The folks at Nanex have done another very interesting forensic analysis looking at the volume on either side of the flash crash, i.e., between 14:43 and 14:45 when the most vicious part of the selling took place, and on the rebound, between 14:46 and 14:49, when the bounce to unchanged took the market right back up. Not at all surprisingly, there is a huge mismatch, at least on the SPY (which, however, being the most traded security in the market, is a pretty good representation of overall volume trends): selling volume is orders of magnitude, and far more concentrated than the bid side on the bounce. This leads Nanex to conclude that "Basically, when the shelling stopped, there was no one left standing with good pricing information -- and when the shorts went to cover and buy back stock they found prices rocketing skyward with almost no effort." In other words, if anyone wanted to created a short covering squeeze by pulling all the borrow (wink wink State Street and BoNY) they could have immediately undone all the damage associated with the 1000 point drop in the Dow. Incidentally, this is precisely what set off the market on its steady bear market rally back in March of 2009 - a concerted effort, orchestrated by the State Streets of the world, to pull every single financial stock's borrow, creating the biggest short covering spree in the history of our broken stock market.
More from Nanex:
This is a 1-second chart of the Spyder ETF (symbol SPY) showing the sell-off and recovery on May 6, 2010. We included the tick volume (the count of the number of trades for each second) plotted at the bottom. What this chart shows is that the recovery occurred not due to strong demand, but rather because there were few sellers left. If it were due to strong demand, the volume would have been much much higher. This is a simple matter of supply and demand.
In other words, nobody wanted to buy stocks at the bottom: and the rebound had nothing to do with HFT intervention as the now luckily retired Jim Simons claims, but merely a push to incite yet another flare up in short covering. What would be interesting is charting the VWAP on the SPY pre crash and post crash - we are confident the differential would be well over 300 Dow points, once again confirming that supply and demand no longer matter in stocks whose prices can plunge at the whim of a few SPARC cores.
Below is the visual representation of the move: