Thoughts about an EA / Can we put this up from scratch ?

[xxl205]

Hi guys,

 

I am currently having some thoughts for a very simple (in my eyes) EA ... but I haven't found anything like it until now but maybe missing out...

 

I have a manual system called Trend Rider v3, which I use successfully right now. It is mainly based on EMA crosses for entry and exit in M5 and M15. I kept opening and closing my trades in the following manner, and this would be my requests for an EA :

 

1. Time Slot : adjustable, maybe 8 am to 6 p.m. GMT+1 or anything else, maybe 24/5 ?

2. Open 1 trade with adjustable lot sizes WHEN a certain EMA crossover apperars (Fast 4 Slow8 or so)

3. EXIT this trade when the REVERSE EMA crossover signal appears ... and OPEN again the reverse trade ...

4. SL adjustable, maybe 30 pips for GBPJPY or else pairs ...

5. Thus only ONE trade per pair open in the direction of the EMA crossover ! down = sell, next up signal, buy ! Nothing to care about the exit ... running 24/5 ... do I dream or could it be that simple ???

 

So to sum it up, this would go zigzag over the chart, pairs with lower spread are better off but hey, even with 5.5 pips I made around 20-50 pips on the GBPJPY rides without ever getting stopped out !

 

Last week I made 300 pips with this simple rules on GBPJPY M15 ... but I have a regular day job to do and can't trade that style anymore ...

 

It would be sooooo simple ... surf the cream out of each wave, get some crap too, but hey ... get the fat rides also ! My result came of +380 pips -80 pips neg = 300 total ...

 

So if any of you capable guys would be coding this EA suitable for all timeframes. I would load up that system to get analyzed. Or point me to an EA that already exists ...

 

Please help me to get us all rich(er) ^^

 

/xxl205

[Rio]

Re: Thoughts about an EA / Can we put this up from scratch ?

 

Should be a piece of cake to code.

 

I'd just need more info on exactly what EMAs we are talking about and perhaps a chart with examples so we can see how you would manually trade this.

[xxl205]

Re: Thoughts about an EA / Can we put this up from scratch ?

 

Hi Rio,

 

thanks for your reply ! Here is the package :

 

hxxp://[email protected]/download.php?4wmkxcmmmtn

 

pwd is middle part of this forums' URL ... ixxo-ixvest ... without the x of course :)

 

I have put the TrendRider3 into the archive, together with the wierdEMA.mq4 EA.

This EA nearly does the trick, but relies on MACD for entry and exit trigger, and some variables I cannot change cuz I can't code :((

 

For how I trade this : I stick to the TRider3 template and have set the EMA crossover fast 5 slow 8 to ring a bell (sound on in the options). That is all ;)

 

As I have gathered some more information, the EMA crossover indicator is maybe repainting. But for me I have always pulled the trigger on any signal instantly and opened the trades around 10-15 sec later.

 

So ... basing on this EMA crossover signal I'd like to use an EA ... it should work, simply but perfect. Look at those 300 pip waves on GBPJPY M15/M30 last week.

 

Now Rio, if you could kindly tell me your thoughts and analyse the package ... I would leave you a lot of Kudos for that, and surely some others too ... [-O<

 

/xxl205

[toritoo]

Re: Thoughts about an EA / Can we put this up from scratch ?

 

that pwd is not working

[Rio]

Re: Thoughts about an EA / Can we put this up from scratch ?

 

I can't open the package.

 

I've tried many variations of the middle part of the URL for the password and they all failed - except typing it all in capitals which then showed me that the archive is corrupt anyway.

 

Just use : indo-investasi (in lower case) for the password and find some way to reliably produce an extractable archive.

[saendee]

Re: Thoughts about an EA / Can we put this up from scratch ?

 

:-? can not work out all this pass word protection :-??the package open for me but can not open explorer url

[xxl205]

Re: Thoughts about an EA / Can we put this up from scratch ?

 

Sorry guys ...

 

http://[email protected]/file/aljzjmem0zm/TrendRider3.zip

 

Just replace the @ in the URL, this time I have used no password.

The old archive including pwd worked 4me nevertheless, dunno what there was.

 

I have now already found an EA which seems to do the things I like it to do, name is UniversalMACrossEA. Has all the settings I wished for, but had no time to implement it right now. It is also a bit older, I did not search for the right things before I guess.

 

Sorry for the fuzz ...

 

/xxl205

[Rio]

Re: Thoughts about an EA / Can we put this up from scratch ?

 

I got the package this time, although the trading strategy seems vague.

 

Also, I can get contact from the author of the UniversalMACrossEA, but even he wants some extra work done on it because it is not perfect.

 

Some more info on the trendrider strategy would be great, so i can code it into something.

