My EA Gift to Traders | Page 34 | 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

  • #661
  • Feb 25, 2015 10:18am Feb 25, 2015 10:18am
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.
Attached File(s)
File Type: txt TG_EUR_240a_Jan2014_Feb2015_EmaEvent_Candlesticks_RSI5Bars_SLMultiplier_TWMultiplier_BreakEvenFi.txt   13 KB | 257 downloads
File Type: txt TG_EUR_240b_Jan2014_Feb2015_EmaEvent_Candlesticks_RSI5Bars_SLMultiplier_TWMultiplier_BreakEvenFi.txt   13 KB | 244 downloads
  • #662
  • Feb 25, 2015 11:20am Feb 25, 2015 11:20am
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}
Alex, I looked at both set files and they looked identical.

Also I couldn't understand that the ticked items which I presume were the optimizations, had values in Column 1 "Value" which were outside the range of the optimization values. for example StopLoss = 200, optimization range 10 to 100 step 5. I have highlighted such values and wonder how you arrived at the value of 200. Same with TakeProfitMultiplier=4

Was the optimized date range Jan 2014 to Feb 2015 as in the filename or Jan2010 to Feb 2015? I couldn't verify the back tests because I don't have tick data back to 2010 but on tests with data 2014-2015, both graphs were unconvincing and showed losses.

Quote: "The win ratio is extremely high thanks to BREAKEVEN. High drawdowns is due to breakeven are considered losers."
I don't understand what you are saying. Maybe you could highlight with graphs/reports.
Attached Image (click to enlarge)
Click to Enlarge Name: 2015-02-26 02_44_39-TradersGift.jpg Size: 128 KB
  • #663
  • Feb 25, 2015 3:00pm Feb 25, 2015 3:00pm
Quoting Dave_C
{quote} Alex, I looked at both set files and they looked identical. Also I couldn't understand that the ticked items which I presume were the optimizations, had values in Column 1 "Value" which were outside the range of the optimization values. for example StopLoss = 200, optimization range 10 to 100 step 5. I have highlighted such values and wonder how you arrived at the value of 200. Same with TakeProfitMultiplier=4 Was the optimized date range Jan 2014 to Feb 2015 as in the filename or Jan2010 to Feb 2015? I couldn't verify the back tests because...
Look to be honest i really haven't made any use of the Start , Stop , Step in optimization i let them default. The only that i set is 200 SL and Takeprofit of 4 which means 800 pips and setting 200 MinPipgain in order the Stoploss multiplier stop when even 200 pips gain occurs.
I really until today had that impression for the EA and the settings , if you could explain what Start , Stop , Step stands for because the meaning i got but how they function probably not . A detailed example will be much appreciated.

Tested period was from 2010 -Jan to 2015 - Feb my bad in the file name.
The difference is only in the MaximumRiskPercentage and StoplossMultiplier .

What considers the Breakeven is something you had added and then removed . My suggestion was Breakeven not to be counted as loss but as neutral for the progression. In case of breakeven not to increase the lot but leave the lot size same. It just helps draw downs to be smaller and increase steps.

Alex
  • #664
  • Edited Feb 26, 2015 1:06am Feb 25, 2015 4:52pm | Edited Feb 26, 2015 1:06am
Quoting alex8585
{quote} Look to be honest i really haven't made any use of the Start , Stop , Step in optimization i let them default. The only that i set is 200 SL and Takeprofit of 4 which means 800 pips and setting 200 MinPipgain in order the Stoploss multiplier stop when even 200 pips gain occurs. I really until today had that impression for the EA and the settings , if you could explain what Start , Stop , Step stands for because the meaning i got but how they function probably not . A detailed example will be much appreciated. Tested period was from 2010...
The concept of optimization is running through a series of possible values to find the best combination of values

for example for EUR running an optimization of SL and TP may produce best result SL=60 and TP=100, AUD may be 40 and 75.
Whilst RSI period = 14,high level = 70, low level = 30 are the generally accepted defaults, each currency can have its own characteristics.

