Asymmetric vs. Symmetric Trading Systems: Asymmetry or Symmetry?
As noted in Trading vs. Investing, when I speak of "trading", please note that I am implying buying and selling something. Trading is defined as: the act or process of buying, selling, or exchanging. The term "trading" is not related to the frequency of buying and selling, but instead the act of buying and selling.
Symmetric Trading System
A symmetric trading system is one that has an entry and an exit that is the same. It's a system or strategy that sells a position with the same signal it uses to buy it.
It may buy when the price is above its 50 day moving average and then sell when the price falls below its 50 day moving average.
If you apply symmetry to chart patterns, it may buy a Point & Figure Double Top breakout and sell a Point & Figure Double Bottom breakout.
If you apply symmetry to fundamental ratios, it may buy if the P/E is below 20 and sell if the P/E is above 20. You probably get the picture. A symmetrical trading strategy is one that assumes prices rise and fall in the same fashion, so it treats them the same with the same buy and sell signal.
Asymmetric Trading System
An asymmetric trading system is one that treats the entry and exit differently, so it has an entry and exit that are different. For example, it may buy above the 200 day moving average, but it sells if the price falls below the 50 day moving average. Asymmetric trading systems may have an entry signal that is different from the exit signal. In fact, an asymmetric trading system may exit a position applying a totally different method than the one it uses to enter. For example, I enter with a directional price indicator and exit with an unrelated risk management system.
Why would you apply an asymmetric trading system that has a different entry signal than its exit signal? If you believe there is a difference in the uptrend than the downtrend. An asymmetric trading system treats rising markets differently than falling markets. That is, it assumes a fundamental bias between the positive movement and the negative movement. For example, U.S. stock market returns are asymmetric: they present a non-normal distribution. Stock market return data has “fat tails”: the downside tends to be much greater and occur more often than “normal”. Stocks tend to behave differently on the upside than the downside. You can simply view an index price chart to see this. Stocks tend to have more gradual upward trends, but they collapse and cascade on the downside as investors' panic. In The truth about asymmetric market returns: fat-tails and black swans I pointed out a study that finds:
These findings prove that risk estimates based on the normal distribution, even with volatility clustering, systematically misrepresent the true nature of the risk being taken. More sophisticated modeling techniques are necessary in order to capture widely observed phenomena such as risk asymmetry and fat tails.
In Investor Reaction to News is Asymmetric we pointed out some evidence of why I believe investors react differently to good news than bad news. Investor reaction to good and bad news is asymmetric: bad economic news is likely to play a more significant role in shaping investors’ behavior than good economic news.
For this reason, our asymmetric trading systems handle rising markets differently than falling markets. If you believe market returns are asymmetric, it probably doesn't make sense to apply symmetry to asymmetric returns. I discovered this long ago through quantitative testing and I have experienced it in real-time portfolio management.
Even if you believe the market you trade doesn't present an asymmetric return distribution, there are other reasons to apply an asymmetric trading system that sells positions differently than it buys them. One of those reasons is the exponential nature of losses. As I discuss in the Asymmetric and Exponential Nature of Losses: losses compound asymmetrically and exponentially. That is, losses are not linear because the compounding of losses doesn’t occur in a straight line fashion. You can probably see why those who apply mathematical equations like those used in Modern Portfolio Theory underestimate risk. Their equations and methods don’t give enough respect to the big moves at the ends. Those big moves are the factor that makes all the difference if you want to capture big gains and avoid large losses. A -50% decline needs a 100% gain to get back to even. Yet some markets, like stocks, show a history of a -50% decline occurring quickly, but that 100% gain to get back to break even seems to come much slower. The recent stock market history is a good example. The S&P 500 stock index declined over -50% in a year and three years later it remains about 20% below it's prior high.
You may guess that I probably exit positions differently than I enter them. If investors want to pursue a positive asymmetric return profile that results in a larger total return than the downside declines it takes to get it, then they should consider that we don't want balance and symmetry in all things. Some things in natural science, including the human body, look better with symmetry, but we can’t say the same about investment returns.
An asymmetric trading system is a quantitative portfolio management system that buys and sells with the understanding that the dynamics may be different when a trend is going down than when it is going up. Investors who seek an asymmetric return profile should be sure their investment manager or strategy is one that understands and exploits asymmetries.
Asymmetry™, Asymmetry Investment Program™, and Asymmetry Capital Partners™ are registered trademarks of Shell Capital Management, LLC
Comments (4)
Read through and enter the discussion by using the form at the endChuck Lowman - January 24, 2012 10:22 AM
Based on this post could you say that; Your exit point or stop would generally or always, be tighter than your profit exit point. If so do you believe that the tighter the stop point, the greater the loss will be in time, position sizing being equal? Under what circumstances would a predetermined stop point be greater than a predetermined exit, profit point?
Mike Shell - January 25, 2012 4:31 PM
Chuck, for me, all of your questions are a function of a complete system. It's not that someone can tell another if an exit point should be tighter or loser. Those answers are a function of the overall system. I'm speaking of a quantitative trading system that is applied systematically. That is, as opposed to a discretionary strategy where a person makes intuitive decisions as they go.
For me, those things were answered years ago in the development and testing process. However, we can say this:
1. If you believe the markets decline faster and sharper than they rise, then you may consider the possibility intuitively that your exit may need to be faster than your entry. In Tennessee, we all that "common sense", and it's sometimes a good starting point. :)
2. "do you believe that the tighter the stop point, the greater the loss will be in time". A tight stop point doesn't reduce your risk if it's too tight, say, within the normal daily noise of the market. If a price swings plus or minus 5% in a normal day, then a stop within that is sure to get stopped. If you do that very often, you'll have a lot of small losses which may lead to a larger overall loss.
3. A systems predefined exit point, that is, the initial risk amount at the point of entry, may very well be very different than it's exit point after a position has become profitable. There are all kinds of different exit methods. Though I don't discuss how our systems exit, I will tell you that there ere multiple exit systems involved and they are asymmetric, meaning they are not the same as say, the one that determines how much we risk at the beginning.
All of these things were determined during the design and development stage. The only way to understand how a method or parameter works is to test it and then use it in real time. I tested thousands of exit methods to determine which we'll use as part of a complete system. It's a function of the other parts of the system and our objectives for total return and risk control. I decided nearly a decade ago exactly how I'll exit my positions today. I don't sit around trying to figure out when to sell or what's going to happen next.
Tim - January 25, 2012 5:28 PM
Great post Mike. I know a lot for the systems i have looked at are more "binary" in which they looked at sectors and determined if the light switch for each was either on on off. Trying to add some positon sizing to the equation did help in some circumstances but didnt eliminate the cause as described in your article. That is using a an asymmetric exit strategy that doesnt use the same buy indicator in reverse as a sell indicator.
Again, good article and definately lends itself to some good independant thinking...
Mike Shell - January 25, 2012 9:38 PM
What you described is more typical. In fact, it's the method applied in most academic studies of relative strength and momentum. Keep in mind that what I described is a concept that defines a factor of my actual systems that have been managing real money for seven years. For me it's well beyond the hypothesis, testing, and simulation stage.
It seems many systems that are publicly known or written about are incomplete. Maybe they are made public by inexperienced researchers. You may consider that no one has any incentive to make public any specific methods that may actually work. For example, if you write to popularize a moving average, you may find that it no longer works so well. If we know it's become popular, we can take advantage of the liquidity that occurs at the signal.