This is what panic looks like:
The chart (click on it to enlarge) traces the price of SPY from 2:32 p.m. Eastern to 3 p.m.
The big dip began to pick up speed around 2:42 p.m., reaching its low at 2:45 p.m., before reversing quickly back up to the level where it began. The 133-tick chart traces a new bar every 133 price movements, so it's a quite detailed level of resolution.
I'm seeing market news stories about stocks plunging more than 1,000 points before partially recovering, with no context regarding the duration of the move.
It was a panic event. It was over in three minutes.
The AP story is the self-serving nonsense from reporters that backs a really dramatic lede that isn't justified by the events.
AP wrote that a technical problem hastened the selling, but then in the next sentence took it back with a statement that even so, emotions were running high in response to the Greek debt crisis.
Implication: It was the Greeks, not the geeks that caused the move.
The news agency needs to learn about time zones. At 2:42 p.m., when the event began, it was 9:42 p.m. in Athens. People were getting ready for bed.
It was 8:42 p.m. in Frankfurt, where the European Central Bank is headquartered, and Berlin, capital of Germany, a major player in the Greek bailout.
Europe wasn't producing news at 2:42 p.m. New York time.
Reuters has just moved a story saying the drop was caused by an erroneous trade entry -- a typo, if you will.
Once a false price like that hits the trading systems, the computers will kick in, and a full fledged panic will ensue, although mainly driven by machines, not humans.
Bravo, Skynet. We are not our own masters.
Where is John Connor when we need him?
The point to take away is, the move wasn't significant. It was an entry error compounded by hair-trigger trading assumptions built into software. It points to problems that need fixing, but it's not significant.
Update: AP later altered their story to include the erroneous trade hypotheses and brought more balance to the analysis.
No comments:
Post a Comment