When you tick the box "Optimization" you are telling the Strategy Tester to cycle through all combinations of boxes you have ticked in Expert Properties and test all possible values in combination with each other, Since the boxes were ticked in your set file, I assumed that you had run the settings through the optimizer. That is why I could not understand the values outside the range of settings. For example in the TakeProfitMultiplier, ticking the box means that the when the optimizer is run, it will cycle through the settings 1.5 to 2.5 in steps of .1. Once all the settings are optimized, you simply double click on a row in the optimized results tab to transfer those values to your inputs "values" column and then when you can save those optimised settings including the range of values used for the optimization.

The ability to run 10,000+ optimizations in a couple of minutes is the essence of the optimizer. In your set file, with the boxes you have ticked the are 2,276,000 combinations of which the optimizer does a best fit scenario and narrows down the selection to 10,496 combinations using a genetic algorithm (testing tab default in expert properties) and gives the results in about 5 minutes per year of testing for your set file.

There is simply no point in picking values out of the air and that is the essence of the EA. If allows you to test thousands of scenarios in minutes. I ran a test the other day and there 4,500 billion combinations but the genetic algorithm narrowed that down to 10,496 combinations and the test took about 20 minutes. If I had deselected the genetic algorithm it would have still been testing next century.

You have highlighted the difference in the sets was MaximumRiskPercentage and StopLossMultiplier (sorry I didn't notice). Just using those as an example alone. If you just optimize those values and nothing else MaximumRiskPercentage for .2% to 2% step .1% and StopLossMultiplier 1.5% to 5% you have 19 * 8 = 152 possible combinations then add in stop loss from 50 to 200 step 10 = 16 combinations now were up to 16 * 152 = 2432 combinations and we haven't even started factoring the RSI combinations, MinimumPipGain and BreakEvenPips.

The point is if you don't optimize - you are "beating around the bush" that is wasting your time. Naturally your results need to confirmed with out of sample back testing, forward testing and tick data testing. This combined with any combination of strategies and filters (in combination with look back periods) give you many options to work with. There are fancier commercial programs such as StrategyQuant which give you genetic combinations but with TradersGift you have complete control with the combinations available and the EA is already written.

As I have mentioned previously, regarding the non counting of break evens in betting progression, you are suggesting that break evens be overlooked but +30's or 40's be counted (with the MinimumPipGain set at say +50). I'm not inclined to do that because the logic is inconsistent.
  • #665
  • Edited 11:56pm Feb 25, 2015 11:34pm | Edited 11:56pm
I have now incorporated the suggestion from Post #641 to include an option to progressively decrease the value of the stop loss as a trade moves further into profit. This may alleviate concerns of a good trade going bad and being hit by a trailing stop Loss. It may not apply to all scenarios and some of my back testing indicates that the fixed stop loss safety cushion is best whilst other tests suggest a scaling back of the stop loss value as a trade shifts further into profit.

If you are optimizing, I would suggest values of 0% to 2% step .1% (per 5% increase in value of the trade). The input is right at the end of the inputs so existing set files will not be affected. The default value is 0%, that is, a fixed stop loss value.

The other change relates to removing the EA from a chart where not only the info panel is removed, but also trade lines are removed. It helps clean up a chart after you remove the EA. The result is that if you need the trade lines and the info box displayed after a back test, you need to set the input 3rd from the end -
PanelVisibleOnExit = 1.

Finally a fix for the triple RSI problem discussed a couple of pages ago.

New download at Post #1
  • #666
  • Feb 26, 2015 1:18am Feb 26, 2015 1:18am
Quoting Dave_C
{quote} The concept of optimization is running through a series of possible values to find the best combination of values for example for EUR running an optimization of SL and TP may produce best result SL=60 and TP100, AUD may be 40 and 75. Whilst RSI period = 14,high level = 70, low level = 30 are the generally accepted defaults, each currency can have its own characteristics. When you tick the box "Optimization" you are telling the Strategy Tester to cycle through all combinations of boxes you have ticked in Expert Properties and test all possible...
Thank you very much for the explanations about the optimizations factors . It seems that i have a lot work to do this weekend , unfortunately for the next two days i will be away from my PC that is why i will do it during weekend.

