This page has been archived and commenting is disabled.

Flash Crash Mystery Solved

Tyler Durden's picture





 

Via Nanex

Flash Crash Mystery Solved

Below are portions of a comment letter submitted by R.T. Leuchtkafer to the SEC on April 16, 2010, just 3 weeks before the flash crash. The second paragraph in the excerpt below, unknowingly describes exactly how the flash crash was started. The letter goes on to alert the SEC on the dangers of High Frequency Trading (HFT), phantom liquidity and other concerns.

Our exhaustive analysis of the May 6, 2010 Flash Crash included:

  1. processing trillions of exchange quote, trade and order book records for stocks, options, and futures on May 6th and other trading days back to 2005 and days since.
  2. matching up the actual 6,438 trades (75,000 contracts) executed by Barclays for Waddell & Reed to the 147,577 trades executed in the CME June 2010 eMini during that time.
  3. drawing from extensive interviews with authors of academic and regulatory papers on the Flash Crash.
  4. conversations with dozens of traders, programmers, investigative jouralists, government officials in the U.S. and abroad, anonymous callers, and "people familiar with the matter" since June 2010.
  5. finding and studying similar crashes in the eMini and crude oil and thousands of other anomalies.

From this, we were able to zero in on the ignition point, or starting time of the crash: 14:42:44. That is the moment when one or more large HFT "market makers" hit their limit of long positions in the eMini Futures (ES.M10), and reversed out - "readjusted their position". Immediately. That aggressive act sucked out a significant amount of liquidity and caused thousands of trading instruments (stocks, options, indexes, futures) to reprice, which severely overloaded all trading systems processing market data (peak message traffic set a record at that time, which was not exceeded for the balance of the day). Overloaded systems caused bad, delayed, and unexpected pricing to appear, which caused other algos and traders to stop trading, removing any remaining liquidity. We know the Waddell & Reed algo practically ceased trading shortly before the 600 point slide in the Dow Jones Industrial Average; selling a mere 1000 contracts in small lots (averaging 6 contract per trade), all on the offer side. No Virginia, the Barclay's algo used by Waddell & Reed did not sell indiscriminately without regard to time or price: it didn't take liquidity either. That was the work of HFT. In short:

High Frequency Trading caused the Flash Crash. Of this, we are sure.

If the above isn't enough damning evidence, in the paper Moore’s Law vs. Murphy’s Law dated March 19, 2013 co-authored by Andrei Kirilenko, the former CFTC chief economist and principal author of the final SEC flash crash report (page 11), finally agrees:

...After buying the E-mini for about 10 minutes, high frequency traders reached their critical inventory levels and began to quickly and aggressively unwind their long inventory at a key moment when liquidity was sparse, adding to the downward pressure. High frequency traders rapidly passed contracts back and forth, contributing to the “hot potato” effect that drove up trading volume, exacerbating the situation. Meanwhile, cross-market arbitrage trading algorithms rapidly propagated price declines in the E-mini futures market to the markets for stock index exchange-traded funds like the Standard & Poor’s Depository Receipts S&P 500, individual stocks, and listed stock options. According to the interviews conducted by the SEC staff, cross-market arbitrage firms “purchased the E-Mini and contemporaneously sold Standard & Poor’s Depository Receipts S&P 500, baskets of individual securities, or other equity index products.” As a result, a liquidity event in the futures market triggered by an automated selling program cascaded into a systemic event for the entire U.S financial market system. As the periods during which short-term liquidity providers are willing to hold risky inventory shrink to minutes if not seconds, Flash-Crash-type events—extreme short-term volatility combined with a rapid spike in trading volume—can easily be generated by algorithmic trading strategies seeking to quickly exploit temporarily favorable market conditions.

That key moment in time was 14:42:44.

Finally, we can close the books on investigating the cause of the flash crash.

Note to journalists.

