Forex Factory (https://www.forexfactory.com/forum.php)
-   Trading Systems (https://www.forexfactory.com/forumdisplay.php?f=71)
-   -   My "Ultron" EA for GBPUSD H1 timeframe (https://www.forexfactory.com/showthread.php?t=840339)

Zakattack Nov 1, 2018 10:20am | Post# 21

Pawel, When can you have the EA ready for M1 and M5? The EA is working great on H1. Here are my results for this week. Up 13% trading .1 lot size on a $1,000 demo account. {image}
If it is working great on H1, why would you want to have EA for M1 and M5. that will have alot of noise and false entries I am sure. Be happy and content with what you got, FOR FREE


Just a word of advice, risking 0.1 on a 1000 account seems to be too much (with about 50 pip SL) you are risking 5% of the account per trade ... just my 2 cents

Pawel.Gawron Nov 2, 2018 5:02pm | Post# 22

Hi Pawel, Great work! I like the simplicity. Basically, it is a mixture of going with the trend and momentum. I am testing ultron.ex4 live at the moment with 0.05 lots. I recognized that the first version ultron.ex4 is way better in backtest than the second version ultron1.2.ex4. Did you also recognize that? Cheers and keep up the good work
Paste this into a code to function

CheckForBuy and CheckForSell

if(Volume[0]>25) return;

Actually, it was necessary to have this because otherwise, it can open a trade in the middle of next candle which is bad. My mistake

Pawel.Gawron Nov 2, 2018 5:06pm | Post# 23

Pawel, When can you have the EA ready for M1 and M5? The EA is working great on H1. Here are my results for this week. Up 13% trading .1 lot size on a $1,000 demo account. {image}
This EA is not scalable to m1 and m5. I'm working on making some ea for a low timeframe, but it's just a "for fun" project so I'm working on it in my free time. I'm not sure if you will earn more on m1 or m5, it's easier to make something with good % of winning trades on the higher time frames. Also, Ultron is my first EA ever, so I'm still learning how to code in mql4, I have some ideas, but I'm not good enough programmer to make them real... yet

aminphandana Nov 3, 2018 2:11am | Post# 24

{quote} This EA is not scalable to m1 and m5. I'm working on making some ea for a low timeframe, but it's just a "for fun" project so I'm working on it in my free time. I'm not sure if you will earn more on m1 or m5, it's easier to make something with good % of winning trades on the higher time frames. Also, Ultron is my first EA ever, so I'm still learning how to code in mql4, I have some ideas, but I'm not good enough programmer to make them real... yet
Hi Pawel, thank's for your EA is good working...
if you have a lot of good ideas, I can help make the coding but your idea must really be able to generate profits, don't waste my efforts on doing it.

jamco388 Nov 3, 2018 7:28am | Post# 25

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....
how much minimum deposit required?
i will try it on demo account for 1 month

TheBestMA Nov 3, 2018 8:03am | Post# 26

.

northwind Nov 3, 2018 9:27am | Post# 27

{quote} Paste this into a code to function CheckForBuy and CheckForSell if(Volume[0]>25) return; Actually, it was necessary to have this because otherwise, it can open a trade in the middle of next candle which is bad. My mistake
Hi Pawel,
Subscribed. Thanks for sharing UltronEA with the community. Very KISS approach.
Two questions:
1. for the code change mentioned above, if(Volume[0]>25) return what lines numbers should it be pasted to?
Lines 30 and 39?
2. Should there be ; after return? So it would read if(Volume[0]>25) return;
Will run forward testing on demo acct.
Thanks.

Pawel.Gawron Nov 3, 2018 9:33am | Post# 28

1 Attachment(s)
{quote} Hi Pawel, Subscribed. Thanks for sharing UltronEA with the community. Very KISS approach. Two questions: 1. for the code change mentioned above, if(Volume[0]>25) return what lines numbers should it be pasted to? Lines 30 and 39? 2. Should there be ; after return? So it would read if(Volume[0]>25) return; Will run forward testing on demo acct. Thanks.

Here is a fixed version for those who don't know how to do it
ULTRON.mq4

Pawel.Gawron Nov 3, 2018 9:37am | Post# 29

{quote} Hi Pawel, thank's for your EA is good working... if you have a lot of good ideas, I can help make the coding but your idea must really be able to generate profits, don't waste my efforts on doing it.
It's hard to tell if they will work without hardcore testing, you know that on the market there is no "maybe". If I will have more data I will contact you Also I treat this "hobby" as a learning experience so it's not a big deal if I will have to learn more to write them.

ivosgarcia Nov 3, 2018 9:42am | Post# 30

Hello Gentleman, i am following the Ultron EA progress would like to test it live?

northwind Nov 3, 2018 10:22am | Post# 31

{quote} Here is a fixed version for those who don't know how to do it {file}
Wow, fast response. Thanks Pawel.
Since I want to test on multiple pairs, another change I made was to make the magic number user setable.
//#define MAGICMA 20131111
input double lots =0.01;
input int hour1 = 6;
input int hour2 = 21;
input int MAGICMA = 20131111;

Malfy Nov 3, 2018 10:35am | Post# 32

1 Attachment(s)
Pawel, there's a way to contact you? I have a very old EA that i created making mistakes with a builder online... It's a disaster but inside the disaster there's something great happening. I'm not able to code (that's why i was using a builder) and so i can't isolate and refine this little magic that i see in the behaviour of the EA. Can you analyze it and help me, maybe adding the idea to your EAs?

