Robot Lab Project | Page 12 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Financial Markets Observatory Lab 1,095 replies

Indicators Lab 236 replies

Rel's Lab 103 replies

Nikola's Trading Lab 247 replies

Spud's Lab Accident 22 replies

Attachments: Robot Lab Project
Exit Attachments

Robot Lab Project

  • #221
  • Jan 3, 2015 8:46am Jan 3, 2015 8:46am
Quoting gimin
Dear sir : sorry for my ask... the Corr_Diver_01 Robot have MT4 ver ??
Once we reach a final version on MT5, I will convert it to work on MT4, which then can be used for Forward Testing on MT4.
You know, I had to switch to MT5 for this robot so that we can do back testing, because back testing robots that work on multiple pairs is not possible on MT4.
No guts, no glory
  • #222
  • Jan 3, 2015 9:01am Jan 3, 2015 9:01am
Quoting abokwaik
Correlation Divergence Trading (Two Pairs){image}
Still testing and seeing positive results on D1 and H1 so far but will update when I have done plenty of tests.

Will share results by tomorrow evening
We miss 100% of the shots we don't take
  • #223
  • Jan 3, 2015 9:14am Jan 3, 2015 9:14am
Please post test results with various pairs from 2010 then. Thanks friends.
  • #224
  • Edited Jan 4, 2015 2:01am Jan 3, 2015 9:25am | Edited Jan 4, 2015 2:01am
Quoting B-Naire2B
Please post test results with various pairs from 2010 then. Thanks friends.
I am out of the office now. Can do it tomorrow.

However I don't think it's necessary, on this Raw version any way.

Remember, I am using very simple (probably inaccurate) method for divergence detection between the 2 pairs.
The robot needs to provide:

  1. multiple ways of divergence detection using one or more indicators or maybe a method like I posted earlier in excel.
  2. support negative Correlation.
  3. better lot size for each pair, it should be inversely proportional to their volatility which can be measured by ATR

...So again, this just a start-up version of this robot. Stay tuned ..

No guts, no glory
  • #225
  • Jan 3, 2015 11:55am Jan 3, 2015 11:55am
I meant for future references. I myself cannot do it, because either my internet is slow or smth else. So just asking if people are going to test other versions or pairs on their MT5 please do so for 2010-2014. Thanks.
  • #226
  • Edited 9:04am Jan 4, 2015 8:53am | Edited 9:04am
Correlation Divergence Trading (Two Pairs)
Second Version
Corr_Diver_02 (MT5 Only)

Updates :
1. Input parameters :

  1. ATRPeriod = 50 ATR period.

2. Robot will use the Fixed_Lot_Size for the first pair, and calculate relative lot size for second pair based on ATR ratio between the 2 pairs.

  1. for example if Daily ATR(50) for EURUSD was 60 pips, and for GBPUSD was 80 pips, and the the Fixed_Lot_Size=0.4 then the ratio will be 60/80=0.75, so the lot size for GBPUSD will be 0.75 * 0.4 = 0.3

I tried different indicators but found CCI(2) is the best one to detect pairs divergence, so I will stick with it for now. I haven't tried the calculations I used in excel yet, once I have free time I will try it (I think it will be more natural than CCI).

If you guys have better ideas on divergence detection, you can post it here.

ROBOT USE: I recommend Daily Time Frame, with MIN 1.5:1 Reward to Risk Ratio. I prefer also using Equity Increase/Decrease Percentages over SL and TP, so you can set SL and TP to valid large values that can not be hit.

When using lower time frames, Lot size and Equity Increase/Decrease Percentages should be changed accordingly.

You can test with the following pairs (positively correlated):

  1. EURUSD and GBPUSD
  2. EURJPY and GBPJPY
  3. EURJPY and USDJPY
  4. AUDUSD and NZDUSD
  5. and so on

Next version will include and option for trading negatively correlated pairs.

Sample test result of this version.