The HFT lobby will vehemently deny any blame for causing the flash crash and will use a number of straw man arguments, eventually enlisting the SEC final flash crash report which named Waddell & Reed as the cause (W&R). Here are some key points to remember:

  1. Always ask for data used to back up any claim. You will find a lot of people are long on talk, woefully short on data. Data is difficult to obtain and takes a lot of expertise to work with: use that to separate the wheat from the chaff.
  2. The SEC publicly stated numerous times that it took them 5 months just to receive and organize the vast amount of market data. 5 months from May 6th is October 6th, yet they published their report on October 1st. When did they have any time for analysis?
  3. W&R used Barclays to execute the algo (a fact not mentioned in any SEC report).
  4. The regulators (CFTC) first interviewed the Barclay team that executed the W&R trades on October 14, 2010 - yet published the final flash crash report 2 weeks earlier, on October 1, 2010. Why this has never been mentioned in the media is a mystery.
  5. Barclay's algo only sold at the offer, never crossing the bid-ask spread during the crash: it was entirely passive, and did not move the order price lower once entered. This type of passive algorithm cannot cause a market crash on its own.
  6. Barclay's algo only sold about 1000 contracts during the entire 600 point slide in the DJIA. It did not accelerate selling as the market began to drop. To claim otherwise is simply untrue.
  7. Barclay's algo sold heavily after the eMini circuit breaker ended and the market had bottom: a period of maximum chaos, severely delayed feeds and minimum liquidity. Yet market prices climbed.
  8. The SEC final flash crash report incorrectly uses the word liquidity 249 times in 89 pages. It is worthwhile reading about this monumental logic error.
  9. Barclay's and W&R gave the actual trade executions to Nanex for analysis and we were able to match up those trades to the CME time and sales so we have a true and accurate picture of how the algo executed. We know exactly which trades came from W&R and an exact picture of how the market looked when each one executed.

This is a very complex subject and lobbyists will use that to bamboozle you. Feel free to ask questions at the email below. Also, be sure to watch this award winning documentary (Money & Speed) which captures Eric Hunsader of Nanex on film formulating what turned out to be, the actual cause of the flash crash.

 


- advertisements -

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.
Wed, 03/27/2013 - 20:01 | Link to Comment Pairadimes
Pairadimes's picture

Who the hell else is in the market?

Wed, 03/27/2013 - 20:13 | Link to Comment EscapeKey
EscapeKey's picture

sitting ducks

Wed, 03/27/2013 - 20:51 | Link to Comment notbot
notbot's picture

A+ stuff, Nanex.  Great work. 

Wed, 03/27/2013 - 21:13 | Link to Comment aka Gil
aka Gil's picture

It doesn't matter how nicely you present a gift to the SEC, they don't give a fuck because they don't work for us. Just ask Harry Markopolos if you don't believe me.

Wed, 03/27/2013 - 20:15 | Link to Comment Groundhog Day
Groundhog Day's picture

Don't worry no one will go to jail or be indicted for the billions lost.  Just another day in the neighborhood at the SEC.  They are now just  a wholly owned susidiary of JPM and GS

Thu, 03/28/2013 - 00:04 | Link to Comment putaipan
putaipan's picture

hutzpah bitchiez!

Wed, 03/27/2013 - 20:02 | Link to Comment Zero_Sum
Zero_Sum's picture

Of course, 'Leuchtkafer' is a pseudonym. Still curious who actually wrote that letter.

Wed, 03/27/2013 - 23:46 | Link to Comment AlaricBalth
AlaricBalth's picture

Leuchtkafer is German for firefly. The personification of a small light in the darkness, maybe.

This link takes you to his (or her) Bibliography of HFT Studies.
http://blog.themistrading.com/wp-content/uploads/2011/11/HFT-Bibliograph...

"The bibliography attached to this note includes nearly a century’s worth of evidence-based research discussing the effects of market maker scalper business models on markets around the world. Several papers describe how scalpers are not just simple passive market makers, as sometimes imagined, but trade aggressively on information they glean from their many privileges.

Several papers also describe how unregulated or unconstrained market maker firms manage inventory, and how, in volatile markets, these firms can exacerbate price volatility by trading as quickly and aggressively as they can, just as happened in the Flash Crash. The bibliography also includes evidence- based research papers that explore the benefits of market maker regulations and obligations in stock markets that have tried them, including, somewhat ironically, the U.S. itself before deregulation.

Finally, several papers discuss the effects of high frequency trading of all types on today’s markets."

