Should the Flash Crash Change the Way You Use Stops?

喀拉峻草原

(奇摩原貼：2010/07/15)

評析:  一般書上常說的是: 機構. 大戶或法人傾向在收盤前作買賣動作, 不去理會盤中的波動, 他們要等大勢底定之後才動作, 因此散戶採取MOC(Market on Close)策略有點像在學習 [大人], 感覺上應該會帶來好處. 這篇文章卻用美股的資料提出了證據上的批判! 喜歡在現貨收盤後才做停損動作 (MOC, Market on Close) 的操作者, 或許可以得到一些啟示. 但美國市場屬性和台灣不同, 加上人家沒有漲跌停的限制, 是否窄化了MOC優點不可得知, 需要由台股自己的歷史資料去做驗證. 理性上比較可能的思考是: 如果是長線交易者 (看的 time frame 較長, ex: 以週為單位) 或許比較能夠享受到 MOC 的好處!

Articles from  http://www.iitm.com/Weekly_update/Weekly_475_May_19_2010.htm#tradingtip

by D.R. Barton, Jr.

The May 6th “Flash Crash” was notable from several perspectives: psychologically, technically and also historically. The financial markets went into a brief period of “selective liquidity” on that memorable Thursday afternoon two weeks back. 

For many traders, the biggest tangible results were stops getting hit and very inefficient fills.  In some cases, stops were hit that had huge slippage (slippage is the difference between where an order was intended to be filled and where it actually got filled).  Newspapers had stories of folks trying to liquidate positions and getting $100,000 less than they expected because prices were moving so fast.

While the event spawned numerous conversations for me, the two questions people have most frequently asked are, “What caused this monstrous price movement?” and “Should I quit using protective stops altogether or at least change to close only stops?”

As to the first question, no one has come up with a single, definitive cause.   Most likely, there wasn’t one trigger but a series of events.  Last week, we looked at the confluence of events that formed a type of “perfect storm” of market illiquidity to cause the panic drop and immediate snap back.  Click here to read that article.

But now on to this week’s key question:  Should the flash crash and evolving market conditions change the way you use protective stops?

Analysis of Stop Strategies

First, let’s consider the question:  Should we even use protective stops at all?  Protective stops, along with proper position sizing strategies, are a trader’s main line of defense against too much risk in any given trade.  For practically all types of trading, the use of protective stops is mandatory for good trading discipline.

Now,  let’s address a somewhat thornier question: Based on what happened to some stops on the flash crash day, should traders and investors switch to closing stops rather than intraday stops?

As with all sophisticated questions, there are advantages and disadvantages to both sides.  Let’s take a look at them.

Standard Market Stop Orders 

The standard tool for a protective stop is the market stop order.  For a sell stop, the order is triggered when the instrument trades at the selected stop price.  At that time, the sell stop order becomes a market sell order.  Your order is then filled by the next available buyer at the market price.

For instruments with sufficient trade liquidity, there is rarely any slippage on market stop orders and the trades are usually executed very close to, if not at, the desired price.  For example, in the very liquid S&P 500 e-mini futures market, my experience has been that at least 95% of all stop orders are filled at the desired price. In less liquid instruments, however, slippage may produce order fills at a price different than your stop because of the typically larger spread between the bid price and ask price.  In fast moving markets, stops in all kinds of issues tend to experience significantly more slippage (in relative terms to their liquidity) than in normal markets.

With that background out of the way, we can look at the advantages and disadvantages for standard market stop orders.  The major advantage of a stop market order is that it gets you out of your position immediately when the stop price is hit.  Then, if the price continues to move against you during the rest of the trading day, you have safely exited at the price that met your stop criteria. 

The main downside is mostly a psychological one:  price can trade down to your stop (in the case of a sell stop), taking you out of the position and then rebound while you sit on the sidelines.  This is what happened to so many traders with stops during the spike down and recovery on May 6th.

Market on Close Order

This type of order is designed to execute as close to closing price as possible.  In this type of order (the sell version just to keep our examples consistent), if the price trades at or below your Market on Close (MOC) order price, a market order is triggered in the last minutes of the trading day to exit the position at the first available bid price.

This is considered a more sophisticated order type and is not available from every brokerage platform.

The advantages and disadvantages basically mirror those for the standard market stop order.  On the plus side, if the market has excursions during the trading day that take you below your stop price and then back to higher prices, your position is preserved.

On the negative side (if the instrument goes below your stop price during the day), the price could continue to head down for the rest of the day.  In this case, you suffer the additional loss of the price moving down until the position is closed at the end of the trading session.

What’s a Trader to Do? 

It’s a classic standoff.  Do you risk having your position exited during a potential back and forth price movement, or do you risk having a much larger potential loss if the price moves hard against you over the course of a single trading day?

