Wow - very impressive - broke the 1,000,000 mark.
|Thread Tools||Search this Thread|
DislikedI,m trying to copy these to my desktop to load. For some reason it won't copy as the yellow yeild sign. I have copied many before. Are these files good? any ideas?Ignored
DislikedThe EA is optimised for Eur/Jpy. Please try on 1Hr time frame with my settings as mentioned in the first post. Or you can try your own settings.Ignored
DislikedOk Aqib, this optimisation apparently works for EURJPY, but since we have no further info from you on the algorithm/methodology I cannot really say if it is luck with that pair or not. For example, minor changes (0.5)to the buy level will take you from 22% drawdown to 85% drawdown. Also, the number of trades over 3 years was 150 using the H1 bars as you said. That is extremely low and I personally would only use an indicator for this kind of frequency, not an EA to trade on my behalf.
Please provide either a description of what you are doing, what these "buy" and "sell" levels mean or at least alternative settings to use with other pairs.
P.S: I am a computer scientist as well, so don't worry about overly technical descriptionsIgnored
DislikedI have the same problem. It is not being drawn. EA works but results are quite terrible I am afraid with 90% accurate data on H1 charts GBPJPY and EURUSD.
In short, maximal drawdown was 90.% and profitable trades (total) was 13%. Default values used (risk = 10). Will try EURJPY as well, although /JPY crosses have similar action.Ignored
DislikedYou are getting low trades because only trades with high probability are executed. The buy level is that you are taking trades with 0.7 (70%) are likely to be profitable.
This EA is designed for eur/jpy and is no luck it is performing good.Ignored
DislikedThanks for the reply,
But the % of profitable trades is unfortunately less than 50% even with the EURJPY alone. The number of consecutive losses is more than twice that of wins..And when you increase the Buy_level to .8 it goes down, and drawdown even goes as far as 84% for .85!! Maybe the algorithm needs some polishing in this estimation business?
Also, if changing the parameters (stop loss..etc) is not sufficient to adapt the algorithm to GBPJPY, which is almost IDENTICAL in price action/movements, then your AI is definitely very complicated indeed. I notice also that the EA is most profitable in 2006-7, and after seeing/wrinting hundreds of EA's I can tell you this usually means that this period's curve fit much better the others, which is very dangerous.
Anyway please let us know if you decide to share the indicator or the algorithm, and good luck!Ignored
DislikedDo realize though that evolutionary algorithms are used to find optimal parameters for an algorithm. It basically achieves the same results as brute forcing, mostly just a little worse, but much faster. That's the only thing evolutionary (genetic) algorithms mean - optimization of values.
So basically this is curve-fitting. If it's a good system, that doesn't matter, but it is curve-fitting.Ignored
Dislikedi have just started forward testing and it is giving very good results. My backtest results have been very good too. But I dont rely on backtest results. You can backtest yourelf to get an idea.Ignored
DislikedSo I'm forward testing this EA (default settings) on a demo account. It took its first trade yesterday @ 18:00 GMT long on the EUR/JPY. It bought 11 lots @ 163.06 with a S/L @ 162.56. The trade is currently still open reached a high of 163.28 (05/01/07 03:00GMT) and is now sitting at a loss @ 162.87. I think an optimization with a T/P would be useful. Any plans for that in the future aqibsherwani? I know you're already planning on a trailing stop feature which would be nice.Ignored
DislikedThanks for the work you have done on the indicator and EA. I have dedicated a computer to running only your indicator and ea. Since Sunday I am up $437. on my demo account using GBP/JPY. I have it set also for EUR/GBP and USD/JPY but it hasn't taken a trade on either as of yet. I have my buy level set to .5 and risk at 20, everything else is set at default. Do you plan on allowing users to see the indicator at work any time in the future? Thanks again.
DislikedOk, so im trying out the EA and it initiated a buy a couple of hours ago.
The funny thing is that Take Profit is set to zero. This can't be normal can it?
See attached screen shot
|Thread Tools||Search this Thread|