Topic: When to optimise a 15 min strategy? after a day, a week, a month?

The one million dollar question for me is, if I found a good M15 strategy (for me that means a good Sharpe Ratio, low DD), when is the best time to optimise it?
And, how is better to optimise it?

The idea is to survive in time with the strategy and for this, I think, is good to optimise it according to the market.
First of all there is the discussion about the " how many bars to use". If I optimise for the last 8 months, that means 20.000 bars. is it enough? is it too much?

then, is it good to optimise with "out of sample testing"? how much? 10%, 20%,30%?

Now, the example. let's say, we have a 15min strategy which is working good for the last 8 months, and it was also good for me for the last 2 weeks, now, is it good to optimise it for the last 4 months with 10% out of sample testing? 10% means aprox. 2 weeks...

many thanks for your help...

Re: When to optimise a 15 min strategy? after a day, a week, a month?

My thinking is that the more bars the better so as to cover whatever changes the market may be undergoing over a substantial period.  Then the ea will have a better chance of surviving the OOS period.

I use for a 5 minute chart 50000 bars and I start several months back, and then see what it does as I move it ahead one month at a time. As long as Sharpe ratio looks reasonable for each period, I keep it even though daily profits may fluctuate.

I am not sure what to do with OOS, I just use the 30 per cent and it seems to be working, I have no idea what is correct to do.

I find that if I start with a long enough period and get a high enough Sharpe Ratio to begin with that the robot will survive many periods into the future. Of course, the daily income is probably lower than if I was to use a shorter period.

Soon there will be a new version of the program which will make the System Quality Number available with each item in the Top Ten..... and that will be a good guide as to what we have to do, it will probably change how we regard our Ea's and cause us to adjust our criteria...