Quote from fluttrader:
I had the same impression when he came in my thread "How to define a trend 100% objectively ?"
Some people like to impress others.
Jerry, please continue with your interesting posts,links
Ok, but if the group is serious in discovering something, lets actually do something concrete in contrast to intellectual posturing and tossing theories around. This is entertaining and sometimes educational but you can't deposit a theory in the bank. Profits on the other hand can be converted into a bank deposit.
So I'd propose a group test to see if a market has random or non-random behavior and hence tradable characteristics.
It might work like this:
1) Take a data file of say 10,000 bars.
2) Break it into segments: system development (5000 bars), Out of Sample (OOS) performance (5000 bars) and a future segment unknown to all participants
3) Each participants apply their standard trading methods to the test part of the file. Play with it, tweak you parameters or models.
4) When finished apply the same to the OOS
and post the results here
We all have ideas and methods, the best of which we don't share, with a few exceptions. So I'm not suggesting that we discuss how we get whatever results we end up posting, just the numbers.
People may choose to define what they do in general terms: chart patterns, random trades, NN models, modified Turtle Trader or whatever. Some general categories would give some nice information in that is the performance of the chartists is 50% better than any other group it may indicate something that I missed about charting.
If everyone looks at their results and the results of others and says "gee, that's pretty random" then the random theory of the market is supported. I'm sure someone here with great knowledge in that area will volunteer to run the confidence levels and tell us how random or non-random it is.
If on the other hand if 50% the participants show results with a Profit Factor better than the other 50% thought theoretically possible then the Non-Random hypotheses has a bit of support and bears further investigation.
I'm sure there will be some who will have a theoretical reason why a simple concrete experiment is a useless idea or a waste of time. To these I'd say talk is cheap and unimpressive, action and results speak louder than words. If necessary we can move the project to a different thread or a different forum to avoid the distraction of professional critics.
Any thoughts?
Jerry030
