Subscribe to DSC Newsletter

AnalyticBridge competition: investigate the spectacular stock market collapse of May 6, 2010

This is the largest one-day drop in history, and it is known as the flash crash. In 6 minutes Dow lost 600 points then recovered. The difference between high and low on May 6 was about 1,000 points, and Dow went below 10,000. Is it a technical glitch, foul play or natural cause?

To investigate this massive collapse, you are asked to:

- download publicly available daily stock prices (volume, open, low, high, close) for all stocks that are listed on the NYSE or Nasdaq (you can get the data from Yahoo Finance and download it using a web robot -- there is about a few thousand stocks)
- is the collapse evenly spread among all stocks? Any abnormality? What would the worse drop (for an individual stock) be if the average is 10%, given the number of stocks in NYSE or Nasdaq?
- see also this posting.


- The winner will be featured as our Member of the Month ($250 award)
- An additional $750 prize is offered
- The winner will be invited to become a member of AnalyticBridge's Executive Club and will receive our certification (none can be purchased)
- This represents a terrific career boost as we will make public announcements, press releases, etc.


Email your detailed analysis to Dr. Vincent Granville at [email protected] by December 15, 2010. Messages larger than 2MB will not be accepted. The subject line should be AnalyticBridge Competition 2010.

Views: 716

Replies to This Discussion

Actually, the question is: Is the flash crash the result of foul play, data glitch or some system failure, or natural cause. One could for example try to test the following assumption, using statistical theory:

Hypothesis: the drop was natural (and occured e.g. because we reached a level where a large number of stop orders were in place, resulting in cascading sells in a very short time period), but the recovery artificial (e.g. because such massive drops never recover smoothly based on historical data, and probably sell orders were reversed by the stock exchanges, and it is clear that the market could no longer support 11,000 for Dow based on the last 6-8 weeks).

I expect participants to use publicly available data that can be downloaded e.g. from Yahoo Finance. This data might be sufficient to reach robust conclusions, particularly if it involves a large number of stocks and indexes. But the reward will go to the analyst who provides the best arguments and deepest / most robust analysis (even if the conclusion is "I don't know" or "there is some chance - say > 45% - that based on my analysis, the conclusion is xyz"), regardless of the data set being used, and regardless of the methodology (be it very sophisticated statistical models, extreme value theory, Monte-Carlo simulations, or model-free methods such as pattern recognition or pure data mining), potential comparison with similar drops that occured over the last 50 years.

A winning solution will have several of the following: cross-validation, data quality analysis, good understanding of the mechanisms at play, ability to process moderately large data sets, good design of experiments, model fitting and model selection, justification of the randomization or sampling processes being used.
See also this interesting article discussing spread between ETF's and individual stocks exploited by high frequency traders (arbitrage).
Simulations such as those displayed at could help determine the frequency of flash crashes with a 10% intra-day drop.
This competition seems very interesting - it's too bad I discovered it with just ten days remaining!

I'm already competing in another competition, but this one seems much better. Vincent, have there already been a number of high-quality submissions, or do you think there is room for late entrants?
We haven't had many responses, and we will extend the deadline by 3 months. This will be mentioned in our August newsletter, and elsewhere.

Thanks Amit,
On Wednesday September 8, 2010, 1:09 pm

WASHINGTON/NEW YORK (Reuters) - Regulators probing the stock market "flash crash" last May still have not uncovered a single cause but will point to "stub quotes" and other previously identified issues as having exacerbated the market's dramatic drop, according to two sources familiar with the probe.

A third source said the U.S. Securities and Exchange Commission is still asking about a "smoking gun" that might explain the May 6 crash, when the Dow Jones industrial average plunged some 700 points before sharply recovering, all in about 20 minutes.

Regulators are soon due to issue a follow-on report on the crash, which rattled investors worldwide and exposed flaws in the high-speed electronic marketplace.

So far, the report by market regulators does not contain a lot of new information and is expected to repeat earlier findings that a number of events caused the crash, two sources said. The sources requested anonymity because regulators are still collecting data and finalizing the report.

The sources said the report will point to stub quotes -- orders placed by marketmakers that are well off the market prices for stocks -- as one of the structural issues that contributed to the plunge.

"Quote stuffing," in which large numbers of rapid-fire stock orders are placed and canceled almost immediately, will not be fingered as one of the causes of the crash, sources have said.

The SEC is increasingly probing market data from other trading days, looking for possible problems with excessive numbers of buy and sell orders, the third source said.

The SEC has already adopted a pilot program to help prevent a repeat of the crash. That circuit-breaker program pauses trading in a single stock if that stock is in crisis.

The SEC also wants to ban stub quotes and is expected to propose such a rule in the near future.

(Reporting by Jonathan Spicer and Rachelle Younglai; editing by John Wallace)

Source: Yahoo Finance.
New article in the New York Times: Ex-Physicist Leads Inquiry into Flash Crash

WASHINGTON — As a doctoral candidate in physics at Princeton two decades ago, Gregg E. Berman spent a year and a half in a laboratory searching through subatomic data for an elusive particle called the heavy neutrino.

Now, from his small office at the Securities and Exchange Commission here, the former physicist is busy completing a similarly painstaking task, supervising a team of more than 20 investigators who have spent the last five months scrutinizing reams of stock-trading data and hundreds of interview transcripts in an effort to figure out why stock prices went into free fall for 20 terrifying minutes on May 6.