Wed, 03/27/2013 - 22:51 | Link to Comment fuu
fuu's picture

Leuchtkafer = Hacker Flute and Leather Fuck

Wed, 03/27/2013 - 20:09 | Link to Comment Bear
Bear's picture

Now can I sue for the $11,000 that I lost that day? Or does the explanation conveniently come after the statue of limitations expires?

Wed, 03/27/2013 - 20:47 | Link to Comment sitenine
sitenine's picture

Fuck NO, you can't sue! In fact, you should slap yourself silly for having money in this Fed sponsored and SEC sanctioned casino in the first place.

Thu, 03/28/2013 - 00:01 | Link to Comment Pure Evil
Pure Evil's picture

Sure, you can sue, but expect to lose another $11k (or more) to the lawyer.

Wed, 03/27/2013 - 20:10 | Link to Comment Schmuck Raker
Schmuck Raker's picture

SEC response: Move along...nothing to see here...return to your dwellings.....we have assumed control

Wed, 03/27/2013 - 20:50 | Link to Comment thatthingcanfly
thatthingcanfly's picture

Kim Kardashian was spotted wearing a baby-bump-hugging dress today.

Wed, 03/27/2013 - 21:56 | Link to Comment cifo
cifo's picture

She still goes to the gym?????

Wed, 03/27/2013 - 20:12 | Link to Comment ultimate warrior
ultimate warrior's picture

It's just what free markets do....move along

/sarc/

Wed, 03/27/2013 - 20:24 | Link to Comment Umh
Umh's picture

I think that this sentence can be applied to a number of different government regulatory areas.

"Someone once suggested the real scandal is often not what is illegal but instead what is legal."

Wed, 03/27/2013 - 20:30 | Link to Comment disabledvet
disabledvet's picture

First comes the bamboozlers...then it's full on balderdash. After that it's "dummy up and talk to legal."

Thu, 03/28/2013 - 00:29 | Link to Comment putaipan
putaipan's picture

is this a good time to mention the monsanto protection bill?

Wed, 03/27/2013 - 20:36 | Link to Comment notquantumdum
notquantumdum's picture

I locked in 5% profits (plus) in 24 hours, buying the lows in various securities during the flash crash after the bottom, and then selling them the next day.

But, I am strongly opposed to BS regulation and enforcement like this seems to indicate!

I would much rather have not made the profits, and not had the flash crash happen.

Wed, 03/27/2013 - 20:42 | Link to Comment nanex
nanex's picture

Good one!

 

Now, photo copies of brokerage statements or it didn't happen.

 

nanex

Wed, 03/27/2013 - 20:47 | Link to Comment notquantumdum
notquantumdum's picture

sorry, don't care if you believe me when I tell the truth.

I would rather stay as anon as poss.

Wed, 03/27/2013 - 21:13 | Link to Comment BackOffice Slut
BackOffice Slut's picture

you can easily stay anon. Black out name, address, and all that important stuff. I mean, if i made 11k that day i'd show that shit off!

Thu, 03/28/2013 - 00:22 | Link to Comment Pure Evil
Pure Evil's picture

You don't have to show the brokerage statements. Just explain the process of how you went about buying the lows in various securities after the flash crash in excrutiating detail.

If you like, you can use the charts from this ZH post about the flash crash to explain to us how you were able to buy at the bottom and sell the next day, especially since the bottom lasted all of about 2.5 minutes.

Now, don't be shy, and try to remember. If you can't dazzle us with your brillance, then dazzle us with your bullshit.

Thu, 03/28/2013 - 00:31 | Link to Comment notquantumdum
notquantumdum's picture

You seem to exaggerate in my recollection.  I remember having longer than 2.5 minutes to decide.

Look at where the older lows and highs occurred, farther back, it's amazing how useful that is.  Look for volume traded, that is very important.

I seem to remember hours being involved, but maybe my memory is not quite remembering well.

Thu, 03/28/2013 - 01:08 | Link to Comment Pure Evil
Pure Evil's picture

Yeah, your memory sucks like a dementia patient at a skilled nursing facility.

So, are you trying to imply that a FLASH CRASH lasts hours?

Don't try to snake us with mumbo jumbo about older highs and lows and volume traded. From Nanex's own charts the FLASH CRASH lasted all of approximately six minutes.

