osfx.RMI.Inside portfolio EA | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

RMI Candles... 0 replies

RMI the child of RSI HELP! 0 replies

Wanted : RMI indicator 8 replies

Portfolio Management 34 replies

Portfolio management 3 replies

Attachments: osfx.RMI.Inside portfolio EA
Exit Attachments

osfx.RMI.Inside portfolio EA

  • #1
  • First Post: Jul 3, 2017 1:05pm Jul 3, 2017 1:05pm
This thread refers to cubbybgood's thread https://www.forexfactory.com/showthread.php?t=641014 and shows the results of the EA, I created on MT5 for this strategy.

Unfortunately, FF doesn't support MT5-trade explorers and myfxbook doesn't really work well with MT5. So I decided to use fxblue.

Version 01 - basic rules (RMI + inside candle) + volatility filter, Pairs: AUDCAD,AUDUSD,CADCHF,EURCHF,EURGBP,GBPCAD,GBPCHF,NZDCAD,USDCAD
https://www.fxblue.com/users/25031122
  • #2
  • Jul 4, 2017 2:53pm Jul 4, 2017 2:53pm
Thanks for all the interesting tests. I would do testing as well, but unfamiliar and not educated enough yet. Seems like I may have to build my own. I have just started reading the MQL4 book and watched some Jimdandy videos. I would like to have this Inside candle/RMI EA as well but it looks like I will have to work for it...I will be looking forward to your threads. Thanks
  • #3
  • Jul 22, 2017 1:09am Jul 22, 2017 1:09am
Hi Osfx,

I am following your system results with interest.

I have made progress in developing my EA of cubbybgoods method (attached below).

Attached Image (click to enlarge)
Click to Enlarge Name: Screenshot2.png Size: 42 KB


I found the only way I could stop a complete draw down was to use the following rules:
SELL:
If inside bar entry is true & inside bar close is Above the RMI 70 level
Close any open buy orders
execute sell order

I notice on your tester results you don't seem to have many stop outs from the 500 pip stop. During the trending periods where price would go beyond 500pips with very little pull back, the system would open 20+ orders from the 20 pip grid entry's, then I would get 10+ stop outs at 500 pips each when the market does not give 50 pips beyond the average entry price. The only way to avoid this was by changing the rules as above. Did you have to change the original rules or add any rules to make it work for you?

Blix.
  • #4
  • Edited 11:05pm Jul 22, 2017 1:48pm | Edited 11:05pm
Quoting Blix
Hi Osfx, I am following your system results with interest. I have made progress in developing my EA of cubbybgoods method (attached below). {image} I found the only way I could stop a complete draw down was to use the following rules: SELL: If inside bar entry is true & inside bar close is Above the RMI 70 level Close any open buy orders execute sell order I notice on your tester results you don't seem to have many stop outs from the 500 pip stop. During the trending periods where price would go beyond 500pips with very little pull back, the system...
Hi Blix,

I see, you're on the right track !

The settings of my portfolio are entirely different from cubbybgood's, and indeed, SL is hit quite often in my tests, but the power of portfolio makes up for those losses. To close all trades on appearance of an opposite signal is an interesting idea, I will test it.

I made the decision to share my rules, but I will not post them here. This is just to prevent anonymous visitors from dropping by and grap the settings.
Instead, I will share to those, who have been already active in the original thread. The "Links / Website" section in my profile shows a link to the rules, which is visible for buddies only (= two members, that subscribed to one another on FF).

Regards Oliver
  • #5
  • Jul 23, 2017 1:42am Jul 23, 2017 1:42am
" but the power of portfolio makes up for those losses"
This is a great tip for me, thank you.
By the way, I cant edit the post I made earlier but the exact implementation of the the rules is not an opposite signal (inside bar setup) but just a close in the 30 zone to exit all short positions.

From your posts it sounds like MT5 is a better platform for testing portfolio trading, I will look in to it.

