Daily Blog

It’s Alive! The NYMEX flash crash and Algorithms Gone Berserk

Just after 2pm eastern time on Monday, the CME/NYMEX (GLOBEX) markets for U.S. crude, gasoline and heating oil futures shut down.  Data feeds to trading terminals ceased. The timing was bad.  It was a half-hour before the close.  Brokers that have been jacked into their terminals for six years (since the launch of “parallel electronic trading”) rushed to the usually comatose oil-futures pit to execute orders that couldn't be completed electronically.   Humans traded with humans. (“It’s Alive!”)  Thinking about how this must have played out, I just can’t get the final scene of Surrogates, the 2009 Bruce Willis movie out of my head.  All those traders in their pajamas walking dazed into the pit.

But seriously, how did this happen?  Is it important? 

Here are the facts as reported by Reuters: “CME’s GLOBEX platform froze at 2:04 p.m. EST.  Trading resumed at 3:15 p.m., after the exchange had cancelled all daily orders. The CME calculated settlement prices based on deals from the NYMEX trading floor, which has continued to operate despite the fact that well over 95 percent of all trades are now done electronically.  The CME said "technical issues" had caused the outage, and a spokesman was unable to provide further details. The outage affected crude oil, heating oil and RBOB gasoline, but not natural gas”.

That explanation leaves a lot unsaid, like what or who caused the problem.  The Zero Hedge blog gets to these answers, but you’ll have to deal with some techie jargon: “On February 13, 2012, starting 13:59:57, quotes for crude oil began queuing. At 14:00:35, all of the queued quotes were sent at once. Again at 14:01:08 the same 38 second block of quotes sent earlier was sent again -- old timestamps and all plus a few new quotes. Again at 14:01:18, all quotes since 13:59:57 were sent again. This repeated 12 times.  From a programmer’s perspective, it looks like a system problem caused a blast of quotes that corrupted a memory queue causing the software to believe the queue was full all the time”.  The result was chaos for automated systems that trade the price difference between oil futures and exchange-traded funds (ETFs) like the United States Oil Fund.  It took 75 minutes for GLOBEX to come back on line.

Got that?  In English, that means that GLOBEX’s trading software went berserk.  Two years after the May 2010 Flash Crash, another trading system went wacky –this time in the energy markets.  Unlike the 1,000 point swings in 2010, the pricing implications were minimal.  (Some blamed Monday’s $0.50/Bbl increase in crude prices on the glitch). But the implications are just as significant.  At this point we don’t know if the problem was totally internal to CME, or if the problem was caused by an errant computer trading system.  But we do know that a software glitch crashed the largest, most actively traded energy markets in the world. 

Since there was no major market catastrophe, there were no front page headlines.    But that doesn’t mean that this is not a bellwether event.  This crash happened when the markets were relatively calm and quiet.  The situation might have played out much differently if the market was in turmoil. 

The moral to this story?  Pay close attention to the magnitude and frequency of electronic market glitches.  One of these days the consequences might not be so benign.

And one more thing.  Perhaps NYMEX should keep those old physical trading pits around.  When electronics fail, there is certainly something to be said to falling back on humans.