A Proven Simple Strategy (2MAs, 1 RSI) | Page 5 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Renko trading using proven strategy 696 replies

Help Needed to Debug Existing EA with proven strategy 760 replies

Daytrading/scalping with high leverage - my proven strategy 8,734 replies

Looking for proven scalping strategy to program 1 reply

where is the "simple proven system" thread gone? 3 replies

Attachments: A Proven Simple Strategy (2MAs, 1 RSI)
Exit Attachments

A Proven Simple Strategy (2MAs, 1 RSI)

  • #81
  • Jan 20, 2016 5:02am Jan 20, 2016 5:02am
Quoting Soros
{quote} so far would you say this is a good strategy or not, judging from your experience with EA's and the results we have?
The strategy enters on retraces in the direction of the main trend. So in principle it is good.
However, there are 2 issues that can make or break the strategy.

1. Correct identification of the main trend. Using MA(200) is one way, but how reliable is it (on different time frames).
2. Correct entry on end of retracement. Using OB/OS levels is one way, but how reliable is it.

Few ideas to consider :

  1. Delayed Entry :

    1. option 1 : Instead of entering once RSI is above/below OB/OS levels, we delay the entry until we spot RSI divergence in the direction of the main trend
    2. option 2 : Instead of entering once RSI is above/below OB/OS levels,we delay the entry until RSI is out of OB/OS levels in the direction of the main trend

  2. Main Trend Identification:

    1. This is a big issue, and might require multi-time frame analysis.

No guts, no glory
  • #82
  • Jan 20, 2016 5:07am Jan 20, 2016 5:07am
As for modeling quality, this is a test with 90% I made on another PC.

This test using Exponential MA:

Attached Image (click to enlarge)
Click to Enlarge Name: 2ma1rsi_90%_ema.png Size: 56 KB


And this one using Simple MA:

Attached Image (click to enlarge)
Click to Enlarge Name: 2ma1rsi_90%_sma.png Size: 56 KB
No guts, no glory
  • #83
  • Jan 20, 2016 6:41am Jan 20, 2016 6:41am
I dont know why but how i get different result from abokwaik with the same setting :|

EDITED: wrong image, fixed
Attached Image (click to enlarge)
Click to Enlarge Name: Screenshot1.png Size: 82 KB
  • #84
  • Jan 20, 2016 6:54am Jan 20, 2016 6:54am
Quoting Sants
I dont know why but how i get different result from abokwaik with the same setting :| EDITED: wrong image, fixed {image}

Your Modeling quality is much lower than his, the reason why is most likely because of the quality of tick-data he is using compared to yours. I am guessing, if you were to DL better quality tick-data, your performance would be much different!

@abokwaik; thoughts??
When trading: "Empty your mind, be formless. Shapeless, like water."
  • #85
  • Jan 20, 2016 6:56am Jan 20, 2016 6:56am
Quoting abokwaik
{quote} The strategy enters on retraces in the direction of the main trend. So in principle it is good. However, there are 2 issues that can make or break the strategy. 1. Correct identification of the main trend. Using MA(200) is one way, but how reliable is it (on different time frames). 2. Correct entry on end of retracement. Using OB/OS levels is one way, but how reliable is it. Few ideas to consider : Delayed Entry : option 1 : Instead of entering once RSI is above/below OB/OS levels, we delay the entry until we spot RSI divergence in the direction...

Thoughts on adding an ADX indicator? I've been trading manually with one and it get's the job done!


SIDENOTE; Platinum is setting up for a sell sign on the H1 chart. Looking to short soon!
When trading: "Empty your mind, be formless. Shapeless, like water."
  • #86
  • Jan 20, 2016 6:57am Jan 20, 2016 6:57am
Quoting Sants
I dont know why but how i get different result from abokwaik with the same setting :| EDITED: wrong image, fixed {image}
Different Required Margin on EUR/USD.