If you go look at a few of the charts I provided through the link, YOU, and everyone else can see the truth and the proof provided by Nanex and ZH.

Still waiting for the detailed post explaining how you executed trades from your mother's basement when computers co-located to the exchanges running HFT algo's weren't allowing any outside trades to be executed since they were widening the spreads making it impossible for people without computers co-located to the exchanges running HFT algos to get accurate bid/ask spreads. Go back and read the last paragraph from the excerpt. What that paragraph is trying to say is that the co-located computers running HFT algo's were using up all the exchanges run time so that no other computers except those co-located were able to obtain time slices with the exchange computers. Effectively blocking all other executions.

Please explain your exact process for the times between approximately 14:53:30 and 14:48:30.

If a birth certificate can be forged so can brokerage statements.

Wed, 03/27/2013 - 21:08 | Link to Comment notquantumdum
notquantumdum's picture

'And, I respect your presumed integrity for asking.

Wed, 03/27/2013 - 20:45 | Link to Comment slightlyskeptical
slightlyskeptical's picture

Same here. Have to admit that all my buys were GTC orders put in weeks earlier, that I had completely forgotten about. Turned out to be a nice day.

Wed, 03/27/2013 - 21:08 | Link to Comment fonzannoon
fonzannoon's picture

I was sitting right there when that crash happened. I went after it and tried to buy a few things. I put in market orders. Stupid move apparently. They got filled, but only on the way back up at way higher prices than I was staring at when I hit submit. Ended up being one giant non event.

Wed, 03/27/2013 - 21:13 | Link to Comment nanex
nanex's picture

I have heard this same story from many people. Fidelity wasn't able to give fills until hours after the market close. There are lots of people out there fuming. Still.

'

nanex

Wed, 03/27/2013 - 21:17 | Link to Comment fonzannoon
fonzannoon's picture

You don't know how much so many of us appreciate what you do nanex. Thank you.

Wed, 03/27/2013 - 21:29 | Link to Comment notquantumdum
notquantumdum's picture

You, are correct, I believe.

I have multiple brokers with multiple accounts, and I seem to remember them going on and off-line at different times.

It was nice having more than one broker, for sleeping at night.  But, it didn't seem to actually end up helping me much given than I never really ended up wanting to conduct a trade during the relatively somewhat short outages which I experienced -- and I wasn't intentionally short-term trading at that time, thankfully.  I think I might have been somewhat lucky.

Outages are hated, but they do happen, from time to time.

Thu, 03/28/2013 - 01:56 | Link to Comment Tapeworm
Tapeworm's picture

If you are the real NANEX, you have my eternal respect for what you have done to lay it out for the rubes like me. No kidding, what you have done should have been the talk of every Sunday morning MSM TV gab fest.

 What sickens me the most is that NANEX did the work gratis for the gombit regulator filth and they do all to bury the information. That is why I see no hope.

 Nevertheless, NANEX is tops on my Hero list. I cannot think of anyone else that has done more.

Wed, 03/27/2013 - 21:16 | Link to Comment notquantumdum
notquantumdum's picture

I HATE market orders with a capital H!

But, I will enter them sometimes, every once in a while, if and only if desperate enough, in a totally adequately liquid-security, when I can't seem to obtain an execution any other way, time and time again.

What else are they good for?

(PS, Doesn't that almost have to mean that I am daytrading, if that is my criteria?)

Wed, 03/27/2013 - 21:19 | Link to Comment notquantumdum
notquantumdum's picture

'should say, ". . . my criteria for needing to use a market order", so who else needs a market order.

Wed, 03/27/2013 - 23:04 | Link to Comment dark pools of soros
dark pools of soros's picture

market orders are fine when you are waiting to dump the yearly vested golden handcuffs your corp zombie job grants you.. like gruel you chug it as fast as you can

 

Wed, 03/27/2013 - 23:22 | Link to Comment notquantumdum
notquantumdum's picture

They're apparently short-termers, also, huh?

Maybe focusing more on the longer term would be useful?

