Random rewards and hooked monkeys

Mark Douglas brings out some interesting concepts in Trading In The Zone. One of the things he talks about is effect of random rewards. If you teach a monkey to do a task and consistently reward it every time the task is done, the monkey quickly learns to associate a specific outcome with the efforts. If you stop rewarding it for doing the task, within a very short period of time the monkey will simply stop doing the task. It won’t waste its energy doing something that it has now learned it won't be rewarded for. However, the monkey's response to being cut off from the reward is very different if you start out on a purely random schedule, instead of a consistent one. When you stop offering the reward, there's no way the monkey can know that it will never be rewarded again for doing that task. Every time it was rewarded in the past, the reward came as a surprise. As a result, from the monkey's perspective, there's no reason to quit doing the task. The monkey keeps on doing the task, even without being rewarded for doing it. Some will continue indefinitely. I think it’s very true even of humans. And this problem gets even more exaggerated in trading.

Suppose you observe some setup. When you trade it, there is always 50% chance that it will go in your favour. Now, that’s random reward. When a trade goes in the money you never know if it was because of your analysis or it was pure randomness. What makes things even more difficult for a trader is the “experiment” (the trade in this case) can rarely be reproduced because of dynamic nature of markets. Markets never provide exactly same conditions. For example, for a currency pair it’s it extremely difficult to find exactly same conditions at two different times. Within a day, for example, a pair will behave differently at different hours. Market characteristics will change dramatically as major markets open and close. Next day, obviously, very different conditions will prevail.

How does one, then, know if a particular setup is indeed working and rewards are not random. One solution is to be extremely skeptical of any results. For example one may reject anything that gives less that 80% positive trades. Another option is to be very specific about market conditions where a setup is applied so that trade can be reproduced with little variation in conditions. No surprise then, I have seen some accomplished traders using these techniques. One very good trader recently told me that any trade that goes against you even slightly is bad trade. That’s some strictness! But that was not all. Oce he suggested a trade and went on to add that trade expires in 15 minutes- meaning if entry trade is not activated in next 15 minutes don’t take the trade. No wonder, he has been exceptionally successful in filtering what works and what doesn’t work in trading.

Desclaimer: http://indiangridlock.blogspot.com/2010/12/disclaimers-for-all-posts-in-this-blog.html