Help Needed to Debug Existing EA with proven strategy | Page 19 | 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

  • #361
  • Apr 11, 2018 3:29pm Apr 11, 2018 3:29pm
Quoting jamco388
{quote} I use the same setting on my post 282. I will try for 2 week with start balance $500
hi mate which version of EA you have used?

cheers
Andy
  • #362
  • Apr 11, 2018 3:42pm Apr 11, 2018 3:42pm
Quoting Philip Lee
Pete, please read my last post 297 - it is also meant for you. Pete, be patient and follow up with the development here on trading the EA on hybrid mode - it is a far safer way to trade this EA. Do NOT use v1.6 to trade live. Cheers Pete.
Hi Philip,

Thank you for your advise. Been backtesting on the EA hybrid mode but did not get the positive result like others had. Could you advise what was the timeframe chart used on your post #274?
Meantime, my forward test on EA v2 also did not do well with much drawdown locked between the buy and sell trades left open. I was then try to set the auto Buy & Sell to false to prevent any further trade open but surprisingly it immediately closes all the open trades with much losses around -$550.
The only best result I could get is from back testing on 2.4v under 5m chart replicating settings from verocity but that too does not seem fit my appetite due to the difference deposit of my initial capital.
I do hope that someone can come out with a manual guide settings in near future as the variable options seem expanding quite rapid
Attached Image(s) (click to enlarge)
Click to Enlarge Name: StrategytestV2.4_5m.JPG Size: 114 KB
Click to Enlarge Name: StrategytestV2.4chart.jpg Size: 66 KB
  • #363
  • Apr 11, 2018 4:31pm Apr 11, 2018 4:31pm
Quoting Verocity
Backtesting is by no means perfect AT ALL, we must take it lightly. It will however give us a good idea if we have things functioning properly before forward testing.
I agree 100% I always believed it was useless!
But I think there is value in it. It helps find out if the EA is doing everything it needs to do. But I still feel this can be done forward testing. It will just take a bit longer. Also back testing gives an idea about different possibilities and opens our minds to think of tweaking and improvements. It can still never be used as indication for furture profits at any level.
  • #364
  • Apr 11, 2018 7:15pm Apr 11, 2018 7:15pm
Pete,

You have over traded. How could $3,000 cope with 1.0 full size lot. I had already repeatedly stated that you need $1,000 per 0.01 lot traded. So, in your demo account, you should only have set the suborder lot size at 0.03 and NOT 1.0 - you are way way over-traded (since you started the test with $3,000 only).

Pete, reading your posts I can tell that you are a total newbie to forex. Pete, DO NOT EVEN THINK ABOUT TRADING LIVE EITHER WITH THIS EA OR ANY OTHER EA. You will get burned in no time. Be patient and learn first - I do not wish to be too blunt here but its very obvious that you are not ready to trade live yet. Be patient and learn first. There is no easy quick money in forex.

Sorry to be so blunt Pete.

Cheers Pete.
  • #365
  • Apr 11, 2018 11:35pm Apr 11, 2018 11:35pm
Quoting densol
got a question here guys, for the live account, if you wanna activate buy one cycle only, you also need to set buy to true? if not buy one cycle not work? where in backtest its possible to do and what is risk reward fixed capital in the ea is for?
sorry i am busy lately with other project..
no free time anymore..

1) Continous buy after profit target
- BUY=true

2) Buy until profit target then stop trading
- Buy=true
-One_Cycle_Only=true

3) Buy until profit target then restart trading back on the following day
- Buy=true
-One_Cycle_Only=true
-Restart_Next_Day = true

4) Risk : Percentage % of risk for lot size calculation if not using Fixed_lot
exp: 1 actually each entry is made with 1% of risk when it hit the stop loss

5) Reward : Percentage % of profit target
exp: 1 actually profit target of 1% from fixed capital/current money

6) Fixed Capital : if you set not using compounding. Risk will be calculated base on this Fixed Capital.

Happy Trading Guys..
Enjoy the profits..
It is like atm machine..
Dont be greedy.. (Lamborghini aventador)
Just take the profit you need for food / clothes then withdraw from broker..
Enjoy the profit you just withdraw..
buy some food and clothes for yourself, your family and neighbours..
The more you give..
The more you will get..
It will make you happy..