Thu, 03/28/2013 - 00:38 | Link to Comment dark pools of soros
dark pools of soros's picture

and a day late to the next crash = ?

lock in gains and spread it around elsewhere

 

Thu, 03/28/2013 - 00:57 | Link to Comment notquantumdum
notquantumdum's picture

Spreading risks out well is apparently almost always a good thing to do.  That is exactly what I have been trying to suggest in recent posts, if not do myself, given the obvious challenges in trying to do this, even though one will not necessarily have a whole lot of certainty as to how things will play out . . . usually.

If only our "leaders" could do so well.  They seem to always maximize the opposite.

There is nothing wrong with sticking to one's goals, even if challenging, I think.

'Don't let the bums get you down, when you know you're right!

Thu, 03/28/2013 - 17:31 | Link to Comment StandardDeviant
StandardDeviant's picture

lock in gains and spread it around elsewhere

...including (compliance rules permitting) using part of those gains to buy some cheap OTM puts on the thing!

If you're the CFO, say, there's an obvious moral hazard.  But if you're just another Joe Employee, whose salary (and benefits, pension, etc.) are entirely dependent on the company, your investment portfolio should act as a hedge, not push you even further long in the same direction.  Just ask any ex-Enron employee.

Thu, 03/28/2013 - 00:12 | Link to Comment putaipan
putaipan's picture

bergoo bisches!

Wed, 03/27/2013 - 20:51 | Link to Comment notquantumdum
notquantumdum's picture

Instead of "BS" regulation and enforcement, maybe, too much bad, and not enough good, regulation and effective enforcement, might be a better way to phrase things. . . .

Wed, 03/27/2013 - 23:03 | Link to Comment notquantumdum
notquantumdum's picture

'Received a number of negative ratings, but i'm certain i was adding liquidity during these events (based on the above def) because I was only entering limit order which were outside of the current market prices, during the flash crash.

'Complain if you will, but what else is one supposed to do?

Thu, 03/28/2013 - 05:43 | Link to Comment YC2
YC2's picture

Not being a robot, my impression that day was that throwing a market order into that chaos was asking for trouble. Slip city.

Thu, 03/28/2013 - 01:08 | Link to Comment GOLDTRADERRR
GOLDTRADERRR's picture

@notquantumdum

Every crisis provides opportunities.

Be glad you profited.

These crisis opportunities will become more frequent

rather than become fewer as the system collapses.

http://goldtradercommentsaugust2010.blogspot.com/

Wed, 03/27/2013 - 20:51 | Link to Comment Sathington Willougby
Sathington Willougby's picture

Kirilenko runs a communal agenda with his "anti-profiteering" pre-disposition towards "economics".  What pseudoscience drivel is this brought by flower children?

 

Please, refrain from entering voluntary contracts if you don't have the fortitude to to take responsibility for your capital and risk exposure.  Perhaps you should invest in government bonds.

 

Two types demand zero risk returns.  Neither reach the level of human being.  One would be human stealing and the other human begging. 

Wed, 03/27/2013 - 21:02 | Link to Comment knukles
knukles's picture

Oh come on, it was more fun blaming Waddell & Reed.

Thu, 03/28/2013 - 01:29 | Link to Comment piliage
piliage's picture

Exactly. It was all the fault of those BBQ eatin' hicks in KC, who actually according to the video bought an order at the very bottom and sold at the very top, not the zillion high liquidity market moving millisecond ordes by 'unnamed large banks protected by the sec due to disclosure rules'. Spit

Wed, 03/27/2013 - 21:02 | Link to Comment IridiumRebel
IridiumRebel's picture

I don't know what any of this shit means. I just buy the fucking dip.

Wed, 03/27/2013 - 21:03 | Link to Comment Miss Expectations
Miss Expectations's picture

Bravo!

Wed, 03/27/2013 - 21:10 | Link to Comment Doode
Doode's picture

As someone with intimate knowledge of barcap participate algo I must say the above article is not very well informed (both 5 and 6 are horseshit). Particpate uses Iceberg child orders that are directed to cross if within a certain number of seconds no execution takes place. The fault is first with the WR traders and second with people watching the barcap algos executing for not interferring in time. The participation rate was way too high - anything over 50% will move the market and the one WR sent was a motherload.

