At 930AM Wednesday, shares of Kraft Foods Group, Inc (NASDAQ:KRFT) surged by more than 29%, hitting a high of $58.54 after opening at $45.36. Incredibly, it took just 19.93 seconds to notch this gain, a rate of return regular investors can only dream about. While market pundits draw wrong conclusions, and traders around the world scratch their heads, let us take a closer look at the data to see how things actually unfolded.
The prelude to this extraordinary event was the spinoff of Kraft Foods Group, Inc. (KRFT) from parent company Modelez International, Inc. (NASDAQ:MDLZ). As a result of the spinoff completed on 10/2/2012, Kraft (KFT) was split into two companies, which began trading on NASDAQ as KRFT and MDLZ. More significantly, KRFT was added to the S&P500, replacing Alpha Natural Resources (NYSE:ANR) in the venerable index. The implication of this is that every fund and ETF which tracks the S&P500 (and there are a LOT of funds that do this) all need to hold some shares of KRFT, meaning that broadly speaking, KRFT should only move in one direction: upwards. Indeed, KRFT notched solid gains in its first day of trading and should continue to see upwards pressure in the days to come.
Whenever there are institutional buyers involved, high frequency traders are bound to congregate like sharks in the water. This is because institutional buying provides a strong support for frequent traders, thus reducing risk and limiting potential losses. Simply put, if you put your order ahead of a big institutional order, even if the market goes against you, it will hit the institutional order right below you which will prop up prices at that level. Furthermore, on a 'new' stock like KRFT, bid/ask spreads tend to be large as market participants are still performing price discovery, setting up the perfect environment for high frequency scalping.
So, to recap, at open on 10/3/2012, KRFT is a stock with a large bid/ask spread, monitored by hundreds (perhaps thousands) of high frequency algorithms, and had a relatively thin book on the sell (ask) side. Indeed, this is the situation that persisted 17 seconds and 229 milliseconds after open as we can see in the following price table*:
*Historical stock price data reproduced with permission from QuantQuote Inc.
Times are given in millisecond past midnight, so the open is 34200000. Also notice that KRFT is not shortable (it has short status X) so the downside risk for high frequency traders is further reduced. The non-shortable status is likely due to the fact that KRFT is a 'new' symbol. Additionally, KRFT is exhibiting a roughly 1% bid/ask spread making the stock extremely attractive to scalping strategies (where one would buy at 46.00 and sell immediately afterwards at 46.45). The role played by dark pools can also be seen here (trades with exchange code D represent dark pool trades). Algos are actively picking up cheap shares from dark pools where they are fewer participants (and potentially less competition for the shares), then filling market orders from institutional or retail traders on public exchanges.
A mere 50 milliseconds later, an irreversible chain of events begins which destroys the delicate balance we see above. At 17 seconds and 268 milliseconds after open, the inside ask order of $46.50 is promptly filled (or more likely, intentionally pulled), and in successive ticks over the next 50 milliseconds, the ask rises promptly to $48.33, a nearly 4% spike.
At and during this stage, the thousands of HF algorithms monitoring the stock (many which employ momentum strategies), rush to buy given the low perceived downside risk. At the same time, the algos that already have shares suddenly aren't selling anymore, and immediately, the ask side of the book vanishes (the -1 is a proprietary QuantQuote live data feed indicator that is triggered when one side of the book becomes extremely thin). Slower algorithms chase with market orders and before long, there is out of control price appreciation.
A mere 2.7 seconds after the ask price was first boosted, shares are changing hands at $58.54, with a massive $8.53 bid/ask spread. In total, 7466 shares were traded between t=0 and the first trade at $58.54. In the next couple seconds, the top scalping algos dominate the marketplace, raking in approximately $8 per share on each cycle, mostly at the expense of less sophisticated algos and retail traders.
We all know the end of the story; ultimately, all trades above $47.82 were cancelled, citing the "clearly erroneous" trade rule introduced after the flash crash of 2010. However, from the data, there is not a clear error. In fact, the market was behaving completely within the rules of the system!
So what caused the events of today? The short answer is insufficient liquidity. There simply weren't enough reasons for market participants to sell this morning, leading to an extremely thin book on the ask side. Couple this with excessive "attention" from high frequency algorithms, and you have a time bomb that takes only the slightest nudge to set off.
And, that's exactly what happened, in 2.7 seconds, faster than a human trader can ever hope to react. So as a retail investor, what can you take away from all this? The first obvious conclusion is that there are enormous profits to be made through high frequency trading, using relatively simple strategies that don't require rocket scientists to figure out (although usually rocket scientists are hired, just for good measure). For this reason, despite all the negative press and occasional mishaps, high frequency trading is here to stay, like it or not.
So how do traders compete in a marketplace full of computers? The answer, ironically enough, is to not compete. Unless you are prepared to pay for a low latency feed and write software to react to market movements on the millisecond timescale, you simply will not win. As aptly shown by the QuantQuote tick data above, the required reaction time is on the order of 10 milliseconds. You could be the fastest human trader in the world chasing that spike, but 100% of the time, the computer will beat you to it.
The second important conclusion is to avoid market orders. Unless you have access to data that tells you exactly how thin or thick the book is (and even that data is often deceptive), you could literally be standing on thin ice. When submitting an order, always remember that on the other side of your trade, is a faster algo seeking to give you a worse price, and when you submit a market order, you are giving control of your trade to that algo. Take these rules to heart, and you can avoid being the unenviable person who bought KRFT at $58.54.