Saturday, July 28, 2007

How Trading Causes Psychological Problems

The assumption of practically every trading coach and psychologist is that psychological problems and biases can hamper sound trading. That happens of course and is well documented in the behavioral finance literature. Less discussed is the reverse: how trading can contribute to psychological dis-ease.

It's not surprising that this topic is rarely discussed. It's hard to believe that promoters of trading conferences/workshops, educators connected to brokerage houses and software firms, and coaches that assist people with trading would see much upside in promoting the view that trading causes emotional disequilibrium.

My email inbox has seen quite a few notes from traders in the past few days, as the market volatility has taken a toll on their accounts and on their psyches. Traffic on my blog routinely goes up during periods of high market volatility; it soared in late February and early March and is on track this month to set an all-time record thanks to the last few days.

The reason for this, I believe, is that volatility upsets the normal expectations of traders, as markets move more and faster than usual. In the midst of the uncertainty and discomfort of the unknown, traders seek information and turn, among other places, to the blogosphere.

I recently updated the Trader Performance page with a post on non-stationarity in markets and a theory regarding market participation. I think it's an important post, as it offers an explanation for why volatile markets are qualitatively different--not just quantitatively so--from other markets. The difference between trading a high volatility ES market and a low volatility one is every bit as profound as the difference between trading the ES market and trading, say, natural gas futures.

That is rarely if ever acknowledged. If markets become different creatures under varying conditions of volatility, it means that many trading signals, indicator readings/patterns, stop loss levels, and profit targets may become at least temporarily obsolete. And therein lies the cause for distress--and why I receive so many emails during these volatile times: just when the market becomes most enticing (greatest movement) and most risky (greatest opportunity for adverse movement), it behaves differently. What may have worked under conditions of moderate volatility no longer works in the high volatility environment because--as my Trader Performance post notes--a different class of trader, trading a different way, is dominant at those high volume/high volatility occasions.

Let's look at the objective evidence. The above chart takes the absolute value of market movement in percentage terms every 15 minutes for the ES futures (close to close) and sums the values for the past day. This provides a measure of the total market movement (excursion)--up and down--during the day. This is a true, price-based measure of volatility. What we see is that the recent values exceeding 7% are seven times higher than the values we saw in 2006 and almost five times higher than the levels seen just a few weeks ago.

If we define an intraday trend as a movement of a given amount in a given period of time, we can see that the volatile markets of late have offered many more short-term trending moves than prior markets. A discretionary trader who has developed a feel for normal markets--and their intraday reversals--now finds that the market moves far further before reversing and these reversals themselves travel much further, much faster. It's as if a slow-pitch softball player suddenly had to face a baseball pitcher: the game seems the same, but everything from the batter's perspective has changed.

Moreover, the recent period of hugely different movement is not unique. We can see from the above chart that the vast amount of market days have moved a total of 3% or less in their total excursions. At times, however, markets have suddenly moved twice that amount or more.

Consider further: From March 15, 2006 to July 19, 2007, the average 15-minute period in the ES futures traded 37,071 contracts and there were an average of 1893 trades. The average high-low range for each 15-minute period was .17%.

Since July 19th, however, the average 15-minute period in the ES futures has averaged 79,676 contracts and there have been an average of 4963 trades. The average high-low range for each 15-minute period since July 19th has been .34%.

In short, we've seen more than a doubling of volume and trades and a doubling of price movement.

If we just look at the market since July 26th (the last two days of trading), the average 15-minute volume in the ES futures has been 106,391 contracts, and there have been an average of 7418 trades during each period. The average high-low range for each 15-minute period has been .51%: three times the average from the sample going back to March, 2006.

What would happen if you took a golf player and made him play on greens that were three times as fast as normal ones? What would happen if you took a racecar driver and had him compete on a course at three times his normal speed? What would happen to air traffic controllers if planes began moving at triple speed?

They would lose their feel for their work. They would perform poorly. They would become distressed.

Volatile markets compress time. What normally happens in a day can happen in an hour. Once time compresses, performance activities become different. That's why lightning chess is a different game from standard tournament play.

When you have many more traders in there--professional traders who are moving size--markets become faster and traders become like the air traffic controllers in the example. They cannot keep up; they become disoriented. Trading, which had been a source of mastery, now causes discomfort.

