question about a robot

New comment
 

I wrote a robot which performs good only on GBP/USD

but fails miserably on all other pairs

Is that the way it should be ??

 

it uses constant values for T.P & S.L and tested on 1H chart from 10/01/2001 - 10/01/2007 total trades about 800

 
sergeyrar:

I wrote a robot which performs good only on GBP/USD

but fails miserably on all other pairs

Is that the way it should be ??


most of the times, yes... but still, backtesting in MT4 can be missleading, if you have used that. I have noticed that the past data are not correct. If you let run the robot for some time that would be a better check I think.

another thing is how specific is your robot. If it includes indicators crossings it might only work on a specific pair and for only a short period of time.

cheers, P

 
Perica911:


most of the times, yes... but still, backtesting in MT4 can be missleading, if you have used that. I have noticed that the past data are not correct. If you let run the robot for some time that would be a better check I think.

another thing is how specific is your robot. If it includes indicators crossings it might only work on a specific pair and for only a short period of time.

cheers, P


I backtested it for 7 years.... not so short time period... :)

plus I used 90% modelling quality for the test... so its not the issue here

 
sergeyrar:


I backtested it for 7 years.... not so short time period... :)

plus I used 90% modelling quality for the test... so its not the issue here


Did you backtest with MT4 backtester? on a demo account?

if both anwsers are no then you might have something good, but still, non-judgemental trading sistems die with time. market takes care of them, believe me;)

for example see the automated trading chempionship, people there get a certin period when their robot should work and since they know that period they make a robot that actually makes wierd trades but wins competitions, because its easy to make something for the past, but eventualy it will stop working

 
Perica911:


Did you backtest with MT4 backtester? on a demo account?

if both anwsers are no then you might have something good, but still, non-judgemental trading sistems die with time. market takes care of them, believe me;)

for example see the automated trading chempionship, people there get a certin period when their robot should work and since they know that period they make a robot that actually makes wierd trades but wins competitions, because its easy to make something for the past, but eventualy it will stop working


I agree on that one... however my robot was backtested on data of 7 years and made lots of trades!!

it had 60-70 losing trades in a row... but yet with only minor drawdown and profit factor of 1.6

 
sergeyrar:


I agree on that one... however my robot was backtested on data of 7 years and made lots of trades!!

it had 60-70 losing trades in a row... but yet with only minor drawdown and profit factor of 1.6


so you have tried it on other pairs and it simply does not work?

well if you want to keep this discussion going you will have to tell us abit how it works... at least the basic logic, does it involves lots of indicators, structure of previos bars/combination of bars, etc....

 
Perica911:


so you have tried it on other pairs and it simply does not work?

well if you want to keep this discussion going you will have to tell us abit how it works... at least the basic logic, does it involves lots of indicators, structure of previos bars/combination of bars, etc....


it is based on breakout of 300sma (1H-chart)

if it breaks above this sma by 9 pips I buy with TP of 260 and SL of 20

same thing to the other direction

plus I applied some noise filter of 40 pips to filter ripples on that sma....

i did some optimization on 2001-2002 period and got some results

after that I took only the best and ran manually through 2002-2003, 2003-2004, .. 2006-2007

and left with only one consistently profitable setup ( actually two, which I combined into 1 EA )

 

i would say that each currency pair got their own characteristic

like USDCHF and EURGBP are know to be scalping pairs.

 
doshur:

i would say that each currency pair got their own characteristic

like USDCHF and EURGBP are know to be scalping pairs.


if it was facebook i'd put like on that one

thanks

 

Automated forex trading is a complicated field with lots to learn. The robots are not the worst help for traders. Trading systems like this one annaforex(dot)com operate in time frames measured in fractions of a second. It is this extreme speed of decision-making which also makes manually trading such strategies impractical. There are instances wherein the trader is not at his desk and the opportunity suddenly presents itself.

New comment
Reason: