hey, it’s Ok if you dont want to tell me. but i think you need to reconsider your entries a little bit as we all know in trading even just 5 minutes difference in entries could mean huge difference in results
here i gave you a screenshot as to why it looks strange to me.
The price has crossed SMA 60 way earlier than the entry (look at the 4th order which i selected in the trading history section). It says the EA opened the order at 6:08 while it should have made an entry already on 5:50 or 5:55 and locked in profit instead of getting away with loss. If the EA was according to the rules then it should have made profit already instead of losing. I see potential in this EA but the winning rate can be drastically improved imo if it’s strict to the rules and ignore missing setup altogether (for example, i have connection problem at 6:00 then i reconnected at 6:10, it shouldnt open a trade. It should just ignore it). That way im sure your EA will be more profitable
Based on the manual method of trading, this is a perfect setup. I would strongly advise you re-read the 3 Ducks e-book that Captain Currency has produced or, like myself go on his 1-to-1 course that gives an excellent insight into the strategy and introduces better trading criteria which includes methods for better entries.
ok sorry i didnt read his method for the whole pages and honestly didnt even know he has a PDF version. I only read his first page.
btw my IBFX MT4 hanged maybe because i opened too many tabs and loaded custom indies there (in one of the chart tab). I ever experienced this before 2 times and i had to reinstall the application (reopening it or restarting computer wouldnt help. not sure what kind of bug is that, can’t find the same issue on Google). Now everything is OK again but i didnt write my demo account password so i have to recreate new account. Will open a new myfxbook account tonight and re-setup everything. Now i have to go.
and btw you saw 2 lost trades there all of sudden lol because i added EURJPY here and instead of opening TP/SL of 30 pips it was just 3 pips and i got 2 losses instantly lol
Guys I notice the results are not as expected and having written a 3 Ducks EA and quite a few others can I offer some suggestions.
Rsg10z mentioned false entries. I believe you are using pending orders for entry and this makes perfect sense as you get a better fill at the entry price. However, this is also prone to spikes and bar formation. What I mean by bar formation is in my earlier notes, that I posted.
You cant use a Close[0] to enter any trades as you will get a false trigger. The reason being that the first Close[0] of any bar becomes the Open[0], the highest Close[0] becomes the High[0], the lowest Close[0] becomes the Low[0] and the last Close[0] becomes the actual Close[0] five minutes after the first Close[0] on a five minute chart. Either use the Close[1] or Open[0] as most bar open values are very close to the last bar close.
Also the 3 Ducks system is a trending system and a breakout system and therefore both system types have a poor overall win/lose ratio of less than 50% so your reward/risk ratio has to be better than 2:1. So your take profit needs to be at least double your stop. In reality you need to let your profits run as there will be lots of small losses. If your stop is the same as your take profit then you have to win more than 50% of the time and this is not possible with a trending system you have to let your profits run.
I did try to backtest your system on different pairs and thus try different stop values but it would not backtest, something you have done I guess. I suggest you consider the above and do some back testing.
I found market volatility was the main cause of lost trades and was looking for some sort of volatility filter, better trend filter or a stop based on ATR. Maybe the 60 SMA - ATR as a trailing stop with no profit target or trail the stop on the higher lows if long with no profit target.
i deleted my demo account because it was a mess (had an entry without TP and SL). But yes I was losing too back then.
I have been busy in the past one week because of real life business as well as testing my own EA (which is basically based on my manual trades, i mentioned it few pages back here). I hired a professional programmer to code that EA which btw im currently still testing.
I agree with Trader009 about this method. This method need more tweakings.
My suggestion is you code an indicator or combination of indicators first so that you can see if your theory works as you expect in principle on the charts, then use the same coding for your EA. Then you can check your backtest results against your indicator to verify everything is working as expected.
In this way you will then be able to analyse those periods that work and those that don’t and make adjustments accordingly to improve the overall results.
What you are looking for is an area of settings that give consistent results and a high number of opportunities to trade. Then you need position sizing to get the most out of the trade. Cut your losses short and let your profits run.
hi Trader009, i didnt do backtest for my own EA. I do forward testing. My method unfortunately cant be backtested because one of few indicators im using (zigzag) is a repainting indicator and the method rely heavily on higher timeframe trend. you only look at buy signal when the higher timeframe is currently bullish or sell signal when the higher timeframe is currently bearish hence why backtesting using strategy tester or any other automaton method won’t work (because you wouldn’t know at that time what happened at the higher timeframe. also i dont think zigzag can be backtested because it repaints when price breaks its high/low). Also i have custom indicator that’s needed for the method which is about pivot and S/R lines and it also repaint every 24 hours so I’m not sure how to backtest that.
I do forward testing and so far it works good as expected. I just found several bugs (wrong SL placement on one particular trade and one more bizarre situation where the EA didn’t pick the trade in pepperstone demo while it did in IBFX demo. Both should have opened the trade)
about the trend recognition, I do it manually with chart analysis (higher low, lower high, that kind of thing, combined with heiken ashi and price action). basically I set my own EA with trend filter label. so when I write BUY it would only look for buy signal and when i write SELL it would only look for sell signal. this EA won’t be 100% automated. we still need to check the H4 timeframe at least twice a day (because I only trade from london open to london close). Btw, my EA is just to ease my own trades. so it’s actually irrelevant to this thread.
but still thanks for your suggestion. Im looking forward to learn from all senior traders
and if aidysproule or whoever want to continue tweaking 3 duck based EA, you guys can let me know. I would just put the EA in my demo account and let it run. I dont mind backtesting for you guys.
Thanks for the feedback, it sounds pretty complicated.
All the backtesting I have tried I always get better results on EURJPY and USDJPY as these pairs seem to trend better, just a thought.
Not relevant to this thread, but I have spent all day completing and backtesting an EA based on the SuperTrend indicator (Robinson 2008) and again the JPY pairs return the best results. Hope to trade this live tomorrow on a micro account.
I did try and send you a private message to say I could help if you need it.
I now have my own none 3 Ducks EA working live in the market so have some spare time. Writing my own EA has given me some very interesting insights, like only trading on the open and only once per bar, also using the SuperTrend Indicator as a trailing stop and only moving the stop once a minimum amount of movement in the intended direction has taken place. So get in touch if you need some help or a different point of view.
I seriously recommend the book “Expert Advisor Programming” by Andrew R Young to anyone planning on creating their own EA’s.
I find many Myfxbook links that are ‘dead’ (“Sorry, but this system was deleted by it’s creator.”) and the links posted on this thread are no exception… Why not just keep these Myfxbooks active - it doesn’t cost anything, right???