Attached Image (click to enlarge)
Click to Enlarge Name: cd_02_tr.PNG Size: 29 KB

Attached Image (click to enlarge)
Click to Enlarge Name: cd_02_tr_gr.png Size: 9 KB


Full test report attached.
Attached File(s)
File Type: ex5 Corr_Diver_02.ex5   32 KB | 454 downloads
File Type: mq5 Corr_Diver_02.mq5   7 KB | 586 downloads
File Type: rar CD_02_EU_GU_2010_2014.rar   88 KB | 528 downloads
No guts, no glory
  • #227
  • Jan 4, 2015 9:08am Jan 4, 2015 9:08am
Correlation Divergence Trading (Two Pairs)

Something to think of :

Why don't we trade EURGBP based on EURUSD and GBPUSD Correlation Divergence ? or

EURUSD based on EURJPY and USDJPY Correlation Divergence ? ....

No guts, no glory
  • #228
  • Jan 5, 2015 12:54am Jan 5, 2015 12:54am
Quoting abokwaik
Correlation Divergence Trading (Two Pairs) Second Version Corr_Diver_02 (MT5 Only) Updates : 1. Input parameters : ATRPeriod = 50 ATR period. 2. Robot will use the Fixed_Lot_Size for the first pair, and calculate relative lot size for second pair based on ATR ratio between the 2 pairs. for example if Daily ATR(50) for EURUSD was 60 pips, and for GBPUSD was 80 pips, and the the Fixed_Lot_Size=0.4 then the ratio will be 60/80=0.75, so the lot size for GBPUSD will be 0.75 * 0.4 = 0.3 I tried different indicators but found CCI(2) is the best one to...
Hi ablkwaik,

I carried out back testing with same parameters but ended up with a poorer result. Just wonder what I am doing wrong or differently.

Please see attached.
Attached File(s)
File Type: docx Strategy Tester Report.docx   26 KB | 411 downloads
  • #229
  • Jan 5, 2015 2:45am Jan 5, 2015 2:45am
Quoting pipster51
{quote} Hi ablkwaik, I carried out back testing with same parameters but ended up with a poorer result. Just wonder what I am doing wrong or differently. Please see attached. {file}
As far as I can see, Nothing wrong with your test.

I noticed that my test resulted 122 trades while yours resulted 118.

Although History Quality is the same in both tests (91% as MT5 reported),
however your test had 119,362,863 number of Ticks which is much bigger than mine (6,919,522).

Based on the above I think your test is better than mine.
No guts, no glory
  • #230
  • Jan 5, 2015 4:08am Jan 5, 2015 4:08am
Quoting abokwaik
{quote} Second version attached as requested. New Inputs : Time_Frame = 0; // Time Frame Used (0:current time frame) // 0:current time frame. Otherwize input number of minutes for desired time frame // 5 for M5, 60 for H1, 1440 for D1 and so on. Max_Seconds_to_Activate = 60; // Max. Seconds to activate after candle starts As I mentioned earlier, this robot need some filter, but I will leave that to you. Should you find it profitable, kindly share your settings. {file} {file}
Thanks for the updates Abokwaik.

I hear you re: the Filter but as yet am testing this as is.

So far the Daily TF is the most stable. Testing on 20 pairs with 0.02 per $1,000. TP / SL = beginning and ending of the HA "trend".

Still testing and will report my findings as I discover them.

Regards,

Ty
My job: I speculate and manage risk.
  • #231
  • Jan 5, 2015 4:22am Jan 5, 2015 4:22am
Quoting TyLamai
{quote} Thanks for the updates Abokwaik. I hear you re: the Filter but as yet am testing this as is. So far the Daily TF is the most stable. Testing on 20 pairs with 0.02 per $1,000. TP / SL = beginning and ending of the HA "trend". Still testing and will report my findings as I discover them. Regards, Ty
Are you Back Testing or Forward Testing ?