See you next time.. bye..
  • #366
  • Apr 12, 2018 12:55am Apr 12, 2018 12:55am
Quoting montego
M5 eur usd jan 2017 to jan 2018 set, nearly doubles the account from £1000 to £1900 with 22% Drawdown {image} {file}
Hello and Thank You everyone for contributing to this great work and special thanks to Philip Lee for his generosity and Tictac for her artistic work.

I am trying to backtest EU same settings as montego but for years 2009 to 2016 to see these years were profitable as well. Not understood Hybrid mode just yet but non hybrid is not bad at all.

P.S , each year performance will be shared separately.

Best Regards
skat
  • #367
  • Edited 4:48am Apr 12, 2018 2:32am | Edited 4:48am
Hello,

I am searching for a coder, that could make a robot acording my trading strategy. Why i look for coder if the strategy works, well answer is simple i want to avoid psychological factors, because sometimes you see but you think to much and u miss a good signal, so i want to make it automatic. If anyone is interested please let me know your skype, that we could speak. Or you can find my in the skype with this emal. { email address deleted by staff }

By the way i am junior member here so i canot make any PM messages. So please be informed by that.
  • #368
  • Apr 12, 2018 8:52am Apr 12, 2018 8:52am
Hi Philip,

Great system! love the way it trades without much risk, thanks for sharing!
I will invest more time and when i come up with some ideas will share here.
Persistence will get it. Consistency will keep it.
  • #369
  • Apr 12, 2018 8:55am Apr 12, 2018 8:55am
Quoting yalgaar
Hi Phillip I have been thinking about the deadlock situation. I have the following 2 suggestions to handle it:...
Great ideas yalgaar!
I think number 1 is the most interesting suggestion.
Persistence will get it. Consistency will keep it.
  • #370
  • Apr 12, 2018 10:37am Apr 12, 2018 10:37am
Quoting 8Lietuvis8
Hello, I am searching for a coder, that could make a robot acording my trading strategy. Why i look for coder if the strategy works, well answer is simple i want to avoid psychological factors, because sometimes you see but you think to much and u miss a good signal, so i want to make it automatic. If anyone is interested please let me know your skype, that we could speak. Or you can find my in the skype with this emal. { email address deleted by staff } By the way i am junior member here so i canot make any PM messages. So please be informed by...
Just start another thread and describe your idea....I am pretty sure if your ideal is even somewhat appealing, many coders will be happy to help.
  • #371
  • Apr 12, 2018 10:46am Apr 12, 2018 10:46am
Quoting tictac-nur
{quote} sorry i am busy lately with other project.. no free time anymore.. 1) Continous buy after profit target - BUY=true 2) Buy until profit target then stop trading - Buy=true -One_Cycle_Only=true 3) Buy until profit target then restart trading back on the following day - Buy=true -One_Cycle_Only=true -Restart_Next_Day = true 4) Risk : Percentage % of risk for lot size calculation if not using Fixed_lot exp: 1 actually each entry is made with 1% of risk when it hit the stop loss 5) Reward : Percentage % of profit target exp: 1 actually profit...
hey thanks for the explanation, hard to figured out things, maybe i'm just slow
  • #372
  • Edited 11:23am Apr 12, 2018 11:09am | Edited 11:23am
Quoting detector
Hi All, following with interest, moving in line with Velocity's test, played around with some inputs and posted results on EUR/USD 5min chart, D {image} {image} {image} {image}
Attached Image (click to enlarge)
Click to Enlarge Name: Screenshot3.png Size: 52 KB
Attached Image (click to enlarge)
Click to Enlarge Name: Screenshot2.png Size: 88 KB


i test with your setting from 1998, still too choppy, around 60% of it is up, 60% chance of doubling an account when use aggressively
perhaps once we doubled our account and reinvest the profit immediately and if we lose we wait try again in a week then if it fails again we try every month and once it's profitable again we reinvest it straight away, therefore we participate more in profitable period.
  • #373
  • Apr 12, 2018 11:15am Apr 12, 2018 11:15am
Hi,

I wish I could test this EA but all the time backtest stops saying there are some errors, like now with EUrUSD:
2018.04.12 18:13:49.122 TestGenerator: unmatched data error (low value 1.23072 at 2018.03.29 01:25 is not reached from the least timeframe, low price 1.23080 mismatches)


can someone help me with this?

