Help Needed to Debug Existing EA with proven strategy | Page 16 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Help to debug my EA 2 replies

Developer Needed To Debug my EA 2 replies

MT4 EA coder required to fine-tune/debug existing EA 0 replies

Debug code in strategy tester 1 reply

debug message MT4 EA in Strategy Test 7 replies

Attachments: Help Needed to Debug Existing EA with proven strategy
Exit Attachments

Help Needed to Debug Existing EA with proven strategy

  • #301
  • Apr 9, 2018 7:15pm Apr 9, 2018 7:15pm
Because of the Stoch filter you can optimize the depth settings and produce some amazing results (yes the older 1.6). I attached a screenshot of a rough backtest of that.

So now I'm wondering if we can run this in the hybrid setup (if we add stoch). I am still trying to understand the strategy in its entirety, so bare with me on that part.
Attached Image (click to enlarge)
Click to Enlarge Name: Screenshot2.png Size: 60 KB
  • #302
  • Apr 9, 2018 7:15pm Apr 9, 2018 7:15pm
Yes Verocity, I would like dearly to see a MOMENTUM filter be incorporated in the hybrid mode so that the EA would only trade when certain level of momentum is present as theoretically it should reduce the incidence of deadlock and should improve on the overall profitability of the hybrid mode.

Can all the gurus here share your idea first on how to effectively identify momentum before we incorporate it in?? I mean share with us your experience in this aspect - eg which indicator, its settings, which time frame more reliable and how do you use it etc etc etc...

I personally believe that is the right direction to move forward to at this juncture - what do you guys think??? Any other ideas on how to improve the profitability and avoid getting trapped in a deadlock situation???

Common guys, share your ideas here & give your feed back on my suggestion above.

Cheers guys!
  • #303
  • Apr 9, 2018 7:21pm Apr 9, 2018 7:21pm
My thought is a CCI filter added, I prefer it personally. It may just take trying a few top choices to see which is best.
  • #304
  • Apr 9, 2018 7:25pm Apr 9, 2018 7:25pm
Hi Verocity,

Yes, the back test results you posted in post 301 is VERY impressive, especially when the period you used covered the BREXIT period - but Verocity, try to run the Seller on the same period, I am sure it will not do so well although in all probability it could have survived the BREXIT period.

On your suggestion of testing the trailing profit feature, why don't you try it and share your results here???

Cheers Verocity.
  • #305
  • Apr 9, 2018 7:35pm Apr 9, 2018 7:35pm
Oh Verocity, no wonder it produces that impressive results - you had used the D1 time frame!!!

Admittedly, I had NOT tested using D1 time frame as I know in reality, it would take a hell long time before the EA makes a trade but then again, once it entered a trade, it should be profitable as the STOCH signal in D1 is pretty reliable (silly me, we could trade multiple pairs then as the EA would not trade so often).

Verocity, I tell you what, you test it thoroughly in D1 time frame over a longer period using BOTH the legs and also on other pairs and share the results here and if its really viable we can start a new thread on this - AGREEABLE Verocity??? (And all who are reading this, you can test it also - but do NOT be bothered testing it on anything lower than D1 - I Have already done that).

If the test results are favourable, I will post all of my EAs here together with the source codes so that we can develop it further.

Actually Verocity, quite a few people also PM me on this as they, just as yourself, can clearly see the potential of that EA.

Cheers Verocity, hope you take up the challenge (and also all those who had PM me on this V1.6 version).
  • #306
  • Apr 9, 2018 7:39pm Apr 9, 2018 7:39pm
Quoting Philip Lee
Hi Verocity, Yes, the back test results you posted in post 301 is VERY impressive, especially when the period you used covered the BREXIT period - but Verocity, try to run the Seller on the same period, I am sure it will not do so well although in all probability it could have survived the BREXIT period. On your suggestion of testing the trailing profit feature, why don't you try it and share your results here??? Cheers Verocity.
Do you know what the "BE" means under trailing? Also, heres the seller test. It's not awesome, but it made profit, and I bet I could run an optimization to run on it to make it perform better. Thats the same settings as the buyer. I did change the Stoch to 10 instead of 20 so that it only took the major moves. I love that ability. The only reason its drawing down from the looks of it is from the initial 0.01 orders , which in the new version doesn't exist anymore, so it would be even more profitable..
Attached Image (click to enlarge)
Click to Enlarge Name: Screenshot1.png Size: 63 KB
  • #307
  • Apr 9, 2018 7:39pm Apr 9, 2018 7:39pm
I need to go to work now...will chat again in the evening.

Cheers.
  • #308
  • Apr 9, 2018 7:42pm Apr 9, 2018 7:42pm
Quoting Philip Lee
Oh Verocity, no wonder it produces that impressive results - you had used the D1 time frame!!! Admittedly, I had NOT tested using D1 time frame as I know in reality, it would take a hell long time before the EA makes a trade but then again, once it entered a trade, it should be profitable as the STOCH signal in D1 is pretty reliable (silly me, we could trade multiple pairs then as the EA would not trade so often). Verocity, I tell you what, you test it thoroughly in D1 time frame over a longer period using BOTH the legs and also on other pairs and...
I will continue to test on D1 and see what I can find! We should be testing D1 on the 2.4 version as well! So much testing.
  • #309
  • Apr 9, 2018 8:19pm Apr 9, 2018 8:19pm
Quoting Verocity
Because of the Stoch filter you can optimize the depth settings and produce some amazing results (yes the older 1.6). I attached a screenshot of a rough backtest of that. So now I'm wondering if we can run this in the hybrid setup (if we add stoch). I am still trying to understand the strategy in its entirety, so bare with me on that part. {image}
Hi, I think that is the right direction to add a momentum filter. We could also consider one other filter, TSR (30) , which is somewhat like a hull moving average. So seller would have control when below TSR (30) and buyer side when above TSR(30).

