The truth is...IMO...that yesterday, in between microsecond print points, the system came very close to Dow = 0.0 yesterday. You didn't see that, instead you saw a few stocks as low as $0.0001 per share, SOLELY because the automated buys kicked in about the same time as Greek police kicked Greek protestor ass. Whether a few more microseconds would have produced Dow = 0.0 or not, the point is that yesterday proved it is possibly...mathematically possible. Disprove what I say if you can, but ignore it at your own peril.
Speaking as a Math/Physics/CS grad (not MIT). I can't see how the algorithms 'failed' yesterday? The spike down was a result of no one (including the algorithms) wanting to offer tight bands / liquidity with that type of volatility / fear. Causing all the stops, margin calls and band wagon shorts to nuke through the thin levels. Thats the risk of leveraged trading and using wide stops. Its not the job of HFT groups to provide liquidity at all times (they are allowed to shut off if they feel its in their interest). Thats the market maker's job, and even they are allowed to widen spreads so they don't go bust. To me, the market operated wonderfully. Everything oscillates around fair value, somedays more than others, the more advanced the algorithms get, the closer we should stick to fair value (since money can be extracted from volatility).
Really? Google the concept of fair value. Company A with X assets and Y revenue (assuming X > 0.01$*shares outstanding). Will never be worth 0 (or 0.01$*shares outstanding). Even if retards are willing to sell it the stock at 0.01$ for a short period of time. Eventually (or rather quickly) people with functioning brains will buy it back up closer to what its actually worth (until the purchase no longer projects a reasonable return).
I would disagree here, but if anyone was looking for a reason to sell, they got it, yesterday. It's too early to tell what is really bothering the market if anything. I happen to agree with CNBC's analysis that a dumbass trader added two zeroes and actually had that large of a position in his account to sell.
Math, MIT, computers, economists are just stupid nerds. They know nothing about the real world. They don't even know how to get laid. They believe their moronic math models are useful... ...useful for generating trading commissions.
True.However, I think his point was that mathematically, it IS possible for the NYSE computers to be buggered beyond all hope. Sabotage can't be ruled out.
More naysaying. The only true way to guaranteeably (new tradism to add to our knowledge base) win is to become the house so that you're indifferent to the financial outcome. Too hard? How about moronic math models that perfectly price options in real time? Stupid to think that they should want to calculate the price of an option so that they can hedge for "pure gravy" of arbing the market so that they make "reasonable profits" just to receive a commission as you say. Unfortunately, a lot more complicated than you give it credit for. These quants and their models...of course...what stupid fools....to think that they need a price before they even make a trade or post a bid or ask. Yes....what ignorant, uninformed traders. We should never give algos any credit b/c they don't work for me so I'll just tell people to give it up. Then, after I quit criticizing traders, I'll decide to take my money out of my FDIC CD.
I just read an interesting interpretation of Thursday's sell-off, which I think makes some sense. Because the market had risen so consistently throughout the past year, there were a lot of trailing stop orders that had built up into the order book. You were looking at incredibly consistent rising indexes without significant correction for months. Without those normal corrections, there was simply a build up of Sell Stop orders that were eventually triggered en masse when the market started shooting lower. I'm sure this is one factor that contributed to the rapid sell-off.