"Curve fitting!" will most of you say now, i know.
For a number of reasons I am experimenting with the possibilities and the limits of fitting simple and not so simple models to historic price action in an attempt to find at least something. If it doesn't immediately and completely fail after the fitted period, if the average profit factor stays above 1 for at least some time (determined after sufficieltly enough experiments), then the model may not be completely worthless and it may be worth investigating why it worked and trying to extract the essence of it's working part in a cleaner and even more powerful form.
I have read from a few people that they have EAs which they optimize on a regular basis, they claimed that they were profitable with this aproach, there is even some code somewhere on mql4 com for automating this process in mql, but i can't find any real empirical evidence with real numbers and diagrams, other than maybe someone saying in a chat something like "I optimize it every week, it is profitable".
Now here comes my question: Is there actually anybody here who has such a setup running sufficiently long enough to provide some useful evidence? Which type of strategy is it, how often must it be optimized and with how much history, how long has this method been profitable, how does the resulting overall profit curve look like?
For a number of reasons I am experimenting with the possibilities and the limits of fitting simple and not so simple models to historic price action in an attempt to find at least something. If it doesn't immediately and completely fail after the fitted period, if the average profit factor stays above 1 for at least some time (determined after sufficieltly enough experiments), then the model may not be completely worthless and it may be worth investigating why it worked and trying to extract the essence of it's working part in a cleaner and even more powerful form.
I have read from a few people that they have EAs which they optimize on a regular basis, they claimed that they were profitable with this aproach, there is even some code somewhere on mql4 com for automating this process in mql, but i can't find any real empirical evidence with real numbers and diagrams, other than maybe someone saying in a chat something like "I optimize it every week, it is profitable".
Now here comes my question: Is there actually anybody here who has such a setup running sufficiently long enough to provide some useful evidence? Which type of strategy is it, how often must it be optimized and with how much history, how long has this method been profitable, how does the resulting overall profit curve look like?