My EA Gift to Traders | Page 35 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Screen shot indicator free gift 0 replies

DECEMA - Xmas gift! 6 replies

Vegas 1HR GIFT: Big opportunity coming! 5 replies

Attachments: My EA Gift to Traders
Exit Attachments

My EA Gift to Traders

  • #681
  • Feb 27, 2015 3:21am Feb 27, 2015 3:21am
Quoting Dave_C
{quote} Alex It's going to take time to absorb your post but a couple of things I don't understand. By the 4th wager my understanding is that by the bet progression, the wager would be about .2. I don't really understand why your wager stalled at .13. Secondly I don't understand why you are embarking on a risk/reward of 1/3.
I understand totally that you can't make a decision in a few minutes after my post.

For the 4th wager no the progression will be 1.2 to 1.4 max . 0.1 * 1.3 = 0.13 and for the forth 0.13 * 1.3 = 0.169.

Look first order was 0.1 was loss after progression second order lot 0.13 . it closed at breakeven we sustain the same wager. third order 0.13 +20 pips. forth order lot increased according to wager 0.16.

The R:R 1:3 is important because it reduces the wager increasing progression. But as i describded when we will get stopped an order by trailing stop and gain some pips we will subtract them from our overall loss and reduce the R:R accordingly sustaining the same lot size.

The basic idea is increasing in firts place TP to a maximum level of R:R 1:3 and only after that start increasing lot. Because when we increase TP we don't increase our risks but we can't increase it unlimited times because it will become unrealistic and i believe that 1/3 is the maximum because beyond this number TP will never be hit.

Hope this helps , please don't hesitate to make some your addings because it is a ram idea still after some testing we will be able to speak in more specific numbers.
  • #682
  • Edited 8:33am Feb 27, 2015 4:49am | Edited 8:33am
Quote
Whenever you embark on a new strategy, it's best to press "Reset" (bottom right hand corner of expert properties).

I'll keep that in mind. The way I've been testing is running optimization at open prices to get a rough price, then I run that config in an every tick, and if the results show some promise I run an every tick optimization on that config. Then I test run that config on an every tick run just to make sure it's all good. Then at that point I post it here.
  • #683
  • Feb 27, 2015 2:43pm Feb 27, 2015 2:43pm
Quoting Shawn29
{quote} I'll keep that in mind. The way I've been testing is running optimization at open prices to get a rough price, then I run that config in an every tick, and if the results show some promise I run an every tick optimization on that config. Then I test run that config on an every tick run just to make sure it's all good. Then at that point I post it here.
That's fine Shawn. As long as when you embark on a new strategy you effectively zero out all the previous inputs by pressing the Reset button.

In the case that you posted, you had obviously used an older set file and some of the inputs did not match, which can result in mismatched values if extra inputs are added during an update. I'm trying to put the extra inputs at the end of the input list but some of these will eventually end up in money management and at that stage all the published strategies will have to be redone. (It's like inserting a field into the middle of a database table.)
  • #684
  • Edited 4:00pm Feb 27, 2015 3:09pm | Edited 4:00pm
Quoting alex8585
{quote} I understand totally that you can't make a decision in a few minutes after my post. For the 4th wager no the progression will be 1.2 to 1.4 max . 0.1 * 1.3 = 0.13 and for the forth 0.13 * 1.3 = 0.169. Look first order was 0.1 was loss after progression second order lot 0.13 . it closed at breakeven we sustain the same wager. third order 0.13 +20 pips. forth order lot increased according to wager 0.16. The R:R 1:3 is important because it reduces the wager increasing progression. But as i describded when we will get stopped an order by trailing...
Alex, I have to be honest, this is too complicated and falls into a specialized category. I will not be interrupting the logic flow on betting progression. It was your idea to set a minimum pip level and now you want to ignore it again and begin to vary the R:R. The concept as it stands is logical. If you are starting with 1:3 R:R you are defeating your own idea right from the beginning and that was to be in a position to be able to capitalise on a favourable trade. Do you realise 1:3 R:R means that you require a 300% return on risk. I suspect what you really need to say is 1:1.3 R:R that is reward is 1.3 times risk.

We have had this discussion previously (quote: "it closed at breakeven we sustain the same wager") and whilst your idea means well, it detracts from the basic premise of the EA. I give you the tools, how you use them is up to you.

I carefully explained to you how to do optimizations and that is the essence of how to use this EA. You should be optimizing StopLossMultiplier, TakeProfitMultiplier, SetStopLossToBreakEven, MinPipGain and all the other money management variables. In my previous post regarding this matter, I mentioned the 4,500 billion combinations and you acknowledged the extra homework required in this area. That is the correct way to use the EA rather that to build in layered rules.
EDIT: The latest build incorporates an adjustable trailing stop and it can be optimized also. You have more than enough tools to work with.