About breakeven is totally acceptable your decision.
  • #667
  • Feb 26, 2015 5:24am Feb 26, 2015 5:24am
Hello Dave,

Thank you for your work on the Ea looks amasing. I am new Ea's trying to figure it out. I have a strategy I use already that am trying to put into Ea, 14 ma cross 34 ma using mcad 12,26,9 as a filter. I put these into ea but cant get it to trigger trades. I am sorry if am doing something dumb or missing something, am dyslexic and struggle sometimes. I have attached a set file of what I am using here so hopefuly can see the same. I really would appreciate any help in getting this to work. Thank you for your time and help. Once working would happily test and optimise across currencies and time frames and upload those set files for everyone else
Attached File(s)
File Type: zip maset.zip   2 KB | 216 downloads
  • #668
  • Feb 26, 2015 6:20am Feb 26, 2015 6:20am
Hi Dave,

I hope the question I posted didnt make you feel like that, have jsut found the thread and was genuinly struggling with it, am happy to contribute as stated before
  • #669
  • Feb 26, 2015 7:52am Feb 26, 2015 7:52am
mpope59 attached below i believe is what you were trying for. The results weren't good though on EURUSD 1 hour. Anyways you should read through the pdf on page 1 that might clear some things up for you.
Attached File(s)
File Type: txt masetfixed.txt   13 KB | 190 downloads
  • #670
  • Feb 26, 2015 8:04am Feb 26, 2015 8:04am
Hi Shawn

Thankyou
Really appreciate the help

Thanks again
  • #671
  • Feb 26, 2015 8:05am Feb 26, 2015 8:05am
Quoting mpope59
Hello Dave, Thank you for your work on the Ea looks amasing. I am new Ea's trying to figure it out. I have a strategy I use already that am trying to put into Ea, 14 ma cross 34 ma using mcad 12,26,9 as a filter. I put these into ea but cant get it to trigger trades. I am sorry if am doing something dumb or missing something, am dyslexic and struggle sometimes. I have attached a set file of what I am using here so hopefuly can see the same. I really would appreciate any help in getting this to work. Thank you for your time and help. Once working...
There is a very simple answer to your problem. You are using EMA cross and MACD cross both within MaxBarsAgo = 1. That is they have to occur simultaneously. You just need to extend the MaxBarsAgo in one of the scenarios out to say 5.
  • #672
  • Feb 26, 2015 8:13am Feb 26, 2015 8:13am
I didn't even think of the bars ago in that fix, I just saw he had the EMAPriceEvent at 1 and the EMACross at 0.
  • #673
  • Feb 26, 2015 8:18am Feb 26, 2015 8:18am
Quoting Shawn29
I didn't even think of the bars ago in that fix, I just saw he had the EMAPriceEvent at 1 and the EMACross at 0.
Yes EmaCross and EMAPrice are 2 different strategies.
  • #674
  • Feb 26, 2015 8:27am Feb 26, 2015 8:27am
Hello Dave and Shawn

Thankyou for your replies will give it a go, will also upload some screenshots to further clarify what I mean , I am terriable with numbers, hope that images may help

Thankyou once again

Mark
  • #675
  • Feb 26, 2015 8:34am Feb 26, 2015 8:34am
I guess while I'm here, here's my latest optimization attempt. USDJPY 30 min using EMA Cross and EMA Price Event. Back tested to Oct 2014 to present.
Results and set file below.

Attached Image (click to enlarge)
Click to Enlarge Name: USDJPY30MResult.JPG Size: 90 KB

Attached File(s)
File Type: txt USDJPY-EMA-30min-Oct2014-Feb2015.txt   13 KB | 239 downloads
  • #676
  • Feb 26, 2015 8:35am Feb 26, 2015 8:35am
Hi Dave and Shawn
here is screenshot of chart as using it now from this morning session, would illustrate what am doing and see what numbers need to put in to get this working as am using it well for scalping and longer term trends, may be of interest to others.