This is a 4 years backtest on EURGBP 2014-2017 with 10000 Euro capital and 2% risk... Totally unusable at the moment because it's something out of control. Anyway it can clearly shows what i mean. A mistake has lead to some very powerful trend following situations that, even inside a completely wrong EA can give this result with a winning rate of just 28.45%.

Click to Enlarge

Name: EURGBP 2014-2017.gif
Size: 10 KB

Sorry for the off-topic, I hope we can work on it

TheBestMA Nov 3, 2018 11:27am | Post# 33

thanks

robots4me Nov 3, 2018 4:39pm | Post# 34

Very nice -- thank you, Pawel.Gawron. Most of the EAs I download and test rarely perform as well as the author claims -- this one does. Given its simplicity -- really quite interesting...

I recognized that the first version ultron.ex4 is way better in backtest than the second version ultron1.2.ex4
I noticed the same thing. After poking around I found it was due to "lots = 10;". When I changed
input double lots =10;
to
input double lots =0.01;
or
input double lots =0.1;

Then the second version did perform better.

Also, I can confirm the current version of the EA does seem to work best with GBPUSD/H1 -- at least from back testing.

I'll experiment with MT4 optimization to see if anything interesting falls out.

Penguinthief Nov 3, 2018 5:33pm | Post# 35

Very nice -- thank you, Pawel.Gawron. Most of the EAs I download and test rarely perform as well as the author claims -- this one does. Given its simplicity -- really quite interesting... {quote} I noticed the same thing. After poking around I found it was due to "lots = 10;". When I changed input double lots =10; to input double lots =0.01; or input double lots =0.1; Then the second version did perform better. Also, I can confirm the current version of the EA does seem to work best with GBPUSD/H1 -- at least from back testing. I'll experiment with...
The 'input' definition means that this is changeable on your user interface anyway. 10 is just the default value that will show before you edit it for the first time.

jamco388 Nov 3, 2018 5:56pm | Post# 36

add auto lot funtion please

robots4me Nov 3, 2018 6:01pm | Post# 37

The 'input' definition means that this is changeable on your user interface anyway
I know -- I've been developing EAs for a few years. If you read my post, the problem isn't with using 'input' but, rather, using a lot size of '10'. That is HUGE -- and some brokers won't permit a lot size that big -- hence, the EA doesn't trade.

robots4me Nov 3, 2018 6:13pm | Post# 38

1 Attachment(s)
@Pawel.Gawron -- the more I experiment with your EA the more intrigued I become.

I've uploaded a modified version of the EA here:
ULTRON-3.1.mq4

A list of the changes I made:
1. Modified the format a bit to make it easier for me to understand the code -- hope you don't mind.
2. Parameterized the following settings: MA periods, lot size, hours, and magic number (taken from @northwind, above)
3. Removed the hard-coded '60' so that it can be used with any time frame

Using MT4 optimizer I found I could improve statistics even further -- e.g. profit factor > 2 and Win Ratio > 60%. Also, using the proper settings, the EA can be used with other pairs and time frames.

For those who are familiar with using MT4's back tester and optimizer I would encourage you to experiment a bit -- you might be pleasantly surprised. And for those who are not familiar with how to back test or optimize, then I encourage you to spend 5 minutes and learn how to.

Penguinthief Nov 3, 2018 6:14pm | Post# 39

{quote} I know -- I've been developing EAs for a few years. If you read my post, the problem isn't with using 'input' but, rather, using a lot size of '10'. That is HUGE -- and some brokers won't permit a lot size that big -- hence, the EA doesn't trade.
Your post makes no sense. What you're basically saying is that if I make an EA that defaults to 0.1, then make version b that defaults to 1.0, I can fool you into thinking that I've made my EA 10x better...

This '10' is adjustable in the inputs interface, and should be adjusted to the same value that the first version was for any comparisons. Infact, if you ran it the first time, changed this code then recompiled with the same name, it wouldn't even change the value for you.

robots4me Nov 3, 2018 6:25pm | Post# 40

@Penguin -- please learn to read.

I was responding to @CKobi's comment that the second version didn't work as well as the first. I diff'd the first and second versions and found that the hard-coded '10' was the problem.

The fact that my code snippets include 'input' is beside the point -- the second version of the EA posted by @Pawel.Gawron did not use 'input'.

Hey -- if you want to readjust the '10' to something smaller every time you run the EA then be my guest -- set it to '1000' if you like. My preference is for the default value to be something that works.

Gee whiz...


© Forex Factory