Wed, 03/27/2013 - 21:28 | Link to Comment nanex
nanex's picture

Yeah? Must be a different algo then. You didn't even bother to look at the data. We did. We matched every one of those trades and every one executed at the offer. If you have evidence, real evidence, not delusions, then submit it. Otherwise, you're just full of it, trying to bamboozle everyone.

How do you have "intimate" knowledge of the barcap participate algo? 

Say hi to Vijay for me.

nanex

Wed, 03/27/2013 - 21:49 | Link to Comment Bay of Pigs
Bay of Pigs's picture

+1 

Paging Tyler Durden for a bitchslap.

Thu, 03/28/2013 - 00:21 | Link to Comment putaipan
putaipan's picture

wow. last night kruman.

tonight - the high frequency micro forensic, bill black prosicutorial preceeding, imaginary

guilloteenin'- just ends deserved dispatched diggin' algorhythmic sleuths....

i like the company i keep. and, tylers- if you wanna bitchez slap me.... i completley understand.

Thu, 03/28/2013 - 08:39 | Link to Comment Doode
Doode's picture

I know because I wrote parts of it. So I do not know what you matched and where  - initial orders are all passive - then every child order after not getting filled becomes marketable limit.

I called your bluff. And you got nothing.

Cheers.

 

PS. I guess If you knew what you were talking about then you would be writing these algos and not providing market data. Foxriver, BARX, and CS make way more than any market data provider can ever hope to earn. I actually have specific revenue numbers too if you would like to know - even dailies.

 

PSS. Do not want to be mean about it as I am sure you spent a long time on this, but it just shows that analysis that does not involve actual people responsible for these algos could be misleading. Of course, there is a problem of asymetric information where you do not know whom to trust given that the other side holds all the answers, but in the future I think it would be best to consider interviewing (maybe in private) folks who were there and know what really happened.

Thu, 03/28/2013 - 08:59 | Link to Comment ghengis86
ghengis86's picture

I was going to junk your first comment, but then this certainly looks like there may be more to the story than even Nanex has access to...

let's go down this rabbit hole....

Thu, 03/28/2013 - 13:05 | Link to Comment Doode
Doode's picture

Nanex' response is boring and childish at best and unprofessional at worst. "I watched, I saw, there is a paper coming" - if observing the market was enough to understand market behavior then I would agree - it is not. They do not take into consideration latencies, bid/offer dynamics in response to child order postings and changes in child order posting dynamics driven by the algo. The behavior on a high level is relatively simple, but on the tactical level is very complex. The bottom line is the order pushed the market and it should have been halted by either WR or barcap first line or second line or when all else fails third line. They have an OrderWorker monitor showing all the warnings about what those child orders are doing, what participation rate level is selected and etc. They cannot reject the order due to their fudiciary obligation to the client (WR) and have to execute it no matter how stupid the order is. That said someone should have called WR and told them they would cause a serious market moving event by continuing executing that order.

 

 PS. Everything that happened is a product of a finite state machine with a very small degree of randomization built in.

Thu, 03/28/2013 - 13:41 | Link to Comment ghengis86
ghengis86's picture

I hear what you're saying, but that still doesn't address the matched orders from WR/BarClays. Unless the contention is that those executed orders did push the market and if so, how?

Let's figure this shit out and not piss all over each other

Thu, 03/28/2013 - 15:07 | Link to Comment Doode
Doode's picture

This is what I am saying - orders are sent to become marketable limit orders after some time passes and no fill happens. Servers are windows C# located in NJ sending orders against market makers located in Chicago - there is a latency. Some of those marketable limit orders are bound to be hit for a little size, but once the underlying begins moving each marketable order is almost always behind what is the best bid is. Now what it does have is size - so from MM perspective it looks like a large offer keeps stepping down and pushing the rest of the index down. Makes sense? That is why participation level is key - those child order number is fixed and the number of executions is fixed for each time period with some small variation - the only thing not fixed is the size of each child order. If the rate is too high then each child order is going to look like an elephant in a china shop and start driving the market in a direction of the order. That is as old as day trading itself - buy 100 shares on a bid and put an offer out 2 cents higher for the same amount. Then put a 50000 bid a penny below or at the bid - the market would jerk up and take you out, then cancel the bid. Except no one is canceling the bid here and keeps moving with the market. And in this case it was moving an offer down (in fact if I am not mistaken it would have 6 separate child orders stepping down sequentially and hidden Iceberg quantities only work for so long once you hit that offer and it keeps refreshing with new size).

