,hl=en,siteUrl='http://0ldfox.blogspot.com/',authuser=0,security_token="v_SeT2Tv8vVdKRCcG9CCW-ZdIfQ:1429878696275"/> Old Fox KM Journal

Wednesday, August 20, 2003

Feature Article:

Lessons Learned on Computer Simulations and Models

By Ken Long, Tortoise Capital Management

One of the US armed forces enduring strengths is its commitment to rigorous training. A key component of the military training strategy is the extensive use of simulations and models to test theories and ideas under a variety of operational conditions in a risk-free, efficient manner. Before we commit our troops and units to a strategy, you can be sure that we have examined the situation and tested our plans hundreds of times. Furthermore we have rehearsed our actions in simulators that seek to duplicate realistic conditions as much as possible.

Many traders can and do apply a similar concept in their approach to the markets. Extensive use of backtesting to find and understand tradable patterns is the most common. A newer phenomenon is the use of computer simulations to train and test the traders' reactions and decision making under a variety of market conditions. These are designed to make sure the trader can actually stick to the battle plan as designed. When done properly, these training simulations can make a powerful positive impact on your bottom line. Computers and simulations, however are not a panacea. Ultimately a trader has to enter the market and perform in real time.

The purpose of this article is to lightheartedly summarize some of our military's lessons learned with the use of training simulations and models, and highlight some of their limitations so that you can be better informed on the use of these power tools.

Rule 1: Don't believe the 33rd order consequences of a 1st order model (Remember: "a grain of salt!"). Know the limitations of the model and what you can reasonably infer. Remember that models are simplifications of reality that allow us to study different aspects of our world. If you have a simple model (also called a "1st order model") you want to make sure that you don't try to make the model do more than it CAN do. You can try to force more analysis out of the model than it can reasonably provide! (33rd order effects!)

Rule 2: Don't extrapolate beyond the region of fit: (Remember: "Don't go off the deep end!") Know the limits of the environment you are trying to model. A model designed to simulate commodities may not be suitable for currencies or equities.

Rule 3: Don't apply the model until you understand the simplifying assumptions on which it is based and can test their applicability. (Remember: "Use only as directed"). All models require assumptions. This is only a weakness or a problem if you don't know what they are and then try to model and test things beyond the model's capacity.

Rule 4: Don't believe the model is reality. (Remember: "Don't eat the menu"). Another way of saying this is: the map is not the territory, it's simply a tool to help you accomplish your objective.

Rule 5: Don't distort reality to fit the model (Remember: "Wishing won't make it so!"). Don't fall in love with your model or simulation. Know when it's time to move on! The market has some powerful and painful ways of waking us up from our dreams.

Rule 6: Don't limit yourself to a single model: more than one may be useful for understanding different aspects of the same phenomenon. (Remember: "Legalize polygamy"). Of course, going to the other extreme can be just as bad, when you have too many models and indicators and strategies: namely paralysis by analysis.

Rule 7: Don't keep using a discredited model. (Remember: "Don't beat a dead horse"). Stay current, keep looking for new ideas and testing them. When you find something more robust, don't be afraid to let go and adopt the new strategy or new model.

Rule 8: Don't fall in love with your model. (Remember: "Narcissus").

Rule 9: Don't apply the terminology of Subject A to Subject B if it doesn't enrich either subject. (Remember: "New names for old"). Metaphors are such powerful belief concepts that inappropriate ones can take a long time to expunge.

Rule 10: Don't expect that by naming a demon you have destroyed him. (Remember: "Rumplestiltskin"). It takes action and effort to turn tested ideas and sound theories into habitual action.

Simulations and rehearsal tools can be a powerful addition to a traders' kit bag. They are certainly worth the time and effort it takes to use and apply. Don't lose sight of the fact that ultimately you have to go forward into the arena and compete in a live market that is never quite exactly the same as a computer model. Forewarned is forearmed!

Ken Long, a Lieutenant Colonel in the U.S. Army with a Masters Degree in System Development, is a professor of tactics and logistics at the Army's Command and General Staff College. He has developed the Tortoise Method of mutual fund switching, a trading system that takes about five minutes each week and that routinely outperforms the S&P 500 Index

Ken is a speaker at Dr. Tharp's upcoming Infinite Wealth course. He is a trader and writes a daily and weekly market assessment for mutual funds and exchange traded funds. He is a proud husband, dad, and a ju jitsu practitioner.

To learn more about his Tortoise method visit, www.tortoisecapital.com

No comments: