My firm is looking for another third party testing platform. A platform were we can:

A.)   Backtest a strategy

B.)    Optimize it. In other words, we have a strategy with a set of parameters and we want to see what combination of a set of variables within each parameter performed the best across a given set of stocks over a given time period. For example, one system we might want to test might be:

Enter long parameters:

-Stochastic (12 %K periods, 3 %K slowing) <= 20

-Rate of change over last 10 bars is <= -5%

-Current bar low is lower than the low of the last 5 bars

Exit long parameter:

-When 5% fixed trailing stop is hit.

But I’d want to optimize the numerical variables within each parameter to see what combination of numerical variables generates the best performance over say, the last 5 years across NYSE stocks. For example:

-Testing all stochastic <= X levels from 15-25

-Testing all stochastic %K periods from 6-18

-Testing all -x% rate of change values from -2% to -10% and testing those different rate of change values over each of the following previous number of bars: 5-15

-Testing all x% fixed trailing stop values from 3-8.

-Etc…

I realize that for the best performing optimization, “ideal” parameter variables might be outside the set of test variables I specify for each parameter. Clearly, I’d love a test that could find the best performing variable combination for me without me specifying a set of test variables for each parameter. But that might be a lot to ask for. If not, that’s even better to have.