How Facebook's IPO Got Hijacked by Computers

Facebook's IPO taught us one thing: In the financial war against machines, we don't stand a chance.

Friday, May 18th, 11:49 am. For the first time in recent memory, millions of regular people were paying attention to the moment-to-moment fluctuations of a single stock. Most of us had no idea what was really going on — not working in finance, we rarely do — but it didn't seem like Facebook's IPO was going very well. The stock had almost instantly dropped four points from its opening price, and bottomed out at the offering price: $38. Five minutes later, the stock shot back up. It flew around like a stunned bee for the rest of the day, and ended down.

But back to those five minutes. Between 11:49 and 11:54, something extraordinary happened. For about 300 seconds, the computers took over. The stock, which had dropped four points in the five minutes prior, froze in an incredibly narrow five-cent range while two sets of computers put in thousands upon thousands of bids against one another. On one side, the underwriters' computers were offering to buy hundreds of millions of dollars worth of stock to keep it from dipping below the crucial $38 level; on the other, high frequency traders were making veerrryyy slightly higher bids at just above $38 — $38.01, $38.02 — which they would sell, literally seconds later. To get a sense of the staggering volume and speed at which this happened, watch this:

View this video on YouTube

For a few minutes, the most-watched stock in the world behaved like a malfunctioning computer program. The stock that convinced untold thousands of regular people with E-Trade accounts to get back into investing behaved according to rules that literally none of them understood, traded at volumes that none of them could conceive of and effectively followed contradictory orders from two sets of screaming robots. This is what future shock feels like.

Next time your dad asks you if he should invest in a hot tech IPO (because "you know all about this kind of stuff"), show him that video. Ask him how he'd feel to see his purchase flying down that ledger. Confident? Smart? Tiny and stupid? Then tell him, for god's sake, to find a new hobby.


"No human being can comprehend that action in the tape," says its creator, Dennis Dick, a proprietary trader and financial consultant. Dick has gotten used to seeing these bizarre glitches in the market — he predicted the legendary "flash crash" four months before it happened — so much so that he's given some of them names. In the case of Facebook, the stock fell into what Dick calls a High Frequency Tractor Beam. "You've got the underwriters [the huge banks that helped Facebook offer its shares to the public] bidding at $38, and you have the high frequency traders who are trying to get in front of them, and bring it up to $38.03, $38.04," he explains, "and that's how it gets locked in."

"The underwriters are holding it up, and now the high-frequency trading guys are trying to scalp their few cents. They're playing opposite." The result: seemingly inexplicable deadlock.

Knowing the precise details of this five-minute period isn't going to help you become a better day trader, or somehow give you an upper hand. There is no upper hand to be had. Finance computers don't have hands. "High frequency traders do 70 percent of the volume in the market now," says Dick, which means about two out of three market participants aren't even conscious beings. "With that kind of volume it really is humans against the machines to a certain extent," he says. It's worth noting that this was already the second example of bizarre systems behavior that morning: Problems with the NASDAQ's IPO mechanism caused Facebook's debut to be delayed.

The fundamental weirdness of high-frequency trading is well-documented, but tends to feel foreign. Stories about HFT read like dispatches from the elite world of high finance, their meaning abstract and their consequences hard to grasp. But this is Facebook. This is a company whose product everyone uses and which was preordained to be the poster boy of this financial cycle, whatever it is. Facebook's IPO felt dangerously knowable to a regular guy with a few tens of thousands of dollars in retirement funds to throw around, when, in fact, its behavior was utterly unknowable. "It was predictable from my perspective that something like that could happen," says Dick, "but I didn't think they could hold a stock like Facebook down. It just shows you the power that the high frequency traders really have."

This puts a new spin on the old finance adage: "Don't play the loser's game." Except it's not just that the pros will always be better than you at gaming the market on a short term basis; it's that you'll never be faster than a computer, either.

It's too early to know what Facebook's stock is really going to do — the first few weeks really don't count. You can guess that the company will do well and make lots of money, which is a pretty good reason to eventually invest. You can divine that the site has no real business plan beyond "ads ads ads" and decide that's a good reason not to buy.

But try to play the "game" and you'll end up broke. Facebook's stock is, at the time of writing, still whipping around in the wind, currently listing at $34, partly because its underwriters are no longer propping it up. Which, maybe, is for the best: The more casual investors get scalded by Facebook, the fewer will eventually get crushed to death by the gears of the new financial machine.

Update to more accurately reflect the roles of the underwriters in the HF Tractor beam situation — thanks, @nickrizzo.

Skip to footer