The pound has been falling sharply for several days, but for eight minutes in the early hours of the morning of Friday (7 October) the falls were something else, but then they largely reversed. What happened? And why did it happen?

Sterling has fallen from $1.30 at the end of September to just 1.2434 at the time of writing (7.37 BST 9 October). But for a few minutes in the wee hours of the previous Friday, the falls appeared to be the stuff of panic.

It began at 7.07 Hong Kong time, sterling dropped from $1.26 to a $1.1491 in just eight minutes, before recovering – although the pound still finished the day at a price that was lower than the start of day level.

So why the crash?

Part of the explanation lies in the time of the day when it happened. At that time of the morning, when traders in Europe and New York are asleep, or at least not working, but even traders in the East are still wiping sleep from their eyes, volume is low – tiny even.

When volume is low, the odds of one-off events distorting the markets grow.

One theory is that it all began with a so-called fat finger, human error exercising a sale order on a scale that was not intended. At that point algorithms, the so-called algos, stepped in and computer programs designed to sell under certain conditions sold. The programs monitor the activities of other programs, and what followed was a positive feedback loop – a phenomenon so named after what happens when sound is amplified via a microphone but creates that awful screeching noise.

Another theory is that the selling was kicked off after certain programs designed to sell, or indeed buy, reacted when the FT published a story about French President Francoise Hollande wanting the EU to take a very tough negotiating stance against the UK. Whatever the cause, most seem to agree that at a certain point, the algos took over, and the selling became a frenzy, a mad eight minutes of selling until the pound became so cheap that value investors, or maybe algorithms designed to buy when something looks underpriced, took the initiative.

But the currency market is huge – worth many multiples of trillions of dollars – how could it become so distorted?

Some blame the algorithms themselves, programs that facilitate high-frequency trading; when sell orders can be executed in tiny fractions of seconds – when things happen that fast, common sense does not get a look in.

But such criticism misses the point, and it misses the point for three reasons.

First off, the last thing the creators of the algorithms want is for them to distort the markets, and over-sell, as that is just as likely to creates losses as profits for the companies behind the trading. Over time, such algorithms will become more sophisticated, and the AI that is programmed in, may learn how to see the bigger picture and factor in the impact of algorithms at times of extreme volatility.

Second off, algorithms are more likely to have a significant effect on price when trading is low. In this respect, regulators, who have cracked down on proprietary trading by the banks – when they speculate with their own money – may be part of the problem. Banks have been leaving the world of currency trading, which in turn has cut back on liquidity, and increased the odds of market distortions occurring.

On this theme, Ulrich Leuchtmann, head of currency strategy at Commerzbank AG in Frankfurt, told Bloomberg that: “This is not something you would expect in a half-efficient market. . . We have a liquidity situation which has eroded massively over the last few years and policy makers have largely ignored it. All the regulation that we have in place, for good reason, has the side-effect that liquidity in the FX market is much more-shaky and fluctuating heavily, and we have times when it’s extremely low, especially in Asian trading.”

There is a third explanation, maybe the markets were exercising some prescience. Maybe, they were simply selling sterling down to a level that really is the appropriate level. Watch what happens to sterling over the next few weeks. If the pound finishes this year at a price in the region of $1.15, then instead we can simply say that the flash flood was a sign of foresight.