Thu, 03/28/2013 - 09:12 | Link to Comment nanex
nanex's picture

Nice try with the bamboozling. We matched executed orders from Barclays statements to CME time and sales/depth of book. The algo sold very few contracts during the 600 point slide in the DOW. There's an academic paper coming out shortly that found the same thing. 

You wrote "parts of it"? I didn't realize something that simple would require a team to write. You also refer to this algo as if there was only one. Dude, do you even code?  Or are you the guy that makes sure the coffee is always fresh? It's an important job, to be sure.

Out.

nanex.

 

Wed, 03/27/2013 - 21:07 | Link to Comment jmeyer
jmeyer's picture

I was trading the emini500 that day and what is called a " flash crash " was just an ordinary mini panic that any group of trading humans could have caused. HFT and all the faster interconnectedness that we love just accelerated the drop AFTER A POINT OF BID FAILURE. This is fear as opposed to greed and happens all the time in trading. The drop started slowly and then accelerated when the bidders broke down by the ton of sellers. I'm sure all of you traders have seen what can cause a real " flash crash " when some news, usually at 0830 or 1000, causes the market to move INSTANTLY. We're lucky that there's a lot of traders on either side of the last price in the DOM who can buffer things that can cause panicky moves. And also, how many more of these " flash crashes " have occurred since that time ? Check your DOM and be happy that there are thousands of contracts trading on the bids and asks, and don't forget to use STOPS.

Wed, 03/27/2013 - 21:11 | Link to Comment notquantumdum
notquantumdum's picture

i seem to vastly prefer "mental" stops over stop orders.  What do you think?

Wed, 03/27/2013 - 22:28 | Link to Comment Itch
Itch's picture

You're an imbecile pal, "checking your DOM"... you dont see these leaches sucking the life out of everything? Cant you see how sewn up this market is? Your telling people to check thier fucking DOMs for last years price? When you are finished checking your DOM perhaps you might want to check out the $800mm data centers popping up beside exchanges... and your edge is your DOM? Please, wake up and smell the piss on your face.

Wed, 03/27/2013 - 21:42 | Link to Comment jmeyer
jmeyer's picture

ANY stops are fine

Wed, 03/27/2013 - 21:55 | Link to Comment notquantumdum
notquantumdum's picture

Ouch, 'have I been burned by "stop orders" entered!

Maybe, your right, but I'm not sure.

The goal seems to be "defined" risk, being established, and no more, so as long as that is achieved, I guess your suggestion helps.

Wed, 03/27/2013 - 21:59 | Link to Comment notquantumdum
notquantumdum's picture

'sorry . . . "maybe you're right" . . .

Wed, 03/27/2013 - 22:03 | Link to Comment Herkimer Jerkimer
Herkimer Jerkimer's picture

'

'

'

'

What? No fat fingers?

 

•J•
V-V

Wed, 03/27/2013 - 22:24 | Link to Comment FecundaGoat
FecundaGoat's picture

It's not a Stop Loss.....It's a START Loss....Courtesy of HFT Bots....

Wed, 03/27/2013 - 22:33 | Link to Comment CharmaChameleon
CharmaChameleon's picture

Apropos of nothing, possibly, but.

Five months after May 6th is not October 6th.

Wed, 03/27/2013 - 22:33 | Link to Comment Palladin
Palladin's picture

And for those of you that missed it. Here's what it sounded like in real time.

http://vimeo.com/11570780

Seems like just another day on Wall Street. 

/sarc

Wed, 03/27/2013 - 22:49 | Link to Comment BullyBearish
BullyBearish's picture

Careful Boys...$5B POMO tomorrow, Thursday--last day for month/quarter-end window dressing...wanna bet there's a new sheriff in S&P Town?

Wed, 03/27/2013 - 22:54 | Link to Comment ozzz169
ozzz169's picture

when will these idiot polititians do something that is actually useful, and ban this BS.  So there is no upside for anyone other then the owners of the computers.  and plenty of downside for everyone else.  the arbs will keep the markets in balance, without the risk of craziness. 