The code is free and you are welcome to use it as your own and to rewrite any section, but my mission is to keep it simple and provide the basic tools.

I would like to know what others think.
  • #685
  • Feb 27, 2015 5:52pm Feb 27, 2015 5:52pm
Please Dave, keep it simple! There are more than enough variables to find a workable strategy, if there is one.

The problem with using the back tests to find a winning strategy is that it is designed to "curve fit". You are correct in saying that it is practically useless to test over periods longer than a few months (3 or 4 at most). One way to tell if you have a strategy or a "curve fit" is to do a test for a 3 month period (make it the current 3 months), now test that strategy (same time frame) for a different three month period, and yet another three month period. If the first test is really good, but the other fail badly, you almost surely have a "curve fit", and the probably that the next three months are going to be good are slim. If one can find a strategy that gives good results over multiple 3-4 month time periods, that will be a winning strategy in the future. It doesn't have to be extremely profitable for every period, but should give a positive result or at least a near break even, and then have some periods that are really profitable.

I have been testing many of the strategies, and find the best, so far, to be the EMA cross, RSI, and the Continuous Straddle; but all are going to need a filter (which have not found yet) to be consistently profitable. Should I find that filter, I will certainly post it here.

One of the most promising strategies that I have found is using EMA Cross with Renko (offline) charts, has anyone else tried this. Only problem is, I cannot back test (or don't know how to). Forward tests using this strategy are very promising.
  • #686
  • Feb 27, 2015 6:19pm Feb 27, 2015 6:19pm
Hi Dave

I think the EA has more than enough options without making it any more complicated. If we cant find a good strategy with what we have now I don't think we ever will.
  • #687
  • Feb 27, 2015 7:05pm Feb 27, 2015 7:05pm
Quoting mwbarth36
Please Dave, keep it simple! There are more than enough variables to find a workable strategy, if there is one. The problem with using the back tests to find a winning strategy is that it is designed to "curve fit". You are correct in saying that it is practically useless to test over periods longer than a few months (3 or 4 at most). One way to tell if you have a strategy or a "curve fit" is to do a test for a 3 month period (make it the current 3 months), now test that strategy (same time frame) for a different three month period, and yet another...
Great advice. I am definitely not a fan of extended back testing and will include a reference to your post at Post #1. Curve fitting is one of the greatest hoaxes of commercial EA's.

I deleted a tongue in cheek post the other day where I suggested that the wording of Yelland statements was more significant.

I do remember somebody offering to assist with Renko bar back testing and you could Google that situation. If I can find the post or PM I will pass it on.
  • #688
  • Feb 27, 2015 10:08pm Feb 27, 2015 10:08pm
Quoting alex8585
Dave , First set file was tested on open prices from 2010 -2015. Second on every tick same period. The win ratio is extremely high thanks to BREAKEVEN. High drawdowns is due to breakeven are considered losers. I really think next what needs to be looked into is trailing stop options i am quite sure that we can get some better results with the combination of MinPipgain. {file} {file}
This is a great technical discussion over a very nice EA. I am totally with the multiple strategy options and suggest that we stick with them (RSI, STOCH, MA crossings etc.) according to my optimization in back-testing this over the three month period ending in the current month, I can say that the Profit factor is very high with the combination of two or more strategies to ensure a BUY/SELL is confirmed. I usually prefer STOCH turning confirmed with two MA crossings and RSI +70 or -20 with BBs taken into account.

Optimization as Dave mentioned can take forever if the Genetic Algorithm is not ticked, so in the main menu of properties for the EA, make sure u tick it or itwilltake forever to run thru all optimization scenarios (millions in that case) and this would be a waste of time.

Also, I note that, when optimizing, usually the first few scenarios are the best, u can check the optimization results tab while the test is running. usually the best comes first.

I agree with Dave 100%, the next step is to test it with one or two set files resulting from the best optimization cycle on real data, i.e. a demo account running on real tick data which is what I am doing now.

Thanks Dave.
  • #689
  • Edited 6:04pm Feb 28, 2015 3:49am | Edited 6:04pm
Quoting Tradezilla
{quote} This is a great technical discussion over a very nice EA. I am totally with the multiple strategy options and suggest that we stick with them (RSI, STOCH, MA crossings etc.) according to my optimization in back-testing this over the three month period ending in the current month, I can say that the Profit factor is very high with the combination of two or more strategies to ensure a BUY/SELL is confirmed. I usually prefer STOCH turning confirmed with two MA crossings and RSI +70 or -20 with BBs taken into account. Optimization as Dave mentioned...
It's taken about 500+ posts to get beyond , "I don't understand this" "I have a new system" "Could you consider this". Maybe this is the way it had to evolve.

