Thanks RT, I need some inspiration here. I believe you mean 'profitable' auto trading is possible. Really thought I was onto something to earn a living... will look for your thread for more inspiration/guidance.
Thanks to you too 5of7. I trade stocks only.. but those horrible backtesting results has thrown me for a loop. Will look for more threads regarding this..
I trade futures with an ATS and I have spent months upon months doing walk forward optimizations and tests on ideas before committing a single dollar to the venture. Sometimes what looks like a sure fire winner when tested backward plays out horribly and afterward you do the head smack and think "how could I have missed that" etc Anyone who doesn't check their strategy under "real world" situations is crazy .... in my view there is no substitute for it .... Just for the record I trade Eurodollar contracts.
I agree, all new strats are (and should be) forward live tested before they get fed more money. Trader 5of7 @ TheCollectiveFX.com
How do you measure the system's performance to decide when to turn it off/on? Some sort of moving average on the equity curve? Rolling Sharpe Ratio? Number of winning trades calcuation? I've tried some of these things with mixed results. Maybe I'm using too short or too long of a lookback in my moving average of the equity curve.
I already asked a similar question and he replied to it as follows (see quote below). His reply sounds reasonable but also very complicated and it seems like a lot of work in a completely new and challenging direction, and I don't want to get started on this unless it is really necessary and unless I am positive that it will work. Since "Strategy Filtering" is not recommended by everyone, for now I'll stick to the whole of my strategy and to its equity line, with its ups and downs. Besides, I feel that the first point is to make sure you only use strategies that do not show clear ups and downs (maybe by filtering its signals, based on range or other parameters). I mean, I already had to work on it for years as it is,. with just 2 or 3 parameters in my strategies. If I introduce the concept of strategy filtering, the work will double and I will have to work on it for another 2 years.
I'm just adding this link to remind myself that this poll is paired with this other one: http://www.elitetrader.com/vb/showthread.php?s=&threadid=161203
Grail is very much still around and was acquired by Tradestation in June 2010. http://www.thegrailsystem.com/