Required Margin is determined by Leverage:
Required Margin = (Current Price/Leverage)*Standard Lot Units.
My leverage is 200, Required Margin on EUR/USD currently will be (1.0912/200) * 100,000 = 545
So when I set the variable lot Input to 6%, my starting lot was = (10,000*0.06)/545 = 1.1

It seems that your lot size was less because your leverage is lower.
If your leverage is 100. Required Margin on EUR/USD currently will be (1.0912/100) * 100,000 = 1009.
So you need to set the variable lot input to 12%, so that your starting lot size is similar to mine.

I hope it is clear.

Quoting Shemski
{quote} Your Modeling quality is much lower than his, the reason why is most likely because of the quality of tick-data he is using compared to yours. I am guessing, if you were to DL better quality tick-data, your performance would be much different! @abokwaik; thoughts??
Answer is above.
No guts, no glory
  • #87
  • Jan 20, 2016 7:01am Jan 20, 2016 7:01am
Quoting Shemski
{quote} Your Modeling quality is much lower than his, the reason why is most likely because of the quality of tick-data he is using compared to yours. I am guessing, if you were to DL better quality tick-data, your performance would be much different! @abokwaik; thoughts??
As far as i know, modelling quality is just a matter of each broker pricing model and it wont affect the whole tested strategy, isnt it? <-- not quite sure please correct me if im wrong.

I just download the history data from my own broker history centre and this is the result from December 2007
Attached Image (click to enlarge)
Click to Enlarge Name: Screenshot2.png Size: 139 KB
  • #88
  • Jan 20, 2016 7:03am Jan 20, 2016 7:03am
Quoting abokwaik
{quote} Different Required Margin on EUR/USD. Required Margin is determined by Leverage: Required Margin = (Current Price/Leverage)*Standard Lot Units. My leverage is 200, Required Margin on EUR/USD currently will be (1.0912/200) * 100,000 = 545 So when I set the variable lot Input to 6%, my starting lot was = (10,000*0.06)/545 = 1.1 It seems that your lot size was less because your leverage is lower. If your leverage is 100. Required Margin on EUR/USD currently will be (1.0912/100) * 100,000 = 1009. So you need to set the variable lot input to...

If you factor in spread, slippage, news etc...... would you say this strategy is profitable overall? remember backtests dont factor those things in.
I am what Many Dream to be but only a few can achieve, im a part of the 1%
  • #89
  • Jan 20, 2016 7:05am Jan 20, 2016 7:05am
Quoting abokwaik
{quote} Different Required Margin on EUR/USD. Required Margin is determined by Leverage: Required Margin = (Current Price/Leverage)*Standard Lot Units. My leverage is 200, Required Margin on EUR/USD currently will be (1.0912/200) * 100,000 = 545 So when I set the variable lot Input to 6%, my starting lot was = (10,000*0.06)/545 = 1.1 It seems that your lot size was less because your leverage is lower. If your leverage is 100. Required Margin on EUR/USD currently will be (1.0912/100) * 100,000 = 1009. So you need to set the variable lot input to...
thumbs up!! yes i forgot FXCM provide lower leverage, thanks its clear for the differ result.

Shall we proceed to discuss a far long term (2007 and lower) backtesting result
Attached Image (click to enlarge)
Click to Enlarge Name: Screenshot1.png Size: 121 KB
  • #90
  • Jan 20, 2016 7:10am Jan 20, 2016 7:10am
Quoting Soros
{quote} If you factor in spread, slippage, news etc...... would you say this strategy is profitable overall? remember backtests dont factor those things in.
Not without resolving the 2 issues I raised earlier.

Defining the major trend is not a simple job.
Most indicators depend on a number of previous bars, which are not of equal size.
Using range bars, or combining multiple time frames, for trend definition could help.

