Just posting to maintain updates to this thread. I will be forward testing many of them.
Just posting to maintain updates to this thread. I will be forward testing many of them.
Howard:
Are you speaking in general terms or have you ever reconciled the backtested results with your guide analyzed results trade-by-trade?
If not, would you be willing to post your guide backtested results for October (pref. In Excel format) so that somebody can conduct a Strategy Tester backtest to compare against?
I believe that would help Aha and others hugely, especially if there was analysis provided regarding the difference. I would be ready to do the analysis with receipt of the guide backtested transaction logs, EA conduct backtests and obviously stipulated rules document.
To acquire all this stuff correlating correctly, details details and more details are essential! No surprise since the markets are lively!
Aha,
Thank you for your participation to this forum.
You have written many specialists.
Have you come across any that are profitable, and have been profitable within many years, with low/med risk?
Due
Hi FxBabe,
I haven't discovered a mechanical system that may be profitable in extended term. The market situation is constantly shifting, the more I code EAs for mechanical methods, the more I know the significance of trader's discretion. If you want to be successful in long term, I figure you need to enhance your trading skills and find your own system. I really don't think there is a shortcut here.... But what I understand, maybe I haven't seen enough.
There is a huge misunderstanding in the trading world regarding locating The System, profitable over a 100 years without ever having to be touched , once turned on. Unfortunately anyone who looks for such a system does not know that the market is a collection of waves that change in height and width with time. Sometimes the surf pounds smaller and harder and faster, and sometimes it's larger and slower and gentler.
Looking for a this type of system is like picking the most efficent vehicle to get from point A to point B(a journey across mountain ranges, rivers and plains) is a vehicle with racing suspension and tires, and one non equipment. Well of course it can get there eventually; the racing suspension and tires can allow it to onto the highways and the low equipment will keep the car going on the steep terain. However, we all know it makes much more sense to provide the car several gears, and a much more fexible suspension and a set of different tires.
The same is true to your own system. It's simple to think you'll find 1 group of configurations that can make your system decently profitable year in and year out to the ever-changing waves of the market. So instead of optimizing your system over 12, 24, 36 months, optimize it on 4,3, or perhaps even only 6 months. And re-optimize every couple of weeks to maintain your own body fitting the market waves most profitably. Your profits will soar while your drawdowns will collapse. I guarantee.
You will find elements of adaptive algorithm in my trading system.
I feel a lot of adaptive systems are functioning far and wide. Have you been given such a system you would be coining money. You would hardly talk profitable adaptive system.
But I think many people can construct flexible system themselves.
Nice input aicccia (lots of c's in that screen name!) . The notion of adapting one's method to every-changing market requirements is definitely sound. While I adopt that notion, I also believe in market cycles that at times could represent sudden changes from recent history. By way of example, the summertime make exhibit a specific market personality that abruptly changes when Autumn hits. Thus, if it had been the beginning of Autumn and I'm optimizing based on Summer information, it would seem that I would be out-of-sync using the Autumn cycle.
While I truly don't think things are this clear-cut, I really do believe in cycles (not always or just seasonal cycles). The question then becomes how do you go about reconciling this with the notion of easily evolving waves based on past xx weeks of history? Yes, one might consider waves just another cycle. But, if cyclical study shows that wave x ray is typically a small wave followed by wave y that's larger, then, certainly optimizing based on wave x wouldn't be optimal will be state.
Which are your thoughts on adapting the concept of waves with cycles?
Bill
Well said aicccia. I agree with you. That is why human discretion is important -- a good trader knows when to change equipment.
We consistently have the exact same notion here:--RRB-. That is why I stopped coding EAs and began looking at Neural Network solutions. The achievement of Phoenix from the EA championship demonstrates that a tuned up easy system may profit in short period, for specific pair. And I feel an NN system may do better.
I agree with you too, Bill. It's possible to style self-adaptive fine-tuned EAs, but considering the time and energy demands, it's barely doable. Most of us know market is shifting, but the number of parameters we need to use to characterize the change of market conditions? What kind of instrument we need to develop to accurately measure the shift of each parameter? Then what sorts of egy should be used at which market states? These questions could be Ph.D. thesis issues and likely worth a lot of years of research... much easier than finding the grail.