hello, currently i'm developing an EA based on parabolic SAR
the rule is simple,
open a buy position when Parabolic Sar dot lower than current price, hold the posision until the trend reserve (the dot going to higher than current price) & vice versa
i found a big problem for my EA.....
my EA runs very well on the hi-movement price, such as gbpusd last week. But it will become a big problem if the price just ranging (lazy to fly / fall)
i've combined the ParSAR with day_high & day_low to protect my EA for opening position when the chart didn't change too much...
then i put also at my EA, a second ParSAR, you can combine ParSAR step 0,01 with 0,02 for closing a position.
here is information for my EA
step : step for ParSAR #1
max : max for ParSAR #1
step2 : step for ParSAR #2
max2 : max for ParSAR #2
ranging : if day_hi - day_lo > ranging, a position will be opened.
trev : reversal tolerance. if the ParSAR reversed more than (trev) times, the EA will wait for NY market (13.00 GMT) opened or next day Euro market opened (5 GMT) before taking a new position.
false means, all position will be closed at 01:00 GMT
this EA can modify any position, including position that not opened by it.
standar options :
january 2006 : (1jan-14jan) : 318 pips
you can download the file for back & forward testing until 31 Jan 06.
this EA is for gbpusd 1H
tested at Northfinance MT4
Why an EX4 file?
A lot of testers shy away from EX4 for various reasons. IMHO you'll get more feedback if you post the source MQ4 file.
As for myself,no matter how good an EA looks or sound I would not test it if it's in an EX4 format. Furthermore, testing an EA for only 15 days is kind of useless...2 months min before I start to trust it.
It's up to you.
PS: BTW if this EA is going to be commercial, you are posting in the wrong section of this forum. the forum's commercial section is at:
I agree with sadaloma. Since MQL has filesystem-capabilities, you can actuall use to plant a virus/trojan or to spy out sensitive information.
No one in his right mind should execute *.ex4 files, which are not self-compiled or out of a REALLY trustworthy source.
Oh and ... PSAR sucks for Long/Short Open/Close Signals. They hav some signficance for a dynamic StopLoss, but that is about it.
Really look at one week of trading with PSAR. Yes, the big movements you will get (like with 90% of all the other indicators), but you will burn the money in a market which is not trending like crazy.
Parabolic is indeed a great predictor of turning points
I use a custom redesigned parabolic sar i call Kasa to predict turning points. I basically was able to stay clear of most noises, and take signals during active trading hours when the market is likely to make a sustained move. its also great for the indexes
I hope to develope this into EA and this EA could have been a perfect starting point if the owner shared source codes -
See attached graph for kasa turning points
i just put ex4 last week because the code is very ..... unreadable...
now i have put some part to precedure & function, make us easy to read the source-code. And i've uploaded it.
@safari, how about if we join your custom parsar with my EA
it will be a great
btw, i have deleted the second parsar, because it is useless...
can you please post kasa indicator??????
i have recode my EA, so it will more unreadable..... (i think, it is very clear)
it uses parsar as the main indicator, and only can open a position at 7 GMT -11 GMT and 15 GMT - 19 GMT. the rest, EA can only close a position.
EurUsd M30, please use .set configuration for Expert Properties.
backtest result :
1 january 2006 - 20 january 2006 : 312 pips
dec 05 : 340 pips
please help me to forward-test my EA
welcome to any suggestion for improvement
Thanks for the EA.
I will forward test this EA by tomorrow, will let you know the results the next week.
please notice that my EA have some bugs...
i gotta fix it soon....