Quote from cokezero:
I've found that thinking is more effective than observing. When you look at a chart the human mind is wired to find patterns that are not actually there. Also, what works is what you can't see with your eyes but instead have to see with your mind. When I say thinking I mean making an assumption based on my believe of how the market works and test it to see if there is any validity. When I run out of ideas I read. I've found that reading, no matter how useless the materials are, is a good trigger for new ideas. I've built a library consists of totally useless trading books but I still flip thru them once in a while just to get some stimulation for my mind. I also visit websites like ET for stimulation.
Whenever I think of an idea I write it down on my notebook. I've been able to generate a lot of ideas because I'm constantly thinking about the market. Ideas could come anytime and I make sure I write them down everytime so that I won't miss any. When I don't have my notebook with me I type it up on my PDA and later copy it to my notebook. With an abundant backlog of ideas I was able to consistently test about 3 to 4 ideas a day 5 to 6 days a week for the last 5 years. That amounts to about 5000 different ideas. In the beginning I test just about anything I could think of no matter how outrageous it might be and the quality of the test is bad and most of the time spent is pure waste. Overtime I've developed a toolbox of valid assumptions about the market and is able to narrow down the direction of tests and the quality of the test goes up drammatically. The quality of the test follows an exponential curve with completely crappy results for a very long time in the beginning. The best ideas are always my latest ideas as my experince grow.
Experiencing failure thousands of times is no fun and is very discouraging. To drag my ass back to testing and experiencing yet another failure is no easy task. I've found that thinking thru the idea throughoughly and run "simulations" in my mind before the actualy test helps a great deal. For one it would eliminate many ideas not worth testing on second thought and also thru the process it helps to generate even more ideas. More importantly with the mind simulation I was able to convince myself that the idea is a good one and would give me hope and get myself excited enough to get back to the computer and work. Of course most ideas would again proved to be useless. Sometimes it gets to a point that when I think of a good idea I don't dare to test it. I want to hang on to this "good idea" for a couple days so that I can feel better with some hope. I'm afraid once I test the idea it would turn out to be not so good afterall (99% of the time) and so goes the hope with it. Over time I've learned to just treat it as a job and test 3-4 ideas a day like a robot. I no longer hope or get excited or anything. I just test like a robot and goes next next next. It sped up my testing because the hope-excited-defeated cycle is a drag both emotionally and time wise. When I see valid results it goes to my toolbox for further development.
I think the process to develop an edge more or less follows this path. I might be just dumb and I think it could be possible to find something that is useful with maybe a few hundred ideas and do it in maybe a year or two. I would recommend against automatic pattern search and the like. The market is all about the assumptions you make and how close your assumptions are to reality. ATS wouldn't help you bulid a valid assumption and is probably counter productive in that regard. I think there is no substitute for manual research and most importantly lots and lots of thinking and seeing things with your mind.
Good luck!
