My "Ultron" EA for GBPUSD H1 timeframe | Page 6 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Broker Challenge "NDD/STP" vs. True "ECN" and "Mkt Mkr" broker talk 46 replies

Dealing with "Invalid Price", "Requote", "Server Busy" in MT4 3 replies

Timeframe inside current timeframe - is that possible? 2 replies

Seeing higher timeframe indicators on lower timeframe 4 replies

Finding the best "personal" timeframe! 6 replies

Attachments: My "Ultron" EA for GBPUSD H1 timeframe
Exit Attachments

My "Ultron" EA for GBPUSD H1 timeframe

  • #101
  • Nov 6, 2018 7:46pm Nov 6, 2018 7:46pm
Quoting aafx
One important thing: the results are not so impressive if the test is run on a 3y period ... only the last year give good results.
It doesn't make any difference how it traded 3 years ago -- you trade in the present (unless you have a time machine).

No EA trades the same over long stretches of time. An EA uses indicators that are simple algebraic formulas with one constant (sometimes more than one). That one constant is optimized for a particular period of time -- e.g. the last 4 months. Since market conditions change, that one constant should also be "refreshed" by re-optimizing the settings (i.e. the indicator constants) from time-to-time.
  • #102
  • Nov 6, 2018 7:50pm Nov 6, 2018 7:50pm
Can this EA be run on 2 digit and 3 digit pairs?

I can only test it on 5 digit pairs.
  • #103
  • Nov 6, 2018 7:56pm Nov 6, 2018 7:56pm
@bzolmajd -- it was nice that you took the time to make additions -- but you re-wrote the entire EA in your own image. I was curious to look at the changes you made but when I diff'd your version against the previous it was impossible to see what had changed because everything was different.

When you are sharing a piece of code with others it is important to keep in mind you are not alone. Furthermore, this is not your EA -- it is @Pawel's. It is fine to make incremental changes, but I can't use your code to make further improvements because you changed everything -- and I don't have time to analyze the side-effects.
  • #104
  • Nov 6, 2018 9:28pm Nov 6, 2018 9:28pm
@robots4me,,, I did not any changes in his EA, And after these years I know I can not do this , I only sorted the EA codes and put group of codes in different modules for better unbrestanding the job of EA. Thank you for your attention.
  • #105
  • Nov 6, 2018 9:40pm Nov 6, 2018 9:40pm
Quoting Pawel.Gawron
Working long-term EA for GBPUSD H1 in less than 100 lines of code. You can test it, use it, etc. By long term I mean profits are generated in long-term, because EA itself is a day trader. There is always max one trade active, always sl and tp, EA is active in day hours (at least in Europe, so make sure that you use timezone the same as for example Tickmill broker). Also my algo is very selective, so it's not that there is always active trade or something, but 58% win ration in the last 7 years with tp and sl 50 pips each is enough to make money....

Since you said there is only one trade active and the algo is very selective can I know how the trade is being open? what the requirement to make the EA open the trade position? sorry was quite new with that sort of EA.
  • #106
  • Nov 7, 2018 12:14am Nov 7, 2018 12:14am
Quoting robots4me
{quote} It doesn't make any difference how it traded 3 years ago -- you trade in the present (unless you have a time machine). No EA trades the same over long stretches of time. An EA uses indicators that are simple algebraic formulas with one constant (sometimes more than one). That one constant is optimized for a particular period of time -- e.g. the last 4 months. Since market conditions change, that one constant should also be "refreshed" by re-optimizing the settings (i.e. the indicator constants) from time-to-time.
true, therefore the EA should have several set of param that will use for a virtual trading and based on last x time period will adjust the entry ... or do a self discovery of best param to use, ... interesting project but very time consuming.
  • #107
  • Nov 7, 2018 12:23am Nov 7, 2018 12:23am
Quoting aa4x
therefore the EA should have a set of param that will use for a virtual trading and based on last x time period will adjust the entry ... or do a self discovery of best param to use
Yes -- settings that make a difference should be exposed as input parameters. Some are, some currently are not. But that is like one line of code and slowly, slowly it's getting there.