Another way could be to set the system to trade Only Long, or Only Short based on one's techno-fundamental bias.
No guts, no glory
  • #91
  • Jan 20, 2016 7:14am Jan 20, 2016 7:14am
Quoting Sants
{quote} thumbs up!! yes i forgot FXCM provide lower leverage, thanks its clear for the differ result. Shall we proceed to discuss a far long term (2007 and lower) backtesting result {image}
Problem is that you never know how the market will looks like in the future. With my setting it's profitable from 2007-2016. It's good. Of course, in 2007-2010 it could be better. But if we do optimization fot that years so it's not so profitable in 2014 and 2015. What is more important? Last two years or 8 years back? It's a question. My idea is to use 2-3 different setting on the same pair. All profitable from 2007-2016 but one will be more profitable in 2007-2010, one 2010-1014 and the last one 2014-2015. If you use it all together maybe it could work like hedging or something like that. What do you think about it?
  • #92
  • Jan 20, 2016 7:18am Jan 20, 2016 7:18am
Quoting abokwaik
{quote} Not without resolving the 2 issues I raised earlier. Defining the major trend is not a simple job. Most indicators depend on a number of previous bars, which are not of equal size. Using range bars, or combining multiple time frames, for trend definition could help. Another way could be to set the system to trade Only Long, or Only Short based on one's techno-fundamental bias.
For me as i learned to determine a major trend, i looked at D1/W1/MN1 candles and search for the latest ABCD pattern. Either it makes a zigzag to make a lower low and lower high (downtrend) or it makes a higher high and higher low. I think this kind of logic is tangible for coding.

For a minor trend, go to H4 timeframe and below and search the latest ABCD pattern( HH and HL) or (LH and LL). Once a pattern formed, its the trend!

Quoting dafa
{quote} Problem is that you never know how the market will looks like in the future. With my setting it's profitable from 2007-2016. It's good. Of course, in 2007-2010 it could be better. But if we do optimization fot that years so it's not so profitable in 2014 and 2015. What is more important? Last two years or 8 years back? It's a question. My idea is to use 2-3 different setting on the same pair. All profitable from 2007-2016 but one will be more profitable in 2007-2010, one 2010-1014 and the last one 2014-2015. If you use it all together maybe...
Of course we never know my friend, but "every cow barn will smell the same over and over until the cow is replaced". We are looking the "red cable" that flowing through time to time in pricing especially forex market. That red cable is the main thing why Big boys e.g big banks are earning money time to time and increasing in growth. We need to have a rigid setting to make sure it will survive in any financial issues (global crisis, WW3, doomsday are out of factors of course).

Some of my thought, would be glad to hear from yours
  • #93
  • Jan 20, 2016 7:30am Jan 20, 2016 7:30am
Quoting abokwaik
{quote} Not without resolving the 2 issues I raised earlier. Defining the major trend is not a simple job. Most indicators depend on a number of previous bars, which are not of equal size. Using range bars, or combining multiple time frames, for trend definition could help. Another way could be to set the system to trade Only Long, or Only Short based on one's techno-fundamental bias.
Sir i think we are over thinking things here and trying to create a holy grail system which we all know doesn't exist instead of creating a simple system that will experience DD but overall will be profitable which is key, well with that said, let me rephrase my question to make it a little more accurate, looking at the beautiful backtests posted here, do you think it would have made at least similar results during the same months if it was traded live?
I am what Many Dream to be but only a few can achieve, im a part of the 1%
  • #94
  • Jan 20, 2016 7:35am Jan 20, 2016 7:35am
Quoting Soros
{quote} Sir i think we are over thinking things here and trying to create a holy grail system which we all know doesn't exist instead of creating a simple system that will experience DD but overall will be profitable which is key, well with that said, let me rephrase my question to make it a little more accurate, looking at the beautiful backtests posted here, do you think it would have made at least similar results during the same months if it was traded live?
Definitely i agree with you! It costs me many $$$ to come a conclusion: a system with higher win rate does not guarantee a profitable system. All we need is a reasonable R/R and win ratio below 50% would be a sustain one
  • #95
  • Jan 20, 2016 7:43am Jan 20, 2016 7:43am
Each system will experience DD, nothing is 100%, all we need is consistency, a system that will experience some DD but get up and keep moving farward CONSISTENTLY, i dont mind loosing a few battles as long as i will win the whole war, so here we dont mind loosing battles as long as we will win the war, i thing changing and rearranging is the worst thing a trader can ever possibly do, always wanting to be right and over analysis, its a killer of progress, each time we experience DD we seek to optimize and change and in turn we shoot ourself in the foot because we inturn mess with the mathematics and the statistics, so we can never actualy be sure if something works or not because the variables within the strategies are always changing before we can even gather enough data for analysis, so all i wana know is will this perform similar to the backtests during a farward live tests, thats all that matters to me.
I am what Many Dream to be but only a few can achieve, im a part of the 1%
  • #96
  • Jan 20, 2016 7:53am Jan 20, 2016 7:53am
Quoting Soros
{quote} Sir i think we are over thinking things here and trying to create a holy grail system which we all know doesn't exist instead of creating a simple system that will experience DD but overall will be profitable which is key, well with that said, let me rephrase my question to make it a little more accurate, looking at the beautiful backtests posted here, do you think it would have made at least similar results during the same months if it was traded live?
With regard to the tested pair, with testing settings: spread, slippage, re-quotes would not change the results to a loss.
So if you ran the system on EUR/USD with my settings for 2015 you will end up with profit.