Thankyou

Mark
Attached Image (click to enlarge)
Click to Enlarge Name: MA_MCAD_cross.png Size: 49 KB
  • #677
  • Edited 10:07pm Feb 26, 2015 4:03pm | Edited 10:07pm
Quoting mpope59
Hi Dave and Shawn here is screenshot of chart as using it now from this morning session, would illustrate what am doing and see what numbers need to put in to get this working as am using it well for scalping and longer term trends, may be of interest to others. Thankyou Mark {image}
There is provision to close and reverse trades by setting TradeClosureMode to 1. If all conditions are satisfied for a sell signal with a buy trade on, the trade will be closed and a reversal trade will take place. In your situation if you are using both EMA Cross and MACD as factors with the MACDMaxBarsAgo set to around 20 a buy trade would have occurred at EMA Cross up and a reversal trade would take place with the EMA Cross down not at MACD cross down because the EMA Cross down would not have been validated at that stage. The EA covers the AND scenario but is not designed for the OR scenario (that is, "whichever is first").

You would need to set up 2 charts and use the EA in a type of hedging scenario where 1 trade cancelled out the other trade.

Looking at your chart, a much more logical approach is to simply use the MACD cross as a trade reversal (which occurred about 15 bars prior to the EMA cross) and you would have caught a larger portion of the upward move.
  • #678
  • Feb 26, 2015 10:26pm Feb 26, 2015 10:26pm
Quoting Shawn29
I guess while I'm here, here's my latest optimization attempt. USDJPY 30 min using EMA Cross and EMA Price Event. Back tested to Oct 2014 to present. Results and set file below. {image} {file}
Shawn, I had a look at your set file. It looks as though the inputs originated from an old set file because the captions do not align with the inputs. I have enclosed an updated set file with your inputs.

Whenever you embark on a new strategy, it's best to press "Reset" (bottom right hand corner of expert properties).
Attached File(s)
File Type: txt USDJPY-EMA-30min-Oct2014-Feb2015_2.txt   13 KB | 257 downloads
  • #679
  • Feb 27, 2015 2:07am Feb 27, 2015 2:07am
Dear Dave ,

During this morning when i was on my way to work i had a thought to make EA more conservative and reduce risks. I am not trying to beat the house just reduce exposure and drawdowns.

Before i start it is a raw idea and for sure it takes some optimizations

Lets assume that i open an first order lot 0.1 50 SL 150 TP. and order had been stopped out by SL so we have -50pips.
Second order 0.13 SL 50 TP 150. order closed at breakeven.
Third order lot 0.13 SL 50 TP 150 . order closed at +20 pips thanks to trailing stop.

Now our totall loss is -50 +20 =-30 pips

Forth order lot 0.13 SL 50 TP 90. As soon as we get 90 pips we will maintain R:R 1:3 . hit SL for the forth order.

Now we have -50 from the forth order by 0.13 lot and -30 pips by 0.1 lot.totally 50*0.13 + 30 * 0.1= 95pips is now our loss. but as soon as we can't sustain our R:R 1:3 level because out TP will become totally unrealistics. So we lower our R:R to 1:2 and increase the wager accordingly in case that TP is hit we have profit. Of course all the above is preferably to be calculated in percentage, that will make the calculation more easy. Lets assume now that forth order closed by trailing stop +100 pips.

Now we have 95 pips and 100 pips but 0.16 lots so it will be 160 pips = + 65 pips
Fifth order we start all over again.

Let me explain the basic idea as soon as we get some pips we subtrack them from our overall loss and adjust then TP and lot accordingly. Here the optimization feature could help enormously. I think this concept is totally in the spirit of this thread reducing our exposure and risks , and be far away of the martingale concept. We gonna avoid increasing wager unreasonably. Here comes of course and the need of adjustable trailing stop.

If you are intrested in evaluating something like this let me know because i am gonna definetely work in that direction even if it takes to make calculations manually.

Hope you find my concept really realistic and conservative to stay alive on the long term.
Waiting for your responce.
  • #680
  • Feb 27, 2015 2:28am Feb 27, 2015 2:28am
Quoting alex8585
Dear Dave , During this morning when i was on my way to work i had a thought to make EA more conservative and reduce risks. I am not trying to beat the house just reduce exposure and drawdowns. Before i start it is a raw idea and for sure it takes some optimizations Lets assume that i open an first order lot 0.1 50 SL 150 TP. and order had been stopped out by SL so we have -50pips. Second order 0.13 SL 50 TP 150. order closed at breakeven. Third order lot 0.13 SL 50 TP 150 . order closed at +20 pips thanks to trailing stop. Now our totall loss is...
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.
3 traders viewing now