Their long-awaited report on the so-called flash crash, in partnership with the Commodity Futures Trading Commission, is due to be published in the next two weeks.

Mr. Berman, 44, will not say exactly what will be in the report, but he says that it will not simply restate what regulators have already said — that markets were volatile because of worries over the debt crisis in Europe, causing some computerized tradi-ng programs to stop trading, and finally causing computers on other exchanges to misread the pullback as a rapid bidding down of stock prices.

Instead, he says, the report will zero in on a specific sequence of events that preceded the crash. He says it will tell a clear story about what happened in the markets on that stomach-churning day, beyond simply pointing a finger at the perils of the kind of high-speed computer trading that dominates today’s markets.

“The report will clearly demonstrate how market conditions and events prior to the flash crash led to the extreme price moves,” he said.

When pressed, he added, notably, that he had found no evidence of a deliberate attempt by anybody to disrupt markets.

The implications of the report are not merely academic. Ordinary investors, shaken by the brief stock plunge and the lack of an official explanation, have withdrawn money from stock mutual funds every week since the crash. Market analysts say investors want to be reassured about the integrity of the nation’s markets so they can be confident that a nose dive will not happen again.

The Berman report will not be the final word on the matter. Its findings will be used by a group of advisers to the S.E.C. and the commodity futures commission, which will make policy recommendations.

Still, some analysts question whether the report can deliver a simple answer that will satisfy everyone eager for reassurance.

“What everybody would love to hear from the S.E.C. is XYZ trader blew up the market and made a gazillion dollars and is now in jail,” said Larry Tabb, chief executive of the Tabb Group, a specialist on the markets. “The answer, I think, is much more complicated and nuanced and has to do with a lot of different things. I am not sure that everybody outside the industry is going to have the patience to understand that.”

Mr. Berman acknowledges that his team’s explanation will involve a number of things happening at once. It may strike many people as painfully complex, but that is an undeniable result of the byzantine nature of today’s disparate electronic markets and the many players who take part in them.

The report’s conclusions will involve “market participants doing very different things and for very, very different reasons,” he said.

Central to all of this is the fact that stock trading is no longer centralized but instead takes place on dozens of exchanges, all with varying policies and procedures. For example, the New York Stock Exchange has circuit breakers that prevent stocks from rising or falling so quickly that they disrupt the broader market.

Trading was slowed on several listings on that exchange on May 6, while other markets kept trading lower. That lack of coordination created confusion during the flash crash. Since then, the S.E.C. has extended circuit breakers for individual stocks across all markets.

In investigating the crash, Mr. Berman says he finds himself in a position similar to his physics work 20 years ago, when he was collecting huge amounts of data and comparing the competing views of many laboratories on a question dividing particle physics — whether the neutrino, one of the least known and most common elementary particles, actually had mass.

Today he finds himself in familiar territory, sifting through huge amounts of messy and disjointed data, and at the same time reading blogs and e-mails from a wide range of observers, each with a theory about what happened on May 6.

Full article with links available at
You can read the official SEC analysis at

Our contest is still open. And no, the $1,000 prize will not be offered to the SEC.

SEC Releases Final Flash Crash Report - Waddell And Reed Blamed As Selling Catalyst

Tyler Durden's picture

100+ pages of garbage, and yes, Waddell and Reed, and their destructive hedging trade of 75,000 e-minis is the culprit that apparently set everything off: "At 2:32 p.m., against this backdrop of unusually high volatility and thinning liquidity, a large fundamental trader (a mutual fund complex) [ZH: Waddell and Reed]initiated a sell program to sell a total of 75,000 E-Mini contracts (valued at approximately $4.1 billion) as a hedge to an existing equity position." Apparently, its was all evil W&R's fault who traded in conjunction with an even more evil VWAP algo: "This large fundamental trader chose to execute this sell program via an automated execution algorithm (“Sell Algorithm”) that was programmed to feed orders into the June 2010 E-Mini market to target an execution rate set to 9% of the trading volume calculated over the previous minute, but without regard to price or time. However, on May 6, when markets were already under stress, the Sell Algorithm chosen by the large trader to only target trading volume, and neither price nor time, executed the sell program extremely rapidly in just 20 minutes." In other words, there never was, is and never will be any liquidity in the allegedly most liquid market - ES. Good luck to all those who hope to sell the second there is a second flash crash.

And here is the initial mention of HFTs:

HFTs and intermediaries were the likely buyers of the initial batch of orders submitted by the Sell Algorithm, and, as a result, these buyers built up temporary long positions. Specifically, HFTs accumulated a net long position of about 3,300 contracts. However, between 2:41 p.m. and 2:44 p.m., HFTs aggressively sold about 2,000 E-Mini contracts in order to reduce their temporary long positions. At the same time, HFTs traded nearly 140,000 E-Mini contracts or over 33% of the total trading volume. This is consistent with the HFTs’ typical practice of trading a very large number of contracts, but not accumulating an aggregate inventory beyond three to four thousand contracts in either direction. What happened next is best described in terms of two liquidity crises – one at the broad index level in the E-Mini, the other with respect to individual stocks.

Read full article at


On Data Science Central

© 2021   TechTarget, Inc.   Powered by

Badges  |  Report an Issue  |  Privacy Policy  |  Terms of Service