I will tell you now, when I started this thread, I wanted better minds than my self to contribute. I knew I was going to get the questions without the homework. I am particularly grateful to the people who have challenged me to open up my mind.

This is your thread, brick bats or posies, to take the discussion on Money Management (thank you Alex) or other ideas to a new level. I know that there are over millions of traders out there but a huge silent majority. The users v's givers maybe 100:1 so I really thank the people who are taking the time to post.
  • #690
  • Edited 6:26am Feb 28, 2015 5:58am | Edited 6:26am
Here's my latest optimization attempt on AUDUSD 60min using RSI with pivot points filter. I do have a concern with this one though. I noticed it set the takeprofit to 0, but I'm not using the ATR filters at all. Did a back test from sept to present and the results were pretty good, just not sure if that compromises it or not.
Results screenshot and set file below.

Attached Image (click to enlarge)
Click to Enlarge Name: AUDUSD60test.JPG Size: 95 KB

Attached File(s)
File Type: txt AUDUSD-60min-RSI-Piv-Sept2014-Feb2015.txt   13 KB | 250 downloads


Here is another set from the same batch the results weren't quite as good but this one has the takeprofit at 50.

Attached Image (click to enlarge)
Click to Enlarge Name: AUDUSD-60min-RSI-Piv-Sept2014-Feb2015-2.JPG Size: 94 KB

Attached File(s)
File Type: txt AUDUSD-60min-RSI-Piv-Sept2014-Feb2015-2.txt   13 KB | 246 downloads
  • #691
  • Feb 28, 2015 8:41am Feb 28, 2015 8:41am
Quoting Shawn29
Here's my latest optimization attempt on AUDUSD 60min using RSI with pivot points filter. I do have a concern with this one though. I noticed it set the takeprofit to 0, but I'm not using the ATR filters at all. Did a back test from sept to present and the results were pretty good, just not sure if that compromises it or not. Results screenshot and set file below. {image} {file} Here is another set from the same batch the results weren't quite as good but this one has the takeprofit at 50. {image} {file}
Thanks for your contribution Shawn.

In your first file, because you have set the TakeProfitMultiplier at 1.5 and left TakeProfitATRPeriods at 0, TakeProfit is automatically calculated as 150% of StopLoss.
In your second file, even though you have set TakeProfit at 50, the TakeProfitMultiplier has overridden that and set your TakeProfit at 180% of StopLoss.

the following line of code refers
if (TakeProfitATRPeriods == 0 && TakeProfitMultiplier != 0) TakeProfit = StopLoss * TakeProfitMultiplier;

and this is an excerpt from the help file
TakeProfitATRMultiplier - The value of the multiplier to be used with above to
calculate Take Profit. If TakeProfitATRPeriods=0, this multiplier will be used
as a risk/reward multiplier and will be used to automatically set the size of
TakeProfit from the StopLoss value. Suggest a value of between 1 to 2.5 in this
case.
  • #692
  • Feb 28, 2015 11:56am Feb 28, 2015 11:56am
Quoting Dave_C
{quote} Alex, I have to be honest, this is too complicated and falls into a specialized category. I will not be interrupting the logic flow on betting progression. It was your idea to set a minimum pip level and now you want to ignore it again and begin to vary the R:R. The concept as it stands is logical. If you are starting with 1:3 R:R you are defeating your own idea right from the beginning and that was to be in a position to be able to capitalise on a favourable trade. Do you realise 1:3 R:R means that you require a 300% return on risk. I suspect...
In a way you are right . I was looking only for an EA to make the money management for me not the entries .
As i stated from my first post it is only my suggestion and i am fine if you don't want to follow me on that , it is accepted.

Wish you all the goods my friend and all the contributors.
  • #693
  • Feb 28, 2015 5:42pm Feb 28, 2015 5:42pm
1 minute data to the end of Feb 2015 now available at Post #1.
  • #694
  • Feb 28, 2015 10:26pm Feb 28, 2015 10:26pm
Hi Dave,

I'm attempting to set my stop and take profit using the ATR function and it's not working for me. This is the first time I've tried to use it. The attached image shows my settings. Could you tell me if they are correct?

