haven't yet set up an online account, click here to get your online account activated.