As far as "self discovery of best param to use" -- wow, that would be a true game-changer for all EAs that have ever been developed or ever will be developed. It's a nice concept, but no one yet knows how to do it -- so, probably won't be implemented in the next version of this EA.

MT4's strategy tester is usually used to determine the best settings for a defined time period. It's easy enough to use -- people who are not already familiar with it should take 5 minutes to learn how to use it.
  • #108
  • Nov 7, 2018 12:26am Nov 7, 2018 12:26am
Actually there was a try. EA used like 100 different strategies and kept track of each strategy's performance in the last X period of time. The ones that performed poorly was skipped by the EA
Necessity knows no law except to conquer
  • #109
  • Edited 1:59am Nov 7, 2018 1:48am | Edited 1:59am
Quoting 3endlessHope
Actually there was a try. EA used like 100 different strategies and kept track of each strategy's performance in the last X period of time. The ones that performed poorly was skipped by the EA
Yes -- a portfolio EA that includes many strategies but runs on a single chart. Very familiar with them -- see Forex Strategy Builder -- an excellent product (https://forexsb.com/).

However, in the case of this EA, when you consider it takes MT4's strategy tester only a minute or two to compute the best parameters for a given data horizon (e.g. 4 months) and you only have to do that once every month or two, then that seems like a more realistic (and practical) approach. Unless someone is really, really too lazy and doesn't mind losing money...
  • #110
  • Nov 7, 2018 1:54am Nov 7, 2018 1:54am
Quoting robots4me
{quote} Yes -- a portfolio EA that includes many strategies but runs on a single chart. Very familiar with them -- see Forex Strategy Builder -- an excellent product (https://forexsb.com/). In the case of this EA, when you consider it takes MT4's strategy tester only a minute or two to compute the best parameters for a given data horizon (e.g. 4 months) and you only have to do that once every month or two, then that seems like a more realistic (and practical) approach. Unless someone is really, really too lazy and doesn't mind losing...
Strategies were not optimised for a time period, but different market conditions.
Necessity knows no law except to conquer
  • #111
  • Nov 7, 2018 2:14am Nov 7, 2018 2:14am
I am running also EURUSD on live , 2 years backtest (99% quality, but still just a test) not bad. Is anyone able to adjust the EA in order to trade/test also Indices? I tried more FX pairs, but for now just GBPUSD and EURUSD are ok.
Attached Image (click to enlarge)
Click to Enlarge Name: Unbenannt.png Size: 25 KB
  • #112
  • Nov 7, 2018 2:53am Nov 7, 2018 2:53am
Quoting 3ndlessHope
Strategies were not optimized for a time period, but different market conditions.
Where did the data for the 100 different market conditions come from? The only market conditions that matter are the current ones, right? Why would someone use 100 market conditions that differed from the current one?

Also, how would one know which of the 100 strategies is the best performer without actually trading all 100?

What does make sense is to trade 100 strategies in a Demo (or micro Real) account and only allow the best performers to trade in your Real account. However, once you let that run for a month or two and then move on to the Real account, the market conditions may have changed again and the best Demo performer may not perform well in the Real account.

Quoting ea0setup
Is anyone able to adjust the EA in order to trade/test also Indices?
Using MT4's Strategy Tester and optimizer, you can determine the best settings for any pair, any time frame. However, this doesn't mean all pairs and time frames will perform well. This is simple and easy to test and something everyone should know how to do, and not expect others to do it for them.
  • #113
  • Nov 7, 2018 3:03am Nov 7, 2018 3:03am
Quoting ea0setup
Is anyone able to adjust the EA in order to trade/test also Indices?
Using MT4's Strategy Tester and optimizer, you can determine the best settings for any pair, any time frame. However, this doesn't mean all pairs and time frames will perform well. This is simple and easy to test and something everyone should know how to do, and not expect others to do it for them.