Kind regards,
drofwarc
Attached Image (click to enlarge)
Click to Enlarge Name: ATR.png Size: 22 KB
  • #695
  • Feb 28, 2015 11:56pm Feb 28, 2015 11:56pm
Quoting drofwarc
Hi Dave, I'm attempting to set my stop and take profit using the ATR function and it's not working for me. This is the first time I've tried to use it. The attached image shows my settings. Could you tell me if they are correct? Kind regards, drofwarc {image}
You need to set TakeProfitATRPeriods and TakeProfitMultiplier only. (TakeProfit can be left at zero). If you are not optimizing, there is no need for the green tick.

With those settings, say TakeProfitATRPeriods=10 and TakeProfitMultiplier=2. The TakeProfit will calculate as twice the Average True Range over the last 10 trading periods.
  • #696
  • Edited 12:26am Mar 1, 2015 12:12am | Edited 12:26am
The example trading strategies have now been redone and also a new TradersGift Trading Concept has been posted at Post #1.

Included are 4 sets of examples optimized for the 7 Majors over the last 3 months. The optimized values have the green ticks.

1st set - 60 mins - EMA cross with RSI filter up to 5 bars go, SL Multiplier, TP Multiplier, SetStopLossToBreakEven, MinPipGain.
2nd set - 60 mins - MACD main signal cross with CandleSticks filter up to 5 bars ago, SL Multiplier, TP Multiplier, SetStopLossToBreakEven, MinPipGain.
3rd set - 60 mins - EMA cross, SL Multiplier, TP Multiplier, SetStopLossToBreakEven, MinPipGain and decreasing Stop Loss%.
4th set - 1min - Continuous Straddle, SL Multiplier, TP Multiplier, SetStopLossToBreakEven, MinPipGain.

The first 3 strategies have been combined for the TradersGift Trading Concept.

There is a small change to the CloseTrades EA.

You still need to set the StartingEquity and the TakeProfit and StopLoss values (values in your own currency) but the EA will automatically recommence trading with no manual adjustment at the beginning of the MQ4 calendar day using the same StopLoss and TakeProfit and the AccountEquity which exists at that time.
If you wish to adjust the StopLoss and TakeProfit, the inputs for those will have to be amended manually.

The Trading Concept will also be posted on Trade Explorer at the top of the page.
  • #697
  • Mar 1, 2015 12:28am Mar 1, 2015 12:28am
Quoting Dave_C
{quote} You need to set TakeProfitATRPeriods and TakeProfitMultiplier only. (TakeProfit can be left at zero). If you are not optimizing, there is no need for the green tick. With those settings, say TakeProfitATRPeriods=10 and TakeProfitMultiplier=2. The TakeProfit will calculate as twice the Average True Range over the last 10 trading periods.
It's still not working right. I'm wondering - do the settings you recommended also account for the stop? I also want to use ATR for the stop.
  • #698
  • Edited 1:58pm Mar 1, 2015 5:14am | Edited 1:58pm
Quoting drofwarc
{quote} It's still not working right. I'm wondering - do the settings you recommended also account for the stop? I also want to use ATR for the stop.
The help file explains this in full detail. Exactly the same considerations as for TakeProfit!
  • #699
  • Mar 1, 2015 6:18am Mar 1, 2015 6:18am
Quote
In your first file, because you have set the TakeProfitMultiplier at 1.5 and left TakeProfitATRPeriods at 0, TakeProfit is automatically calculated as 150% of StopLoss.
In your second file, even though you have set TakeProfit at 50, the TakeProfitMultiplier has overridden that and set your TakeProfit at 180% of StopLoss.

Thanks Dave, that actually cleared up a few questions I had, and for some reason that made a lot more sense to me then what was in the help file.

Anyways I liked the results from that 60 min test, but I noticed it did start to fizzle out at the end of last year and the start of this one. I'm running some tests using the same filters but knocking it down to a 15 min chart and trying a shorter more recent time period. I'll post the results when it's done (Had it going last night but forgot to check minpipgain so I'm starting it over)
  • #700
  • Edited 7:18am Mar 1, 2015 6:56am | Edited 7:18am
Dave ,

This one set file that i managed to get thanks to your advises on optimization. Here i believe it is time to let the EA to run live to see how it does in real market.

One thing i want to ask you even if it is considered as a favor . Is there big difficulties to make trailing stop adjustable and for testing and for live trading so we can set it after how many pips to be activated and after how many pips to move the stop. It will be much appreciated.
Attached Image (click to enlarge)
Click to Enlarge Name: Screenshot1.png Size: 11 KB
Attached File(s)
File Type: txt EUR_15_Jan_2010_Feb-2015.txt   13 KB | 321 downloads
0 traders viewing now