ATC.Experience, knowledge and practice. - page 7

New comment
 
Martin CHEguevara:
No neural network can show you where the market will go with more than 53% accuracy.
Funds and houses only focus on stocks and indices more than currencies. They stupidly put in a billion and make 10-15% a year and live well.
Forex is a thousand times more complicated.
You can make money on it alone, knowing the structure of any market. Of course nothing will be written about it anywhere. Of course to understand and calculate it, and to see it you need a brain.
But vision is deceptive - where the brain tells you to buy, the robot calculates the probabilities and says 50-50.)
In today's world, no brain can calculate more objectively than a robot.
Of course the brain is primary, but in terms of objectivity of vision the robot is primary.
Domestic robots are no different from industrial robots)
Differences are constructed only by those who believe that complexity is a guarantee of success. But that is not the case. The simpler and more perfect and
algorithm, the better and more reliable it is in the market.

brain-logic-algorithm buy/sell robot

 
all set... all transactions are +... yield 1200% per annum... I'll sell it for a lakh...
 
Vasiliy Ishenko:
...
My experience, knowledge and practice in the field, at the moment I have to trade manually, analyze by myself, "with my own head" + use a couple of exclusive indicators of my own design. Yes, I use a couple of robots (they are even more like scripts), which make decisions about opening/closing at levels or points specified by me personally in advance, but nothing more than that. But it is not possible to write an algorithm using my "manual strategy", because it involves figurative perception and complex graphic analysis.

Before I started writing robots myself I was pondering over this question - is it possible to formalize trading systems where a trader mostly relies on experience and subjective perception of the market?

So far I've come to the following conclusions:

1. Of course - any system can be programmed, if it has clear rules.

2. If the rules are not clear - you can analyze factors that affect interpretation of the rules, and also program it using fuzzy logic and game theory methods.

3. There will always be some advantage for a human, because from time to time "new" situations emerge that were not taken into account whenwriting the robot, and which a human will always see, while a robot most likely can "view" them (this also applies to super-duper "all-inclusive" neural networks, they only slightly compensate for this if there is constant additional training). The robot naturally has its own advantage - which is the greater, the more frequent-trading the system is.

4. Often it does not make sense to program a system even for very successful manual traders - because it is adaptive and is essentially a conglomeration of many simple systems and a superstructure of the decision-making system with such parameters as flair, luck and even courage :) in this sense, as already said here, it makes sense to work with a portfolio of many simple systems, properly configured, it is easier and statistically analyzable, unlike manual trading.

And yes, I have not yet lost interest and I am programming "complex graphical analysis" and I believe there is nothing unusually complicated there, if you really program it)

Vasiliy Ishenko:
...
But my profitability is +30% a month, but it's not just 30% a month, but every month(!) from year to year(!) with no exceptions! Is it possible for a robot to do this? I don't exclude the possibility, but I also don't exclude the impossibility (excuse the tautology) and therefore I don't do any more research or development in this field.

i wonder why the profitability is so high) if my monthly profitability is stable i'd be happy with 3-5%. i'd rather have a 50% slippage and i need more, but it's also very dispersive.

I adhere to the opinion that there is a mathematically valid limit to the profitability, which depends on the "fit" of the system, i.e. the less the system is adjusted to a specific market, situation, period, etc. (I haven't found any scientific research about it, it would be interesting. Perhaps that is why the biggest funds and trading departments of banks do not show space profits.

1234567
New comment