I do not expect anything from anyone.... If I ask is because I tested it already with the DAX and it doesn´t open any trade...got it?
Attached Image (click to enlarge)
Click to Enlarge Name: Unbenannt.png Size: 31 KB
  • #114
  • Nov 7, 2018 3:18am Nov 7, 2018 3:18am
Quoting robots4me
{quote} Where did the data for the 100 different market conditions come from? The only market conditions that matter are the current ones, right? Why would someone use 100 market conditions that differed from the current one? Also, how would one know which of the 100 strategies is the best performer without actually trading all 100? What does make sense is to trade 100 strategies in a Demo (or micro Real) account and only allow the best performers to trade in your Real account. However, once you let that run for a month or two and then move on to...
I didnt say 100 different market condition, but strategies. 100 is just a number, i cant remember exactly how many strategies were coded into the EA but it was around 100. EA kept track of performance of trading signals for all strategies without actually putting up an order for the ones that performed poorly in the past X time.
Necessity knows no law except to conquer
  • #115
  • Nov 7, 2018 4:14am Nov 7, 2018 4:14am
Quoting ea0Setup
If I ask is because I tested it already with the DAX and it doesn´t open any trade
Sorry -- you didn't mention DAX. I assumed you meant other forex instruments. I guess the reason for my sarcastic remark is because you'll notice that about half the questions asked here could have been answered by the poster if they were willing to spend 5 minutes to check for themselves.

Are you using the version of the EA that is hard-coded for GBPUSD or one that supports any symbol (i.e. the current Symbol()).


Quoting 3ndlessHope
I didnt say 100 different market condition, but strategies.
You said, and I quote: "Strategies were not optimised for a time period, but different market conditions. " So, to me, if strategies are optimized for different market conditions then it means you would need data for those different market conditions. Anyways -- let's let it lie. It's a nice idea.
  • #116
  • Nov 7, 2018 8:59am Nov 7, 2018 8:59am
Maybe useful indicator
Attached Image (click to enlarge)
Click to Enlarge Name: Ultron - Copy.jpg Size: 118 KB
Attached File(s)
File Type: mq4 Helper Ultron.mq4   9 KB | 1,183 downloads
File Type: tpl UltronIndi.tpl   26 KB | 1,031 downloads
  • #117
  • Nov 7, 2018 8:59am Nov 7, 2018 8:59am
Quoting Pawel.Gawron
{quote} Just remove hour parameter from this program. But if you will use it on Asian session then you will lose your money. Each pair has its own time, and GU is often too quiet during Asian session.
So the hour parameter is the hour1 and hour2 variable?. So I need to remove the line or need to set value to 0?
  • #118
  • Nov 7, 2018 9:33am Nov 7, 2018 9:33am
Quoting ea0setup
I am running also EURUSD on live , 2 years backtest (99% quality, but still just a test) not bad. Is anyone able to adjust the EA in order to trade/test also Indices? I tried more FX pairs, but for now just GBPUSD and EURUSD are ok. {image}
Can you post parameter seting for EU pair ? Thank you
  • #119
  • Nov 7, 2018 5:45pm Nov 7, 2018 5:45pm
Quoting manhthangvnn
Can you post parameter seting for EU pair ?
For any particular symbol and time frame there is no single setting that is best. When using MT4's optimizing feature in Strategy Tester you get hundreds of results, and you have to pick and choose from many possible (and similar) candidates.

It is best to learn how to do this yourself. It would take 5 minutes to learn and less than that to run an optimizing back test.
  • #120
  • Nov 7, 2018 8:23pm Nov 7, 2018 8:23pm
I agree entirely with robots4me. Just thinking out loud here; EUR/USD may be quite similar to the GBP/USD, in terms of how the EA may perform. This is due to how similar both trading pairs move throughout the typical daily trading session. It may in fact, perform similarly to how the initial Ultron EA was designed. I did not do any back-test, however. So...... Just my two cents.
0 traders viewing now