Hello how to develop the best robot - page 3

New comment
 
I think that to create good experts, I won't say the best, although that is always the goal, you need to have a lot of knowledge and experience with the programming language, its functionalities, and also code organization and cleanliness. You also need to have a lot of knowledge about the forex market, symbols and their particularities, price movements, correlations, and influencing factors. Putting all this together, I believe it is possible to create good experts.
 
just been working on some EA….got this back from back testing…really nice result…almost to good. will test real in lofe setting today. anyone see a usual big difference between test and real market? if so to what extend? 
 
Damiem Marchand De Campos #:
I think that to create good experts, I won't say the best, although that is always the goal, you need to have a lot of knowledge and experience with the programming language, its functionalities, and also code organization and cleanliness. You also need to have a lot of knowledge about the forex market, symbols and their particularities, price movements, correlations, and influencing factors. Putting all this together, I believe it is possible to create good experts.
i think with AI, we need a good strategy…the programming is now done by ai
 
Mark De Vries #:
anyone see a usual big difference between test and real market? if so to what extend?

It depends on the way the EA is coded and one the way it is back tested. Both should match. So if the coded logic is only applied on prices not including the current (0) bar, it allows for testing in 1 OHLC or even open prices. If not, real results will be different from a back test.

But even then, the tester has nuances and peculiarities. What you have found most likely is a tester grail, exploiting the OHLC tick generation mechanism. Before going live, see if testing on every tick based on real ticks yield the same results. 

 
try latency arbitrage
 
Heri Yusufu Kaniugu #:
try latency arbitrage

Be careful with that. Many broker-dealers prohibit triangular forex arbitrage.

As a possible alternative, you could indirectly arbitrage forex futures against spot forex of the same currencies... or indirectly hedge a front month futures contract against a later month futures contract of the same futures symbol--they are not the same instrument.

 
Imran Rajan #:

I think there is no best robot in the world

if there was the best robot, all of people went under support of this robot and then no-one was the looser in the forex

The best robot is the one with the best risk management. losses will always be there in when trading. Enhancing risk management into an expert advisor makes it highly profitable if your strategy is already good. Expert advisors can manage risk better than humans do because they don't have feelings and most of losses incured when trading are because of emotions and feelings. A well defined EA with the best strategy and risk management can make one highly profitable
 
Enrique Dangeroux #:

It depends on the way the EA is coded and one the way it is back tested. Both should match. So if the coded logic is only applied on prices not including the current (0) bar, it allows for testing in 1 OHLC or even open prices. If not, real results will be different from a back test.

But even then, the tester has nuances and peculiarities. What you have found most likely is a tester grail, exploiting the OHLC tick generation mechanism. Before going live, see if testing on every tick based on real ticks yield the same results. 

i gave it a few days....been running the EA on demo accounts Standaard and ECN with 4 different brokers...quite interesting to see the difference in result....already generated about 1000 trades....these markets in gold create a lot of activity...:).......also found indeed that the result posted is actually in the most pristine environment, so ammended that. In demo account, the performance is actually better with a certain broker, other brokers under perform but still perform....but ECN accounts are the prefered account for this EA. Working now on addons based on the current strategy and testing them as well....my trading knowledge and experience were always there...just did not know how to code....AI has now changed this completely....I might be posting the EA on market place in the near future, but only when I have actually seen real live data with live accounts...will also test several brokers in that instance as well.....currently with one broker we are already at a 50K profit since 25th of februari...but will keep the EA running for many more moons to come..


 

First, trade manually until you are fully convinced your strategy works. Only then try to implement everything in your code. There is no other way to build a profitable trading robot.

Your alo-trading skills improve only when you have a reference.

 
Wojciech Daniel Knoff #:

First, trade manually until you are fully convinced your strategy works. Only then try to implement everything in your code. There is no other way to build a profitable trading robot.

Your alo-trading skills improve only when you have a reference.

The fact that profitable strategies exists wich are impossible to trade manually disproves your statement. 
1234
New comment