If Forward, I hope its a demo account.
No guts, no glory
  • #232
  • Jan 5, 2015 5:05am Jan 5, 2015 5:05am
Quoting abokwaik
{quote} Are you Back Testing or Forward Testing ? If Forward, I hope its a demo account.
Oh Forward...and yes it IS a demo account. Armada.

Regards,

Ty
My job: I speculate and manage risk.
  • #233
  • Jan 5, 2015 5:45am Jan 5, 2015 5:45am
Correlation Divergence Trading (Two Pairs)
Third Version
Corr_Diver_03 (MT5 Only)

Updates :

  1. Correlation Factor : 1 for Positive, -1 for Negative Correlation between the 2 pairs
  2. Pairs Ratio : 1st pair to 2nd pair volatility ratio (better entered manually than calculationg using ATR specially on lower time frames)
  3. CCI Extereme Value : If using CCI period 2, leave this to 0 (zero), but when using periods like 20, 30 ... etc, you can use higher values for this parameter (i.e. 100). This parameter will filter the trades, the higher value the less trades.


Test results on EURUSD and USDJPY with -1 Correlation Factor

Attached Image (click to enlarge)
Click to Enlarge Name: cd_03_eu_uj_2014.PNG Size: 29 KB

Attached Image (click to enlarge)
Click to Enlarge Name: cd_03_inv_eu_uj (1).png Size: 9 KB
Attached File(s)
File Type: mq5 Corr_Diver_03.mq5   8 KB | 795 downloads
File Type: ex5 Corr_Diver_03.ex5   32 KB | 525 downloads
No guts, no glory
  • #234
  • Jan 5, 2015 6:49am Jan 5, 2015 6:49am
Quoting abokwaik
{quote} Second version attached as requested. New Inputs : Time_Frame = 0; // Time Frame Used (0:current time frame) // 0:current time frame. Otherwize input number of minutes for desired time frame // 5 for M5, 60 for H1, 1440 for D1 and so on. Max_Seconds_to_Activate = 60; // Max. Seconds to activate after candle starts As I mentioned earlier, this robot need some filter, but I will leave that to you. Should you find it profitable, kindly share your settings. {file} {file}
Happy new Year Abokwaik,
Thanks for your generous effort in helping us out. Am trying to test this on XAUUSD H4 without any filter. Results do not seem bad at all and i hope they can improve greatly if you add a Trailing stop option. At the moment there is that big DD trade cycle that am trying to get rid of and this will help. Plz find attached. Will start forward testing this when TS is added and will share my best settings. Thank you once again for this wonderful thread.
Happy Trading in 2015
Attached Image (click to enlarge)
Click to Enlarge Name: MetaTrader - Pepperstone.png1.png Size: 195 KB
  • #235
  • Edited 12:13pm Jan 5, 2015 11:04am | Edited 12:13pm
Quoting nogames2014
{quote} Happy new Year Abokwaik, Thanks for your generous effort in helping us out. Am trying to test this on XAUUSD H4 without any filter. Results do not seem bad at all and i hope they can improve greatly if you add a Trailing stop option. At the moment there is that big DD trade cycle that am trying to get rid of and this will help. Plz find attached. Will start forward testing this when TS is added and will share my best settings. Thank you once again for this wonderful thread. Happy Trading in 2015 {image}
Thank you,

Attached as requested,

HA_Rider_03

New Inputs:

  1. Stop_Loss_Pips = 100; // Stop Loss in Pips (per order)
  2. Trailing_Stop_Pips = 100; // Trailing Stop in Pips (per order)
  3. Trailing_Stop_Jump = 10; // Trailing Stop Jump (to reduce order modification requests

​Enjoy ...

