Hi, i just have a simple hypothetical question, this is not a test of an actual strategy, i just want to know the method of testing to see if a strategy is beating the market. For example, lets say my strategy makes these point gains monthly on MSFT: Month Point Gain 1 2.55 2 -3.00 3 -0.50 4 1.32 5 3.33 Meanwhile lets say S&P 500 makes these point gains: Month Point Gain 1 23.55 2 15.00 3 -10.50 4 21.22 5 6.33 How would i know if my strategy is generating alpha? can i use CAPM to calculate it? For example, if i can get the beta for my strategy returns and calculate the req. rate of return using CAPM and if the actual returns of my strategy exceed that, can the excess be considered alpha? Can someone point me in the right direction? any help is appricated. shano

when you testing you strategy you must turn your mind on another parameter of strategy like Max DrawDown, profit factor, percent of winning trades and so on. And certainly you MUST check your strategy for the robustness. You may read about it in PARDO.

Just want to add that I think that Pardo's book is probably the best books iv read on systems testing and design. Highly recommened.

5 months is not nearly enough to show you are "beating the market." If you do not include reasonable statistics, then this whole exercise is meaningless

Thank you for the responses. However, I think I should be a bit clearer about what I am asking. I am aware that in the real world 5 periodic returns is not enough for any kind of analysis, I also know that there are other performance metrics and hypothesis tests that can be applied. But my question is unrelated to those and it is this: I can calculate the required rate of return for a stock using the CAPM model. With all its flaws aside, as long as I can give it a beta, an RF rate and a risk premium, it will give me a risk adjusted rate of return which I can compare to the actual returns of the stock to see if it has generated any positive alpha. Now my question is, can this method be applied to evaluate strategies? If so, then given the numbers I mentioned in the original post, how would one go about doing this? If this method cannot be applied to evaluating strategies then what would be the reason? Shano