If you really have a method that works for one market, would you use it to other market without any change? I remember Tom DeMark said his methods could work for any market. But from my own experience there is simply not true.
I personally haven't found this to be true. However, I do know of someone who moved from currencies to stock options but he is a marketmaker with Darkstar Trading. Of course, we know the advantages of working on the floor using different strategies from most of us.
IMHO markets have character and different volatility levels as a % of their values. I think ideas will work over a broad array of markets but the inputs will most likely vary depending on the characteristics.
Yes in fact if your method doesn't work for any market then it is suspicious of overoptimisation. A method must be robust. Mine has been applied to french cac40 and I don't have to change a iota for Dow Jones. I don't even have to optimise parameters. It depends of course on what type of change you mean for what type of model. If it is a stochastic model, it will sure depends on volatility and so you have to change some parameters VALUES but not parameters NUMBERS if not justified ( Mine is not a stochastic but generally it is stochastic) so what you musn't change is the Structure of the model. If not so it can be suspicious. But when changing values you musn't change every time to fit the curve if so it's curve fitting by definition so useless
I only trade es. I think more people who have successful systems in other markets should come and use those same systems on es. I don't know if they would make money, but it sure would make my job a lot easier.
the more markets work with your system, the better. the fewer changes you have to apply to make it run on another market the better. that is what this topic is all about IMO. now that does not mean at all that you could not have a system that works on one market only. basically the trader universe can be divided into two parts: those who come from watching specific markets and those who come from somehow statistical approaches. the first will always claim that coffee and emini are very different pairs of shoes, the second will point out that it is in essence just time series with special features, nothing more. no reason in arguing about it. both are right if people make money with it. and this is possible with both points of view. peace
That's good. I use both systems which depend on randomness and systems which depend on the lack of randomness. Like you said, if they both work...