Thu, 03/28/2013 - 00:02 | Link to Comment tmosley
tmosley's picture

Too big to fail.  Too big to prosecute.  Too big to stop.  Too big to admonish.  Too big to criticize.  Too big to et cetera.

They know, but they suffer under the delusion that to stop the people doing these things would destroy the economy, or suffering under the cold hard fact that if this shit blows up they will lose thier cushy government jobs and their fat pensions.

Thu, 03/28/2013 - 00:26 | Link to Comment putaipan
putaipan's picture

i'm still looking for the pre-9/11 harvard law review's article suggesting "corporate satirsts" be determined 'terrorists'.any help with a way back machine archived linky? ....'cause it's gone.

Thu, 03/28/2013 - 00:39 | Link to Comment ECE
ECE's picture

a nice commentary but not exactly true.   He describes the currently daily Fed prop but the flash crash wasnt a hot potato but  a volumeless range with one very large fat finger.   it has already been discovered.

the high freq BS that we see today should be paid more attention too.  this rally is the greatest fraud ever perpetuated on the american people.  it is about to fail very very hard

Thu, 03/28/2013 - 00:43 | Link to Comment Monedas
Monedas's picture

The bidders didnt demand lower ask prices .... the sellers just spontaneously dropped their pants ?

Thu, 03/28/2013 - 00:50 | Link to Comment resurger
resurger's picture

Algos fucking Algos ... hope benny copter crashes

Thu, 03/28/2013 - 03:22 | Link to Comment MythicalFish
MythicalFish's picture

Question remains: who dumped this much ES on HFTs in the first place? Europe? VIX activity 1pm onwards looked suspicious.

Thu, 03/28/2013 - 03:28 | Link to Comment NaN
NaN's picture

Thank you, Nanex, for expending time and effort for the good of all.

The SEC, where people get paid to foot-drag^H^H^H^H^H^H^H^H^H regulate, deserves no thanks at all.

 

Thu, 03/28/2013 - 06:02 | Link to Comment kurt
kurt's picture

I just added liquidity, in my pants!

Thu, 03/28/2013 - 06:54 | Link to Comment Black Markets
Black Markets's picture

Interesting stuff but the commentary is still fundamentally flawed.

 

To describe things accurately they need to distinguish between bid liquidity and ask liquidity. Not enough people recognise that the bid and the ask are effectively two separate markets with virtual arbitrage across the spread.

Also Algorithms and HFT are not necessarily one and the same. All HFT are algorithms but not all algorithms are HFT.

 

Finally you cannot comment on HFT or algorithms without an understanding of ‘price impact’.

 

The best algorithms in the world aim to emulate a simple 4 stage cycle.

 

<repeat>

1.      

Purchase high volumes with minimum price impact

2.      

Purchase low volumes with maximum price impact

3.      

Sell high volumes with minimum price impact

4.      

Sell low volumes with maximum price impact

</repeat>

 

This is all they do. If you don’t understand methods of affecting price impact in a dual quote consolidated order book then you are talking shit.

 

All this talk of “rebalancing” is bullshit that this guy has dreamt up to explain his own bullshit theory of the algorithm’s motives. Good algorithms manage risk away from the high volume main market by buying and selling highly leveraged derivates and options in alternative markets with lower information value, due to their lower volume. It’s called hedging, something this guys has clearly never heard of.

 

Game Theory and Emergence Theory means that algorithms following this 4 stage cycle will slowly but surely find themselves in synchronization (like women), because they can achieve maximum price impact when acting simultaneously. This synchronization risk is what is new and is also what triggered the flash crash of May 6th. But the real cause of the crash was the limited capacity of the market infrastructure to handle the volume and provide all participants with equal information. This lack of infrastructure led to information gridlock which triggered a brief cardiac arrest of the financial system as many participants, reading crossed price feeds, hit circuit breakers or literally pulled the plug creating even more imbalances, distortions and dislocations.

 

The true villain here is the exchange, who failed to provide sufficient infrastructure to handle the market they were hosting. And anyone who says otherwise is full of shit.

Do NOT follow this link or you will be banned from the site!