Rergards,
Blix.
  • #6
  • Edited 2:27pm Jul 23, 2017 5:00am | Edited 2:27pm
Quoting Blix
" but the power of portfolio makes up for those losses" This is a great tip for me, thank you. By the way, I cant edit the post I made earlier but the exact implementation of the the rules is not an opposite signal (inside bar setup) but just a close in the 30 zone to exit all short positions. From your posts it sounds like MT5 is a better platform for testing portfolio trading, I will look in to it. Rergards, Blix.
MT5 really beats MT4, when it comes to multi-currency-testing, and with the MQL-Profiler, I reduced the time for one pass on 9 pairs for 7,5 years to 2 minutes...
And finally the best: Since yesterday, with build 1641, it is finally possible to create custom symbols and to import external price data. This eliminates one of most important downsides of MT5.

Regards, Oliver
  • #7
  • Jul 24, 2017 4:14pm Jul 24, 2017 4:14pm
Quoting osfx
{quote} Hi Blix, I see, you're on the right track ! The settings of my portfolio are entirely different from cubbybgood's, and indeed, SL is hit quite often in my tests, but the power of portfolio makes up for those losses. To close all trades on appearance of an opposite signal is an interesting idea, I will test it. I made the decision to share my rules, but I will not post them here. This is just to prevent anonymous visitors from dropping by and grap the settings. Instead, I will share to those, who have been already active in the original thread....


Any success with the rules?
ive been watching your fxblue
  • #8
  • Jul 25, 2017 3:03pm Jul 25, 2017 3:03pm
Quoting faatkumo
{quote} Any success with the rules? ive been watching your fxblue
Yes, it's running smooth and giving the expected results so far.
But I keep on going to work on some enhancements.
  • #9
  • Jul 25, 2017 3:14pm Jul 25, 2017 3:14pm
Quoting osfx
{quote} Yes, it's running smooth and giving the expected results so far. But I keep on going to work on some enhancements.



fantastic!
New to cubby, working on this for a couple more days before i start giving some feedback....

Enhancements meaning working on EA? or other indicators/pa in regards to methodology?
  • #10
  • Jul 26, 2017 12:48am Jul 26, 2017 12:48am
Quoting faatkumo
{quote} fantastic! New to cubby, working on this for a couple more days before i start giving some feedback.... Enhancements meaning working on EA? or other indicators/pa in regards to methodology?
Both of them
  • #11
  • Jul 28, 2017 2:38am Jul 28, 2017 2:38am
osfx,

I have been learning MT5 and kept hitting a brick wall until I discovered that my account was based on a hedge accounting system instead of a netted system. I am curios what system you are using? https://www.mql5.com/en/articles/2299
It makes a big difference in the coding for averaged entries; keeping separate stop losses for unique positions, or a stop loss for a aggregated position.

The multi currency aspect has been difficult to get my head around but its starting to come together now. I previously didn't use arrays much but now everything is coded as an array.
  • #12
  • Jul 28, 2017 3:48am Jul 28, 2017 3:48am
Quoting Blix
osfx, I have been learning MT5 and kept hitting a brick wall until I discovered that my account was based on a hedge accounting system instead of a netted system. I am curios what system you are using? https://www.mql5.com/en/articles/2299 It makes a big difference in the coding for averaged entries; keeping separate stop losses for unique positions, or a stop loss for a aggregated position. The multi currency aspect has been difficult to get my head around but its starting to come together now. I previously didn't use arrays much but...
I use the hedging system. In my eyes, this is the only option for a system like this, as it manages every trade separately, similar to MT4. The netting system accumulates all trades of a symbol to one position, which requires an entire different coding - and the killer argument: it only allows hard SL's for the complete position. Unfortunately, the accounting mode comes with the broker and you cannot change it yourself (as far as I know). I have found only 4 brokers so far, that provide the hedging mode.

