Hey folks. I'd like to get some input on the topic of optimization and what others have experienced when going live afterwards.
My thoughts on time frame, is that there is no point in backtesting what would have worked for the past 10 years. I think we are more interested in what has been working for the 3 months, maybe? On the other hand, its probably not best to optimize only the last week, and then think that that will produce similar results. It wouldn't be best to optimize over some crazy week with wild swings, and then have those parameters working for you next week when it could all be quiet for all you know.
So for those who have their edge, who are trading it, are you say, optimizing the last 3 months maybe, BUT doing it over every weekend, or 2? That way your system is always adapting to a changing market, yet not too sensitively. Just a tweak here or there in the end.
Also, how else do you plan. I will assume, that forward results rarely exceed expectations, but maybe they can. So do you also plan, for example, to have enough money in your account to withstand double, or 3x drawdown, or 2x maximum consecutive losers?
Am I on the right track? What do you do? It would be great to hear some feedback on what is working for others.
Of course I can backtest and optimize, then forward test in demo mode on my own, and I will, but I'd just like to hear what others out there are doing too.
best regards,
SP.
My thoughts on time frame, is that there is no point in backtesting what would have worked for the past 10 years. I think we are more interested in what has been working for the 3 months, maybe? On the other hand, its probably not best to optimize only the last week, and then think that that will produce similar results. It wouldn't be best to optimize over some crazy week with wild swings, and then have those parameters working for you next week when it could all be quiet for all you know.
So for those who have their edge, who are trading it, are you say, optimizing the last 3 months maybe, BUT doing it over every weekend, or 2? That way your system is always adapting to a changing market, yet not too sensitively. Just a tweak here or there in the end.
Also, how else do you plan. I will assume, that forward results rarely exceed expectations, but maybe they can. So do you also plan, for example, to have enough money in your account to withstand double, or 3x drawdown, or 2x maximum consecutive losers?
Am I on the right track? What do you do? It would be great to hear some feedback on what is working for others.
Of course I can backtest and optimize, then forward test in demo mode on my own, and I will, but I'd just like to hear what others out there are doing too.
best regards,
SP.