2/16/2007

The search for profitable ideas

Recently a trader shared a day trading system based on event tracking, which had some success and in the course of our email exchange I said may be one needs to look at the market totally differently to build a suite of day trading automated system. Now this got me thinking. My primary focus has so far been on long term trading and so bulk of my understanding about market and research and efforts are focused on that.

To look at a day trading system is something unique and a creatively challenging process. So I decided to approach the problem on several level and will probably offer him a set of 5-8 ideas next week. It has been a very interesting exercises for me because it involved learning about an entirely different area of market and coming up with possible solutions.

To successfully trade the market and find enduring trading ideas requires creativity. Most of the trading systems which work are creative interpretation of market structures or creative workable interpretation of the market anomalies.

There are so many constrains when you are looking at high frequency short term automated system. The challenge is to find a solution which is enduring with a long term structural edge. The edge should withstand the markets ever changing cycles, plus it should exploit the available opportunity for the day to maximum possible extent. In the sense say if the possible opportunity for a given set of stock on a given day is say 1 million, one must extract the maximum possible out of it while managing risk.

Now in most peoples mind trading system development is mostly about data crunching or back testing. But much of the work is really concept development, the physical aspect of putting a system together if you have a profitable ideas is simple it involves entry, exit, and risk management.

I am approaching the problem differently and creatively. Lot of time we fail to find new ideas because we do not think creatively. Now I have some serious background in creative thinking because I spent over 6 years working in an intensely creative environment in the advertising business. In the advertising business your product is completely intangible. There is a constant pressure to come up with new ideas under strictly defined constrains. Plus most of good advertising ideas has to have enduring value to them. The automated day trading system is similar problem.

In my next post I will talk about the kind of crazy and whacky things I did to find possible solutions to this problem.

5 comments:

Unknown said...

A lot of the creativity in daytrading comes from simply observing the market for hours upon hours. Patterns will emerge and you will develop an intimacy with important market internal indicators that is impossible to learn explicitly by reading books or articles. You learn how different indicators interplay, especially certain market internal sentiment indicators and external indicators such as other markets.

One of my favorite relationships that I have discovered is the tendency for options market makers to lower the implied volatility premium when liquidity enters the market measured by internal and external breadth indicators.

To generate solid day trade ideas requires probably about 80% astute observation and only 20% rationalization.

Pradeep Bonde said...

That is a craft based approach to trading. It is not very suitable for automation and scaling.
There are hundreds of automated systems currently running in the market, they use very different logic and based on scientific principal and have bilions of dollars behind them.
High frequency finance is an entirely different field.

Unknown said...

I wouldn't necessarily call it "craft". My own methods involve using descriptive statistics to identify a potential trade idea and then using inferential statistics to test the robustness/validity.

The key is to identify significant patterns that actually translate to a structural or informational edge. Victor Niederhoffer and his traders test over 100 hypotheses/day based on their own sharp observations. Of course the majority get thrown out but the ones that pass the tests are obviously profitable. This idea of "throw everything at the wall and keep whatever sticks" is common among successful quant discretionary and system traders. It all starts with observations, not ideas.

Unknown said...

Brett Steenbarger, PhD on mechanical system creation, " Your best ideas will come from intensive observation, but some of the best ideas are the simplest and most straightforward."

I don't think trading is similar to advertising. Observation #1.

Pradeep Bonde said...

Brett Steenbarger, PhD is a PhD in psychology. Most of the stuff on his site is very basic stuff for building mechanical trading systems. It is more suitable for absolute beginners.

High frequency finance is completely different field. It uses different kind of models and the creativity invloved in it is different.