Judgmental “Heuristics” Or Biases and Developing Your Trading System: Part III By, Van K. Tharp

van tharp bkEditor’s Note: Dr. Van Tharp’s content is timeless and our goal is to continue to share his material with our readers. Today’s tip is material from a past article.

So far in this series, we have looked at biases regarding randomness, which is the tendency people have to seek out patterns where none exist and to invent the existence of unjustified causal relationships. Because people attempt to understand and make order out of the market, they assume that the longer a trend continues, the more likely it will suddenly turn around. This manifests into the “gamblers fallacy,” which is a very common trap that traders fall into and lose money when they do. In Part Two of this series, we examined data reliability as it relates to the degree to which information reflects what is really happening. We focused on representation bias, availability bias, anchoring bias, and hindsight bias.

Today, we continue with four common misuses of sampling variability in relation to system development and finish with some tips to help overcome all of these biases.

Sampling Variability. Most people misuse the basic concepts of sampling theory in making predictions and designing trading systems. The first principle, which is highly abused, is that you can make more accurate estimates of the true population probability from larger samples than from smaller samples. In other words, you can get a much more accurate estimate of the reliability of a trading signal from a large sample than from a small sample. In an example from earlier in the series, Jack said that his pattern predicted a higher market price 35% of the time. The accuracy of his estimate would be much better if it were based on 100 measures of the pattern than if it were based on 20 measures. Unfortunately, most people follow a bias called the law of small numbers. Once they observe a phenomenon occurring a few times, they believe they understand it and know its likelihood.

People tend to form their opinions based on a few cases and fail to revise their opinions upon the receipt of new data, to the extent that they should, based on probability theory. Traders tend to stick to their old opinions rather than updating them as new information becomes available.

We call this the conservatism bias. This points out the importance of doing thorough, objective testing of your market observations on a set of data that is different from the data in which you made the observation.

Traders want consistent information from various sources, such as three oscillators based upon the same data (which of course are likely to show similar results). However, this consistent information will lead to increased confidence, but not to increased accuracy of prediction.

We call this the consistency of information bias. What it means is that traders are likely to add more indicators in order to get more consistent information so they can feel confident about it. But adding more indicators is not likely to give a trader more accurate information. This points out the importance of developing a simple, robust trading system.

A fourth major misuse of sampling variability is that people fail to understand that the amount of variability in a sample is positively related to the degree of randomness in the sample. Once you have observed a relationship in a set of data, it is no longer random with respect to that relationship. The more relationships you observe with respect to various parameters in the data, the less random the data is with respect to those parameters. Unfortunately, system developers frequently make this mistake when they use a sample of data to optimize a system and then test the system on the same data. Once a set of data has been used to optimize a system’s parameters, then it is not random with respect to those parameters. As a result, when you use the same sample of data to test the system, you can expect it to do well in the test, but this has nothing to do with how it will work as a system trading real money. Data must be tested on a sample that is independent from the sample used to observe the original relationship.

How to overcome judgmental biases

You probably cannot totally overcome the effect of the various judgmental biases. One reason is that one of the most prevalent biases is the ego bias in which people decide, “Yes, I understand all of this, but it applies to other people, not to me. I’m a very special person and it doesn’t apply in my case.”  Nevertheless, if you are willing to assume you are human and that these biases do apply to you, then you can take steps to minimize their impact.

Remember, your job as a trader is to find an edge in the markets. You must capitalize on that edge, so you will make money in the long run, while doing everything possible to preserve your capital in the short run. As a result, I strongly recommend that you spend a lot of time writing down your objectives and designing something to meet those objectives.

What is an objective?

Your objective is your goal, your target. It is the thing that you want to attain or accomplish.

Objectives set the roadmap for the entire system development process. How would one know how to get someplace if they didn’t know where they were going first? It is easy enough to see that if one trader had an objective such as “I want a system that trades long-term stocks, that requires my attention only once each week and makes 20% per year” compared to a trader’s objective that was “I want to actively trade my mother’s retirement account for four hours each day, without holding overnight positions” then two completely different systems would be required. The objectives or goals are very different. There are endless configurations of objectives. The point is you need to specifically know what it is that you are trying to attain. And only then can you develop a trading system that will help you attain it.

Observe the markets as an artist would. Be creative. Determine relationships in the market that occur over a wide variety of markets and market conditions. Remember, you are not trying to explain the markets, but just determine some market relationships you can capitalize upon. The more widespread the relationship—does it occur in different markets and different types of markets—the more likely you will be able to profit from it.

Be willing to be unique. Think about how you can best represent the price of the market. Notice relationships in the patterns of price movement that you can capitalize upon. Once you have observed some relationships, figure out how to measure them. If you can avoid common indicators, then you probably have a real edge.

Simple is probably better. Why? Because the more complex the relationship, the more likely it is to be unique to particular markets and the less likely it is to make you money.

Make sure you understand the edge that the relationships you observe in your data give you. Do your observations make sense? How do they give you an edge? Also be sure that you can write down your observations in enough detail so you can recognize them as they occur and not just in hindsight.

Understand money management so you can capitalize on your observations. Trade according to a predetermined plan rather than your emotions.

Be sure to objectively test your observations on extensive market data that is different from the data you used to observe the relationships in the first place. Objective testing is very important because with subjective testing you will tend to see what you want to see. In other words, the market will confirm your expectations.

Many of the psychological issues described in this article are covered in our How to Develop a Winning Trading System that Fits You Workshop and Home Study program. These programs will help you clarify your objectives and then show you how to design a trading system to meet those objectives.

Leave a Comment

Your email address will not be published. Required fields are marked *

Shopping Cart
Scroll to Top