Also, not to be negative, but we must not get carried away by backtest results. Let me elaborate with my story
I thought I had recently created one of the most profitable EA's with my limited coding ability, it was downright simple strategy and every backtest was giving insane gains. I mean, tripling the account every 2 months with max 2% drawdown....It there was such a thing as holy grail, I thought I found it..

I was using mostly the control points or sometimes the every tick to tune the entry conditions. But the quality was always 90.00% or below. Then I came across this: https://www.mql5.com/en/blogs/post/704853
Retested my EA properly this time + forward tested and.... It was not so profitable and blew up the account once... Just be very careful how much trust you put into backtest to show how profitable your system may be in the future. I really hope you can find the right set of filters to this strategy, I would like to try test some of my filters, but I think there is only source code for the older version which is missing some functions I would need to code first before even adding filters and I don't think the 2.4 version source code will ever be shared.

Kind Regards,
Tzamo
"Work hard, put in the effort, and follow your dreams..." -Me
  • #310
  • Apr 9, 2018 8:49pm Apr 9, 2018 8:49pm
Hi Verocity,

I noticed that you used different sums for both the test - $5,000 for the Buyer and $1,000 for the SEller. I must say Verocity, the results are AMAZING!!!

As can be clearly seen that with only $1,000, the Seller had coped so well, so, you actually only need to start off with $1,000 with the Buyer and still got plenty of equity left.

So, you could have achieved the same results as you have posted if you have started the Buyer with $1,000 only which means a return of more than 1,200% in that two years period and it also mean that the total profit made running BOTH the Buyer and Seller (independently) over that period would have made $12,478.56 with only a combined starting capital of $2,000.

That is truly impressive, especially bearing in mind that you had ran both the legs at the same time (independently). That is very encouraging.
  • #311
  • Apr 9, 2018 8:55pm Apr 9, 2018 8:55pm
I can see the potential of the EA trading in multiple pairs on the D1 timeframe - in reality, it will not make so many trades but most trade cycles would be short and profitable bearing in mind that the EA will only trigger a trade upon the EXIT from and not upon entry into the preset STOCH level (default set at 80 for sell order and 20 for buy order).

More tests needs to be done in order to verify this.

I hope others will chip in and start testing it.
  • #312
  • Apr 9, 2018 9:38pm Apr 9, 2018 9:38pm
Quoting tzamo
{quote} I thought I had recently created one of the most profitable EA's with my limited coding ability, it was downright simple strategy and every backtest was giving insane gains. I mean, tripling the account every 2 months with max 2% drawdown....It there was such a thing as holy grail, I thought I found it.. I was using mostly the control points or sometimes the every tick to tune the entry conditions. But the quality was always 90.00% or below. Then I came across this: https://www.mql5.com/en/blogs/post/704853 Retested my EA properly...

This is a very good post which tells the truth about testing. I've used the tick data suite for a few years and even the previous methods used by its author from about 15 years ago when the tools were free. They are very good tools. But I stopped my subscription a couple of years ago. Am thinking of starting again next month (with the ideas here) after I get back home, but for me this is a big undertaking in time for an old guy hehe. I would recommend everyone be very careful about MQ data and not get their hopes up too much until good data is available and forward testing matches the backtesting.

Best to all.
  • #313
  • Apr 9, 2018 9:56pm Apr 9, 2018 9:56pm
Daily hybrid per tick test underway, custom depth settings! Look at this beast!
Attached Image (click to enlarge)
Click to Enlarge Name: dailyhybrid.JPG Size: 139 KB
  • #314
  • Apr 9, 2018 10:36pm Apr 9, 2018 10:36pm
Quoting Verocity
I am still having 0 luck using the hybrid setting. What does BE mean under trailing?
BE = Break even, at whatever pip or probably money value in this case, you want to put in there. I've not tested this or trailing yet though so don't know if it's working properly.
Nice result you have there.
  • #315
  • Apr 9, 2018 10:57pm Apr 9, 2018 10:57pm
Sorry for a possible dumb question. I spent the last 2 days reading all the posts and am getting ready to install the code. Did you ever decide on a max stop loss (pips or dollars)?
  • #316
  • Apr 10, 2018 12:44am Apr 10, 2018 12:44am
Quoting Verocity
Daily hybrid per tick test underway, custom depth settings! Look at this beast! {image}
hi verocity,
Can you share us your settings on such results? Been running backtest on 2.3 hybrid but don't think can produce such great result like yours.
  • #317
  • Apr 10, 2018 1:19am Apr 10, 2018 1:19am
Quoting petewong052
{quote} hi verocity, Can you share us your settings on such results? Been running backtest on 2.3 hybrid but don't think can produce such great result like yours.
Stop backtesting and start running forward tests, the last version is producing some phenomenal results for me
  • #318
  • Apr 10, 2018 1:45am Apr 10, 2018 1:45am
Quoting tictac-nur
{quote} Can you explain more detail what option you want it to have? Currently if you choose to run BUY and SELL at the same time. When BUY reach TP.. BUY will open new virtual PO. It is same for SELL order. If BUY reach TP first. SELL will be still on drawdown. Da you want it to close the sell open order with a loss?
thanks tictac, plz try and contact me or how do i contact u, u can do an ea together.
  • #319
  • Apr 10, 2018 1:50am Apr 10, 2018 1:50am
Quoting Verocity
Daily hybrid per tick test underway, custom depth settings! Look at this beast! {image}
Check on lower timeframes, perhaps you will not get the same results.
Forex; the game of movements not the game of candles.
  • #320
  • Apr 10, 2018 5:36am Apr 10, 2018 5:36am
Hi Magregus,

Care to share your results?
0 traders viewing now