BUT, is that enough ?

In my opinion, (and this is applicable to any system), it is NOT.

This was IN-SAMPLE Testing, with only 42 trades over only 1 year.

Lets go out side the sample just a LITTLE bit, same pair, same settings, 2 years back.

This is the test for 2013.

Attached Image (click to enlarge)
Click to Enlarge Name: 2ma1rsi_2013_loss.png Size: 57 KB
No guts, no glory
  • #97
  • Jan 20, 2016 8:26am Jan 20, 2016 8:26am
Quoting abokwaik
{quote} With regard to the tested pair, with testing settings: spread, slippage, re-quotes would not change the results to a loss. So if you ran the system on EUR/USD with my settings for 2015 you will end up with profit. BUT, is that enough ? In my opinion, (and this is applicable to any system), it is NOT. This was IN-SAMPLE Testing, with only 42 trades over only 1 year. Lets go out side the sample just a LITTLE bit, same pair, same settings, 2 years back. This is the test for 2013. {image}
This is why I wrote my idea about bundle of settings. And runing it together. My settings I tested with data 2007-2016 and imho the results seems better. But probably we need someone with StrategyQuant or similar software because in MT4 there isn't possible to do this type of backtest.
  • #98
  • Jan 20, 2016 8:40am Jan 20, 2016 8:40am
Quoting abokwaik
{quote} With regard to the tested pair, with testing settings: spread, slippage, re-quotes would not change the results to a loss. So if you ran the system on EUR/USD with my settings for 2015 you will end up with profit. BUT, is that enough ? In my opinion, (and this is applicable to any system), it is NOT. This was IN-SAMPLE Testing, with only 42 trades over only 1 year. Lets go out side the sample just a LITTLE bit, same pair, same settings, 2 years back. This is the test for 2013. {image}

1351 mismatches on this bt
I am what Many Dream to be but only a few can achieve, im a part of the 1%
  • #99
  • Jan 20, 2016 10:37am Jan 20, 2016 10:37am
Quoting abokwaik
{quote} With regard to the tested pair, with testing settings: spread, slippage, re-quotes would not change the results to a loss. So if you ran the system on EUR/USD with my settings for 2015 you will end up with profit. BUT, is that enough ? In my opinion, (and this is applicable to any system), it is NOT. This was IN-SAMPLE Testing, with only 42 trades over only 1 year. Lets go out side the sample just a LITTLE bit, same pair, same settings, 2 years back. This is the test for 2013. {image}
Hello Khalil,

Is it possible to separate the faster moving average to ema(Highs) and ema(Lows)? So that the EA can exit longs at ema(highs), exit shorts at ema(Lows).

Thanks....
  • #100
  • Jan 20, 2016 11:00am Jan 20, 2016 11:00am
Quoting GoldTheHun
{quote} Hello Khalil, Is it possible to separate the faster moving average to ema(Highs) and ema(Lows)? So that the EA can exit longs at ema(highs), exit shorts at ema(Lows). Thanks....
Larry Williams style ;-)
0 traders viewing now