The biggest difference between MQL5 and MQL4 is the handling of the trades. Whereas MQL4 only knows orders (pending or open), MQL5 comes with order, deals and positions...

Regards, Oliver
  • #13
  • Aug 3, 2017 4:06pm Aug 3, 2017 4:06pm
Ok, it's time for a summary after launching v01.

We had a decent growth of the balance of about 6%, but the floating drawdown is quite high (max. about 40% of initial balance). This is caused by three pairs (AUDUSD, EURCHF, GBPCHF), which have triggered and then moved the other way without significant pullbacks.
The idea was to trade 9 pairs with the same settings to diversify the risk leading to a smoother equity curve. This didn't work out very well.

Maybe we have to return to an individual examination of each pair, as each of them shows it's own characteristics. This can also lead to drop some of them again.
There are some other ideas, which are worth looking at. One of is to build a "multiple-timeframe cluster RMI". Some of you have already heard of cluster indicators, like CCFp. They rate the strength/weakness of a single currency by relating them to a couple of other currencies. Maybe we improve our odds by only buying/selling pairs with the strongest/weakest currencies involved.

I will be on vacancy with my family for the next 2 weeks, after that, I will keep on working on these ideas. meanwhile i will keep v01 running as it is.

Regards, Oliver
  • #14
  • Aug 5, 2017 5:40am Aug 5, 2017 5:40am
Osfx,
I have gone back to MT4 and the GBPCAD to try out different ideas to reduce draw downs. I was almost able to replicate your and Cubby's results on the GBPCAD. I have smoothed out a few draw downs but am stuck with the flash crash in October2016 and a long trend without much pullback in June2010.
Attached Image (click to enlarge)
Click to Enlarge Name: Screenshot3.png Size: 96 KB


I have tried:
Dynamicly reducing TP as more entries are taken >> this did help smooth some draw downs out but reduces overall profitability.

Exit on opposite RMI signal or exit on opposite RMI & IB signal >>this avoids the very deep draw downs from the original rules but reduces profitability and causes extended periods of choppy balance (I am going to revisit this idea as it looks quite robust if I can boost the profitability)
Attached test from 2007 to 2017
Attached Image (click to enlarge)
Click to Enlarge Name: Screenshot2.png Size: 40 KB


Initiate hedge on high range volatile bars and unwind the hedge when volatility subsides >> this looks promising but I am still working on this code, its a bit buggy , the idea is to avoid the flash crash bars. I have a few different hedging ideas to try out.
---
I am wondering if you found a way to avoid the 2010 draw down, and by the looks of things your volatility filter helped to avoid the 2016 spike, is this correct? Any clues to point me in the right direction is greatly appreciated.

I can also see that different brokers may get you in earlier or later which can make all the difference on being caught in a protracted trend.

Regards,
Blix
  • #15
  • Aug 5, 2017 12:19pm Aug 5, 2017 12:19pm
I am way behind you fellows, I am still trying to build a Inside candle RMI EA so I can also share and collaborate on my testing. I am assuming your EA testing is a inside candle RMI. Did you buy it ?, or code it ?
-The suggestion above on multiple RMI time frames seems like a swell idea!!
- When I get my inside RMI EA , I will most likely only turn it on after I manually take a trade, and let the robot finish the job. That is what I am gonna test out first. But I am sure there it will not be as profitable if I were to just leave it on 24 hrs.

Great post! Thanks , Cliff
  • #16
  • Aug 5, 2017 2:04pm Aug 5, 2017 2:04pm
Quoting Blix
Osfx, I have gone back to MT4 and the GBPCAD to try out different ideas to reduce draw downs...
Hi Blix,

which settings did you use for the test ? Cubbybgood's original ones or those from my document ?
I will then upload a detailed trade history for you to compare the trades at the specific periods with your's. Concerning the flash crash in 10/2016, the M1-rule, described in my document, should have saved you from a big DD. The DD from 2010 is another story - I noticed, that even small parameter changes (or broker price feeds ?) decide, whether you're involved in this DD or not, which is not good and can be regarded as a proof of curve fitting.