Thanks
mojepole.
  • #374
  • Apr 12, 2018 11:46am Apr 12, 2018 11:46am
Quoting mojepole
Hi, I wish I could test this EA but all the time backtest stops saying there are some errors, like now with EUrUSD: 2018.04.12 18:13:49.122 TestGenerator: unmatched data error (low value 1.23072 at 2018.03.29 01:25 is not reached from the least timeframe, low price 1.23080 mismatches) can someone help me with this? Thanks mojepole.
it meant that the value 1.23072 in higher timeframe can't be reached from lower timeframe (1min), its a mismatch of historical data, just try different pairs. or try downloading the historical data (tools- history center).
  • #375
  • Edited 1:44pm Apr 12, 2018 1:02pm | Edited 1:44pm
Quoting densol
{quote}{image}{image} i test with your setting from 1998, still too choppy, around 60% of it is up, 60% chance of doubling an account when use aggressively perhaps once we doubled our account and reinvest the profit immediately and if we lose we wait try again in a week then if it fails again we try every month and once it's profitable again we reinvest it straight away, therefore we participate more in profitable period.
Densol, did you align the times with London open and NY close, we are +2 on server time so my start was 10 and finish 23, think this was posted in the second set, this is where I changed the trading days but I see you have aligned them inputs, (trading days) great work, D
I don't need to be good at the game, only good enough to beat my opponent,
  • #376
  • Apr 12, 2018 1:52pm Apr 12, 2018 1:52pm
Seems like most of the back testing done here by most of you is for non-hybrid mode. I think we should collectively do testing of different combinations and create 1 report so this can be analyzed together. I would like to start a discussion of such combinations first and once we create a list of all the test we want to run we can assign it to each one of us who is interested to help.

We need Volunteers please to start this discussion and take this further.

Just to kick start the discussion....here are my thoughts about what and how I would like us to test:

1) We test this one at least 10 pairs....EURUSD, GBPJPY, GBPUSD, EURJPY....etc etc. Lets come up with this list based on some reasoning.
2) We test this in Buy mode, Sell Mode, Hybrid Mode, Buy & Sell mode.
3) I believe we should just test this on the best data quality...so no cutting corners...and we so all out tests on tick accurate data of M1.
4) All our tests should be for 1 year. We can do tests for last 5 years.
5) Every test should show the settings, the graph, the report...anything else we want to add.

Ladies and Gents-
This exercise will help us understand the strategy and its possibilities much better. It will also help us fix some challenges we are seeing right now.
  • #377
  • Apr 12, 2018 1:58pm Apr 12, 2018 1:58pm
Densol, also noticed the spread was 18, mine was based on the current of 8, also going back over many years the spread may well have been very different, as has been stated, back testing and getting good results can only make us envious of what we misses if it were true, LOL, hope this helps, back testing from 2017/04/25 shows results as follows, based on $1000, D
Attached Image(s) (click to enlarge)
Click to Enlarge Name: TDI.png Size: 18 KB
Click to Enlarge Name: TDI2.png Size: 19 KB
I don't need to be good at the game, only good enough to beat my opponent,
  • #378
  • Apr 12, 2018 5:13pm Apr 12, 2018 5:13pm
The backtesting looks good. But I've had many EAs that back test. Put them live/demo account and they fail misserably

I placed it on the euro today and it didn't take a trade. Also on the Dow and it didn't play nice at all even tuned and backtested on Dow
  • #379
  • Apr 12, 2018 6:01pm Apr 12, 2018 6:01pm
Quoting Redsub
The backtesting looks good. But I've had many EAs that back test. Put them live/demo account and they fail misserably I placed it on the euro today and it didn't take a trade. Also on the Dow and it didn't play nice at all even tuned and backtested on Dow
Agree with you there, looking to run a live test, the only way to find any faith, could you put the set files up for the Dow Redsub, I have been trawling through the visual mode on the tester to see how the EA performs, high volatility seems to catch the best moves, maybe Phillip Lee could confirm this, when it drops back it can catch you out, currently doing as Yalgaar suggested, testing in hybrid mode etc.. D
I don't need to be good at the game, only good enough to beat my opponent,
  • #380
  • Apr 13, 2018 2:54am Apr 13, 2018 2:54am
Quoting detector
Densol, also noticed the spread was 18, mine was based on the current of 8, also going back over many years the spread may well have been very different, as has been stated, back testing and getting good results can only make us envious of what we misses if it were true, LOL, hope this helps, back testing from 2017/04/25 shows results as follows, based on $1000, D {image} {image}
yeah thats what i hate about high freq trading system, the spreads and execution speed plays huge part in result, i found 2 ea that pretty much makes money in all market in history test but on live, not so good. i heard jforex from dukascopy have much better execution speed
0 traders viewing now