Thanks Maestro and all the rest. Just like rewtree3 said about Alan, I also was only testing to see if people on ET are stupid or not. There's a lot of misleading stuff on ET. But there's also a lot of great material. Not easy sifting the wheat from the weeds, is it? Of course, Maestro, people have repeatedly proved most things to be impossible until they became common practice. And, historically, other beliefs that every human on the planet clung to religiously turned out to be ridiculously false. So is diversifying strategies possible or isn't it? If someone did discover how to do it, would they be stupid enough to admit it openly? Or would they keep that secret to themselves and pretend it was true to keep the masses off the right track? Sincerely, Wayne P.S. My last post in this thread. Thanks everyone.
all due respect, but is this really what this man: http://en.wikipedia.org/wiki/William_Ross_Ashby proved? are you claiming that if i have a strategy A and i add strategy B the risk/return features of the portfolio AB is independent from the correlation between A and B?
In his work "introduction to cybernetics" Ashby has discovered that risk/return does not depend on how much strategies are correlated.
Think about it. By using more correlated strategies your risk is going down so is your return. It keeps the risk/return ratio stable.
Here's the link to the full text of "introduction to cybernetics". See for yourself if it disproves anything related to trading. http://pespmc1.vub.ac.be/books/IntroCyb.pdf
The book does not have anything to do with trading at all! Theoretical conclusions of that book, however, have immensely powerful applications in trading! It takes a little bit of work and imagination to make parallels, but it pays off!
The point you make Maestro is that the strategies used must be uncorrelated. One obvious way to do that is run 2 strategies each on separate uncorrelated markets. However, even then, it's important to measure the correlation between markets because all financial markets correlate to some degree due to the world wide economy. So correlation can be measured as a percentage. Obviously, it's important to combine strategies in a portfolio with very low correlation to each other. Some holds true for combining strategies on a single market. Obviously there some correlation will exists but it's critical to use strategies that have very low correlation to each other. Otherwise, you're right, combining strategies on one market will fail. Sincerely, Wayne
No, the point I am making is IT DOES NOT MATTER . You cannot change risk reward ratio by mixing up any strategies unless one of them has inherently higher ratio.
Wouldn't you say that 5 positive expectancy, uncorrelated strategies running together are likely to create a smoother return stream than 1 strategy run with 5 units of risk? With respect, does your point require a definition of risk reward to be made?