Exit on opposite RMI signal is indeed a smart idea, as it leads the characteristics of this strategy into a new direction.

Regards, Oliver
  • #17
  • Edited 8:05pm Aug 5, 2017 7:52pm | Edited 8:05pm
Quoting osfx
{quote} Hi Blix, which settings did you use for the test ? Cubbybgood's original ones or those from my document ? I will then upload a detailed trade history for you to compare the trades at the specific periods with your's. Concerning the flash crash in 10/2016, the M1-rule, described in my document, should have saved you from a big DD. The DD from 2010 is another story - I noticed, that even small parameter changes (or broker price feeds ?) decide, whether you're involved in this DD or not, which is not good and can be regarded as a proof of curve...
Osfx,
For the testing I am using the original parameters with an additional rule. I noticed a few periods where the system only just got out of a basket trade and was very close to a stop cascade (in the case of the 2010 draw down it got caught). To reduce the chance of being caught in the trade I added the additional rule of a dynamic TP adjustment to get out of trades earlier or at a small loss. So as the # of positions increases the NET pip target decreases by a factor of the #positions, this ends up being negative as the #positions increases beyond a certain amount, but it means you need a smaller pullback to get out of a large basket.
This rule still got me caught on the 2010 draw down, so it is a work in progress.
I am cautious about curve fitting, so I am trying to make my mods general rule changes rather than parameter tweaks.

BTW, thanks for the M1 tip. I'll be implementing that.

Cliff,
"- When I get my inside RMI EA , I will most likely only turn it on after I manually take a trade, and let the robot finish the job. That is what I am gonna test out first. But I am sure there it will not be as profitable if I were to just leave it on 24 hrs."

Personally I would use some kind of forward tester software like forex tester2 if I was going to test any manual approach. You need years of testing a negative Reward:Risk ratio strategy to know if it is going to blow up. This strategy relies on very small position size to allow for large baskets of trades, so you cant leverage it very far. If you were just manually testing over a few months you might think you are onto a winner and start increasing position size, then the first real draw down appears and you are way over leveraged. Forex tester would allow you to manually test several years of price action over a weekend and give you a very good idea of the strengths/weaknesses of the system.
  • #18
  • Edited 11:02pm Aug 5, 2017 10:52pm | Edited 11:02pm
Quoting Blix
{quote} Osfx, For the testing I am using the original parameters with an additional rule. I noticed a few periods where the system only just got out of a basket trade and was very close to a stop cascade (in the case of the 2010 draw down it got caught). To reduce the chance of being caught in the trade I added the additional rule of a dynamic TP adjustment to get out of trades earlier or at a small loss. So as the # of positions increases the NET pip target decreases by a factor of the #positions, this ends up being negative as the #positions increases...
HI, Would I be able to use mt4 ea in forex tester or is it different programming?
  • #19
  • Aug 6, 2017 12:45am Aug 6, 2017 12:45am
Quoting cliffvitale
{quote} HI, Would I be able to use mt4 ea in forex tester or is it different programming?
Forex tester is very similar to mt4 in terms of functionality, it does have a similar language to mt4 but not the same (I don't know about the current version). It has been a while since I have looked at it but I know of people writing indicators and scripts to manually test strategies. If you search forex factory you will find examples of how people have used it.

There are other platforms to manually test strategies within metatrader itself, this may be a good option. If you do a few searches there are several third party manual backtesting tools that work with metatrader.

To me it is the only way to efficiently test a manual strategy, that does not give you hindsight bias.
  • #20
  • Aug 7, 2017 4:03pm Aug 7, 2017 4:03pm
So do you guys/gals already built a inside candle RMI EA ? Is this the test results from it? I think its great that you are trying to modify the parameters to your specific desires.
0 traders viewing now