Of course, no one really makes radical changes in the surfaces of golf courses, the speeds of racecars, or aircraft speeds. Markets, however, change their speed continuously--and at irregular and unpredictable intervals. During one period, stocks may trade slow like Treasury bonds; at other times they are volatile like agricultural commodities.

Moreover, highly volatile periods are sufficiently rare that most traders have not experienced enough of them to develop a true feel for their movement. That is why such periods offer as much risk and danger as reward.

Is there an answer to this dilemma? I believe there is: as markets trade faster, traders can reduce their size and moderate their trading frequency, so that each trade and each trading day will not put capital at enhanced risk. That can give the trader sufficient time to adjust to the higher volatility environment and develop his or her feel.

If traders trade their usual size--and especially if they get carried away with the volatility and trade more often--they now place their portfolios at twice or three times the normal risk--maybe more. With the quicker and different movement, drawdowns become far more likely, and they can be deep. It's not at all difficult to lose as much in a single day as one would might ordinarily lose in a week.

Keeping statistics on your trading at low volatility times, moderate volatility periods, and high volatility occasions will tell you whether, in fact, your trading changes as a function of market movement. If your percentage of winning trades differs significantly under these conditions or if you see meaningful differences in the average sizes of your winning vs. losing trades, that will tell you that changing markets are changing your trading.

In markets, as in medicine, above all else do no harm. You can't win the game if you don't stay in the game. When markets change their trends and movement, give yourself time to adapt. Overtrading a volatile market without a good feel for what's happening is a sure path to ruin--and I have the emails to prove it.


Trader Dan said...

Great post. Not a point I've seen covered before. While you direct the issues\questions to the discretionary trader, I'm thinking they have equal merit for the system designer. Agree?

Wen said...

An excellent post. Your description of volatility and the system's performance is exactly something I have been thought about recently. Prior to this month my 30 minutes trading system consistently outpeform my 15 minutes trading system. For this month I have very good result for the 15 minutes trading system but my 30 minutes trading system suffered some small loss.
The decision I have to make right now is whether this high volatility is for here to stay for forseeable future or most likely a very short term event.

Abel said...

Dr. Brett,

A review of the chart posted, and in recounting the two other periods of enhamced volatility covered in the period, indicate that the market will continue to experience increased volatility, as it works 'back into' another period of reduced, or 'normal' levels of volatility.

My own work suggests that on Friday, volatility was approx 3 times normal, (as measured by the ER2, for almost the same period of time as covered in your example). Intererstingly, on Thursday, the level was approx 4 times normal, so that Friday, it was down, from Thurdays levels.

The prior day of max high in my records was June 13 2006, (which was approx 3 x) Before that, is anyones guess, and these levels may in their own way be max all time highs.

The main point I make is that, in a similar fashion as an intraday bottoming or topping is a process, as you have pointed out in recent articles, I believe this period of enhanced volatility is alike in that regard, which your chart also suggests, in that it will take some time to normalize, and remain higher than normal as it does, even as it slowly recedes.

Dr Bruce Hong said...

Great post.

In my experience and from what I've seen on trading websites, most trader's problems arise not from lack of emotional control, but from lack of preparation and experience.

If you don't have a trading plan that's grounded in the fundamentals and you don't know why you're entering a trade (other than to make money) then of course you're going to get into trouble. And, of course, you're not going to like that and so respond emotionally.

It seems obvious to me, therefore, that run-away emotions is the symptom, not the disease.

Brett Steenbarger, Ph.D. said...

Hi Trader Dan,

Good point. I do think these market shifts raise as much havoc for systems as for discretionary traders. The advantage that the savvy system designer has is an ability to obtain early warnings of deviations from historical performance.


Brett Steenbarger, Ph.D. said...

Hi Wen,

Thanks for the note. Volatility does display serial correlation: highly volatile periods tend to be followed by elevated volatility in the short run. Thus I use the most recent market data to try to anticipate coming volatility.


Brett Steenbarger, Ph.D. said...

Hi Abel,

Great, great points. You're absolutely right: this is part of an ongoing process, and it pays to stand back and view the market from this broader, process perspective.


Brett Steenbarger, Ph.D. said...

Hi Dr. Bruce,

Beautifully stated; I couldn't agree more. Thanks for the great perspectives on your blog--