Has anyone researched the topic?  Luckily, yes.  Next week we’ll look at some very interesting research that lends some objective market price movement data to the argument. Here’s the quick answer: for most traders and investors, MOC stop losses make less sense than standard stop market orders.  Join us next week when we dig into the data! 

     Should the Flash Crash Change the Way You Use Stops? Part 2

Last week we talked about the psychological fallout from the May 6th “Flash Crash.”  On that day, the huge swing in the afternoon caused many traders and investors to be stopped out of their positions—and some with very large slippage.  As a result, these folks were out of their positions when the market rebounded only minutes later.

That event led many to ask, “Should I quit using protective stops altogether or at least change to close only stops?”

Let me reiterate my answer to the first question:  protective stops, along with proper position sizing techniques, are the trader’s main line of defense against taking too much risk in any given trade or investment.  For practically all types of trading, the protective stop is a mandatory part of trading discipline.

With that said, let’s dig into some research to see if Market on Close (MOC) stop orders might perform better for traders.

Data Analysis Shows the Way

In Part 1 of this article series, we covered the advantages and disadvantages of using MOC stops.  We learned that one has to balance two possibilities:

1.      Hitting your stop level early in the day if the market makes a wild intraday swing (like it did on May 6th, 2010), and

2.      Having the market move against you until you get stopped out at the close with a much bigger than expected loss.

There are many approaches to take if we want to investigate the usefulness of MOC stops versus standard market stop orders.  To make the study simpler, I chose to look at getting stopped out of long trades—an issue that affects most market participants (i.e., investors, long-term traders, and swing traders). 

So here’s the methodology:

• Review the daily data for the Dow Industrial index dating back to 1928 (20,500 trading days)
• Find the 100 largest range days in terms of percent moves where the close is below the open (the days when you could really get hurt)
• See how many of these days closed in the bottom 33% of their daily range—this should tell us how many of the big range days would significantly hurt those using MOC orders and how many might benefit from an MOC order

Some Surprising Results

Some interesting stats came out of the test:

• The biggest range for a down day was (no surprise) Black Monday (October 19, 1987) at 22.48%
• The “Flash Crash” day (May 6, 2010) was number 10 on the list with a 10.48% range
• 76% of these highest range days closed in the bottom third of their range
• MOST IMPORTANTLY: Of the 24 instances where the close was in the upper two-thirds of the daily range, only 2 of those days would have significantly favored an MOC close—March 26, 1929 and the Flash Crash day (the other 22 instances were either followed by big down days or occurred in the later stages of a big down move that was already underway)

So an MOC stop

• could have caused much bigger losses in 76% of the extreme move days,
• would have made little difference on 22% of the days and,
• would have been the better exit on only 2% of the extreme down days.

The evidence, therefore, overwhelming favors protective stops over MOC stops if your main goal is to protect your positions against outsized losses on days with extreme moves. 

Note: This study in its current form does not in itself form a definitive case against MOC stops. 

Interestingly, this study also clearly points out a pervasive psychological biases—recent event bias.  Our brains tend to value a recent significant event much more strongly than past significant events or even a preponderance of data.  So the fact that some positions were stopped out on a nasty drop-and-reverse day may have caused a knee-jerk reaction in some traders and investors.  To switch to only MOC stops, however, looks to be a solution for a problem that is quite rare.

Again, for most traders working in intermediate and longer time frames, standard protective stops appear to be the most prudent choice.

Should the Flash Crash Change the Way You Use Stops? Part 3

Unusual events capture our imagination. They also expose our biases in a very distinct way.  With the “Flash Crash” still visible in our rearview mirror, let’s look at one bias in particular that reared its ugly head during that brief but significant turmoil.

A number of traders are familiar with the groundbreaking work of Daniel Kahneman.  He is best known for winning the 2002 Nobel Prize in economics for a paper he wrote with Amos Tversky on Prospect Theory.  In that paper, they quantified the decision biases that arise when people approach risk decisions where probabilities are known but individual outcomes are uncertain.

Kahneman is not an economist; however, he is a psychologist.  He also wrote (or, more appropriately, edited) a lesser known work on the summary of research in the area of well-being.

This particular work thematically reinforced the biases that arise in self-reported well being—with a unique twist.  The research showed that self-report of pain is influenced by only two variables: the peak amount of pain and the last level of pain.

Kahneman took two groups of people and subjected them to the same amount of pain.  Then, he subjected the second group an additional amount of lesser pain.  Afterwards, it was the first group who reported more pain, not the second group.  Why?  It seems that even though the second group experienced a greater “sum” experience of pain, the first group’s “average” experience of pain was higher.

Essentially, Kahneman and his colleagues found that the last result experienced (or more specifically the emotion involved with that result) unduly biases our memory and recall.  Therefore, the last experience also disproportionately biases our decision making on future events. 

Kahneman called this psychological behavior the “peak-end rule.”  We simply might call it “last event bias.”  Either way, the concept has significant implications for traders.

