Have anyone managed to make a continiously profitable ea? - page 6

 

On another note, I want to chime in on intuition, learning and coding.

As a student of pedagogy and neurology, and as a student of trading and programming, yes, I believe that trade decisions are made by intuition, and that can serve you because your unconscious can process some pretty complex data a lot quicker than your "conscious" mind can.

And, I absolutely agree with the necessity to make explicit these "rules" in order to include them in your program (unless you have more pragmatic expectations of your program and would be happy with a "simple" automated strategy with lower returns than a pro manual trader could theoretically get). And, making these rules explicit is important in order to test and modify them with intention.

However, I noticed a couple people brought up two different and equally valid ideas about how intuition is formed, how unconscious knowledge is acquired.

One the one hand, much is learned explicitly that then recedes to a type of intuitive knowledge that is hard to express, though it once was easier to express when you did not "know" it as well. A lot of times this type of "rational" learning can happen just at the surface of our awareness so that we are not totally able to express what we are learning even while learning, much less later once we've "mastered" it. Think about "acquiring" your first language as opposed to "learning" a new language as an adult. This type of knowledge acquisition sort of blurs the lines between "conscious" learning and "unconscious" learning, which I will address next.

On the other hand, our emotional centers aid in an unconscious learning that produces intuitions, a type of knowledge that was never explicitly known. These centers can process much more data than our filtered conscious experience can handle, and these emotional centers can take a very complex pattern and "stamp" it with a feeling/impulse so that we react (irrationally) in a particular way in similar situations in the future.

In either case, whether we have learned a thing "consciously" or "unconsciously," that "knowledge" is not necessarily the truth, or it doesn't necessarily serve us. In the case of "conscious" learning, someone might have learned the earth is flat. In the case of "unconscious" learning, person may have experienced trauma and have developed many responses to normal life that do not serve him or her.

From my favorite radio show, Radiolab, here are two podcasts that touch on this: http://www.radiolab.org/story/91642-overcome-by-emotion/ (on the pros and cons and inner-workings of gut feelings), and http://www.radiolab.org/story/91641-how-much-is-too-much/ (on the limits of our fragile "conscious" mind and how easily it gives way to impulse).

All that is to say that I try to leverage both my unconscious learning as a trader and my ability to make the implicit explicit for a program to do what I can't because I am just a silly great ape. I have spent a lot of time watching the market, simply watching it (either real time or by looking back over the history), with an obscene amount of indicators on the chart. I have spent a lot of time experiencing the consequences of trades, good and bad, live and demo, real time and manual back testing, according to rules and according to intuition. I believe this has built my intuition, and I continue to do it. I have also spent a lot of time trying to notice patterns consciously, which is sometimes a matter of consciously abstracting conscious principles and sometimes trying to uncover unconscious impulses. And, I have spent a lot of time trying to put those things into code.

I will continue to do all three, because I think it is the best use of my biological and technological capacities toward building EAs and trading practices. What kinds of practices do others engage in to develop their EAs? How is it working?

I am only now optimizing my first fully automated EA, so it is hard to say if I think what I am doing is working.

 
Does anyone have any names and %Percentage return history of these so-called Professional Manual Traders?
 

Kaleb. That is a fascinating post.

My feeling is that since it's impossible to eliminate our bad impulses/responses (or even to distinguish them from good ones in some cases), then manual trading becomes nigh on impossible to do continually profitably, and for those that manage it for any length of time are susceptible to becoming emotionally or mentally exhausted (or both), especially as the stakes rise. Therefore I see automated trading as the only option. I'm not even sure that one should try to express traders intuition/gut feel and work in to automated strategies for fear of working in bad impulses. That is to say, if it's possible to be sufficiently profitable without doing so, that would be the preferable situation IMO. That would be different however from trying to trawl our unconscious minds for repeatable patterns we may have seen/interpreted that we then code/test test in isolation to see if they have any statistical basis. That could be useful.

 

I think there are two slightly different (sub)topics here:

1. good requirements capture - when converting a manual strategy to a fully automated one. It's often the smallest details that provide 'the edge'. (assuming the EA is based on a manual strategy at all)

2. appreciation that even the simplest requirements may have non trivial implementations - eg a strategy based on H4 GMT, but trade entry must (only/not) occur while London and NY sessions are both open, and new broker is using GMT+2 (no DST). A simple requirement, with a messy but doable implementation, and good luck with backtesting (this is only non-trivial thanks to a completely brain dead way of timestamping quotes in MT4 combined with lack to tools to help),

 
Trevhib:

Kaleb. That is a fascinating post.

My feeling is that since it's impossible to eliminate our bad impulses/responses (or even to distinguish them from good ones in some cases), then manual trading becomes nigh on impossible to do continually profitably, and for those that manage it for any length of time are susceptible to becoming emotionally or mentally exhausted (or both), especially as the stakes rise. Therefore I see automated trading as the only option. I'm not even sure that one should try to express traders intuition/gut feel and work in to automated strategies for fear of working in bad impulses. That is to say, if it's possible to be sufficiently profitable without doing so, that would be the preferable situation IMO. That would be different however from trying to trawl our unconscious minds for repeatable patterns we may have seen/interpreted that we then code/test test in isolation to see if they have any statistical basis. That could be useful.


"for those that manage it for any length of time are susceptible to becoming emotionally or mentally exhausted (or both), especially as the stakes rise."

Talk to any Manhattan bartender, drug dealer or prostitute, and I think you'll find agreement, lol.

"I'm not even sure that one should try to express traders intuition/gut feel and work in to automated strategies for fear of working in bad impulses. That is to say, if it's possible to be sufficiently profitable without doing so, that would be the preferable situation IMO. That would be different however from trying to trawl our unconscious minds for repeatable patterns we may have seen/interpreted that we then code/test test in isolation to see if they have any statistical basis. That could be useful."

What you can express, you can test?

 
ubzen:
Does anyone have any names and %Percentage return history of these so-called Professional Manual Traders?


http://eareview.net/

I haven't looked at many of these, but this is what someone earlier in the thread referenced.

 
KalebC:

What you can express, you can test?


Well yes but is it worth testing everything you can express? That's probably a question that has different answers dependent on which direction you approach it from tbh.
 
KalebC: http://eareview.net/. I haven't looked at many of these, but this is what someone earlier in the thread referenced.
I was actually looking more for records of a manual trader with a couple of years trading history, but thanks for the link. I've viewed Brits_Review in the past and the list of robots seems to change quite a bit. While we're on the subject, I believe that metaTrader's Signals is a much more open process and resource.