Attached File(s)
File Type: mq4 HA_Rider_03.mq4   12 KB | 675 downloads
File Type: ex4 HA_Rider_03.ex4   16 KB | 523 downloads
No guts, no glory
  • #236
  • Jan 5, 2015 1:33pm Jan 5, 2015 1:33pm
Quoting abokwaik
{quote} Thank you, Attached as requested, HA_Rider_03 New Inputs: Stop_Loss_Pips = 100; // Stop Loss in Pips (per order) Trailing_Stop_Pips = 100; // Trailing Stop in Pips (per order) Trailing_Stop_Jump = 10; // Trailing Stop Jump (to reduce order modification requests ​Enjoy ... {file} {file}
Goody!

Great job abokwaik!

Demo currently at 50%+ "floating" profit...and counting. Before this, the demo had a 25% DD. Therefore at this point all currently open positions can be closed for a Net of 25%+ for the month...and the a vacation can be declared for the rest of the month but this is not how we roll.

I will leave the trades running to see how the EA will handle itself. I expect that much more of the present "floating" profit will be eaten up (hopefully temporarily) when more positions are opened in the prevailing HA "trend" direction and when the D1 Heiken Ashi "trend" of all 20 pairs eventually turn.

As an example: the demo currently has two GBPJPY positions with a combined pip-haul of close to 700pips...this is not counting other pairs. So, maybe we may have something here...too early to say though.

By the way, I am still using HA_Rider_01. Somehow I haven't gotten around to upgrading yet...but I will soon.

Thank you abokwaik.

Regards,

Ty
My job: I speculate and manage risk.
  • #237
  • Jan 5, 2015 3:39pm Jan 5, 2015 3:39pm
Quoting abokwaik
{quote} Thank you, Attached as requested, HA_Rider_03 New Inputs: Stop_Loss_Pips = 100; // Stop Loss in Pips (per order) Trailing_Stop_Pips = 100; // Trailing Stop in Pips (per order) Trailing_Stop_Jump = 10; // Trailing Stop Jump (to reduce order modification requests ​Enjoy ... {file} {file}
Thanks Albokwaik for the quick response. However including a fixed stop and ts makes it worse and will lead to heavy drawdown. I am completely with you we need some kind of filter. Am forward testing it on gold usd starting tomorrow. Will be posting results.
Thanks once again!
Ty,
To me H4 results on XAUUSD are more stable than day. GL on yo forward test...hope we get the best settings and pairs.
  • #238
  • Jan 5, 2015 4:02pm Jan 5, 2015 4:02pm
Albokwaik, sorry for asking too much. It seems EA trades one trade per candle...Is this the case? And if so, can u set it to have various trades as long as the trade spacing has been reached..Thanks in advance.
  • #239
  • Jan 5, 2015 4:12pm Jan 5, 2015 4:12pm
Quoting nogames2014
{quote} Thanks Albokwaik for the quick response. However including a fixed stop and ts makes it worse and will lead to heavy drawdown. I am completely with you we need some kind of filter. Am forward testing it on gold usd starting tomorrow. Will be posting results. Thanks once again! Ty, To me H4 results on XAUUSD are more stable than day. GL on yo forward test...hope we get the best settings and pairs.
Usually it's better not to use SL for systems based on Close&Reverse concept like HA_Rider because your SL is actually your TP.However HA_RIDER_03 has several parameters when setup intelligently they can produce good results. Namely distance between orders, SL & TS. One scenario is : Distance > SL = TS.As for Forward testing I didn't move any robot to Forward testing yet. They still need some work. Once I reach a workable version of a robot I will link it to a Trade Explorer on this thread.
No guts, no glory
  • #240
  • Jan 5, 2015 4:18pm Jan 5, 2015 4:18pm
Quoting nogames2014
Albokwaik, sorry for asking too much. It seems EA trades one trade per candle...Is this the case? And if so, can u set it to have various trades as long as the trade spacing has been reached..Thanks in advance.
I am not in front of my computer at the moment (phone texting) . But as I remember there is a parameter for minimum pips between orders, although the original idea by Ty was to open on each candle.
No guts, no glory
0 traders viewing now