Last Result Bias in Trading

Let’s discuss how we might find Kahneman’s theory at work in our trading.   Say you’ve had a trading system that's been performing consistently well for a period of many months.  Then, an outside event hands you an out-sized losing trade—perhaps overnight news created a big gap against your position and stopped you out for a loss two or three times bigger than you planned.  Or maybe you experience a huge and unusual move like the “Flash Crash” that gave many traders significant problems.

When the next trade set-up comes along, your mind clearly remembers that last painful experience.  You hesitate.  You contemplate changing your stop strategy to try to protect against big losses.  You question whether holding any trade overnight is really worth the potential gaps.

All these extraneous thoughts happen regardless of the success of your trading strategy leading up to your latest trade.  One recent negative event overshadowed the nice aggregate performance that your trading strategy had provided over time.

Last event bias is particularly troublesome for intraday traders, if for no other reason than it can affect their trading more because of the frequency of their trades.  Useful mental states and thought processes can quickly be replaced by negative ones that then override time-tested strategies.

What to Do

Can we overcome this subtle and rather insidious bias?  Yes.  How?  We must begin with a good sense of personal awareness, which can be developed by constantly looking at yourself from an outside party’s perspective.  Once you are aware that a decision blocker (e.g., last event bias) is creeping in, you can overcome it by dismissing extra decision loops and sticking to the established game plan.  If you haven’t looked at them in awhile, review Van’s daily Top Tasks of Trading, which can be quite helpful in minimizing this bias and others. 

Should the Flash Crash Change the Way You Use Stops? Part 4

Last week we looked at the peak-end rule popularized by Nobel laureate Daniel Kahneman.

Kahneman and his colleagues found that people remember and quantify past experiences (whether pleasurable or painful) based on just two factors: the peak level of pain and the pain associated with the last (i.e., the most recent) event.

This means that decisions are based on emotional recollections that are either the extreme or the most recent, not an average of every event. 

It is easy to imagine the effect of peak emotional states when it comes to investing and trading decisions.  One huge win or devastating loss can color our thinking and decision making for years.

Indeed, many investors and traders spend countless hours trying to recreate that one monster winner or working on ways to avoid the loss that nearly busted the account.

Peak Pain or Pleasure—Now What?

Peak pain in trading and investing is almost always a combination of psychological errors and position sizing errors.  Between those two sources, I think there are three primary types of peak pain trading experiences:

1.      A breach of stops. Moving a protective stop (repeatedly even) or just eliminating it altogether has led to more account blow-ups than perhaps all other issues combined.  Simply stated, exit discipline is vital to staying alive as a trader. 

2.      Too big of a position.  Often, such a breakdown in stop discipline accompanies a position sizing violation (e.g., entering a position that is too big from the start).  Then, the trader seems forced to wait for the trade to recover.

3.      Adding to a losing trade.  Nearly as often, big problems happen when traders add to a losing trade in hopes of conjuring a more favorable breakeven point.  How do you think that typically works out?  

Avoiding these peak pain experiences simply involves sticking to a written plan that must include the stop loss point for each strategy and a position sizing model for each trade/investment. 

Every trader already has at least one painful experience in their memory banks.  Has anyone ever been hit by an “outsized” gap against them?  Sure.  Is it likely to happen again in the future? Absolutely.  How big of a problem is it, though, if there’s a plan in place with appropriately sized positions?  Even a big gap should cause only moderate harm rather than devastating damage.  While a recent painful loss may tempt us to change our plans, we have to ensure that we make plan changes based on a statistically significant sample of events and not one very painful experience. 

In a similar vein, trying to recreate pleasurable trades causes us to concentrate on the events that seemed to cause the original great trade.  This means we often go back to the same stock or sector, effectively ignoring other better candidates.  It also can lead us to concentrate our efforts searching for the same perfect confluence of events that may not happen again for quite some time.

The Role of Personal Responsibility

It’s a common occurrence for a painful trade to leave us wanting to change trading rules based on that single event.  We can make the situation even worse, however, by trying to avoid personal responsibility for the painful trade.  If we blame the bad experience on the markets, a news event, the market maker, some floor trader, our broker or anyone else rather than taking full responsibility ourselves, we will most likely try to avoid the situation or alter systems or strategies instead of working on our personal discipline or other psychological issue that were actually the problem.

Think back to a highly-memorable painful trading experience of your own.  Were you responsible for the outcome or did it just happen to you?  If it seems to be the latter, try to think very objectively and consider each decision point in the process where you could have made a more constructive choice. Where could you have done something differently that would have reduced the pain of the final outcome?  In almost all instances, there is at least one area (and sometimes several) where you could have made a different, more useful decision that would have reduced the financial and emotional impact of the event.  Once you’ve identified one or more decision points, own your part in creating the result you got.  Take responsibility and dig for creative ways to avoid that decision breakdown in the future.