After that I may work on the UniversalMACrossEA to see how it goes

 

After that there are SO MANY manual strategies out there that don't have EAs, I will be busy with this hobby for a long time #:-S

[xxl205]

Re: Thoughts about an EA / Can we put this up from scratch ?

 

ok, glad you have it now.

 

For the trend rider strategy : I used the only manual there is, the JPG inside. Told me how to use it, plus a bit of "feeling" :)

 

But mainly I trade according to the first indicator at bottom, which is the BBStop / RSI div / Stoch-Indicator. The other 3 down below confirm me in choice, but I do trade also 3 different colours of those bars once in a while, but never against the BBStopbar colour. I could try to describe more, but I am still a rookie in this buisness :-/

 

But thanks anyway until now Rio, your help is much welcome and appreciated ! ^:)^

[Rio]

Re: Thoughts about an EA / Can we put this up from scratch ?

 

Yikes, a bit of "feeling" to use when making trades? This is a robot I'm trying to create here....

It appears that there are multiple strategies going on here.

 

Ok Ok, if this Jpeg is all I have I guess the way to code this would be to collect as many signals from the indicators as possible and add up the signals into an overall "score" by which trades are taken. Then I'll add some inputs into the mix so you can filter in certain types of "scored" trades.

 

Say you only want to trade when certain indicators line up, but with or without other signals... for example. After that you can optimize it to your hearts content finding out what makes a good setup.

 

Seem good? :shand:

[xxl205]

Re: Thoughts about an EA / Can we put this up from scratch ?

 

Yikes, a bit of "feeling" to use when making trades? This is a robot I'm trying to create here....

It appears that there are multiple strategies going on here.

 

Ok Ok, if this Jpeg is all I have I guess the way to code this would be to collect as many signals from the indicators as possible and add up the signals into an overall "score" by which trades are taken. Then I'll add some inputs into the mix so you can filter in certain types of "scored" trades.

 

Say you only want to trade when certain indicators line up, but with or without other signals... for example. After that you can optimize it to your hearts content finding out what makes a good setup.

 

Seem good? :shand:

 

Of course that seems good ;) But yes, if you could code an EA for EXACTLY what the JPG says this would be "IT" ... if all the indicators say same colour (3x red / 3x blue) AND the RSI div / Stoch is in right mood, then go for the trade.

My "feeling" is of course based on the indicators :) we're not in the casino here where 50/50 is good enough ;)

 

I enter a trade in M15 general IF :

 

BBands (downmost indicator) colour must match direction (blue up, red down trend)

Stoch 8,3 must have shown top or bottom of a "wave" it has done

RSI div. must have at least touched or torched the Stoch wave (yellow, black) 1-2 times

 

Then it usually turns into my direction 80% of the trades. If a trade goes against me, I have my MM strategies to recover it, with some kind of martingale elements, e.g. GBPJPY goes against me means I open double lots each -35 to -40 pips, up to 3 steps. Additionally each step has a lowering SL, 1st 50 2nd 30 and 3rd step 15 SL, after that I take a deep breath and decide what to do. But that's another story, I am using this strategy since 2-3 months now and am quite happy with it, tripled my account this far...

 

So Rio, if you wanna go for this fine work, I won't hold you up :) and feel free to contact me anytime, skype & email also available.

 

Have a nice weekend so long

 

/xxl205

[Rio]

Re: Thoughts about an EA / Can we put this up from scratch ?

 

OK, great that helps I a lot.

 

I've been doing some thinking about the best way to replicate a certain manual trading "feeling" into a bot, as after all, us humans make judgments based on a number of positive indicators weighed up against other considerations.

 

What I am suggesting is that various indicators are given "weights", or a "score", so that perhaps a divergence is given a higher score than a stoch, or whatever. Eventually as we tally up the indicators, we reach an overall score... a little like having a poker hand... and we filter out trade setups below a certain score.

After that we can optimize to our hearts content... and we can optimize the indicator score weights as well as overall score to find the best scoring set up -- effectively "teaching" the bot how to manually trade.

 

I've never seen an EA coded like this, so this should be a fun experiment anyway

[xxl205]

Re: Thoughts about an EA / Can we put this up from scratch ?

 

Good morning,

 

yeah that is a very good idea ! As I said, I can mostly only explain something I do but not code it into an EA. Fun strategy to score those indis, cause then the weightening can be adjusted (to perfection? :)

 

It goes a little away from my "always in" strategy I intended to mean first, but I looked once more at the charts whilst trading. Thus I saw the real high lag of the 5/8 strategy, sometimes up to 40 pips until it paints an arrow. This is not very effective. I usually trade 10-20 pips faster in one direction than the arrows do...

[Rio]

Re: Thoughts about an EA / Can we put this up from scratch ?

 

Hi again, bumping this thread back up.

 

I've started work on it... so far only adding the heikin ashi bars into the indicator soup. I'm still looking at some of the other ones to see if they would be valid candidates for inclusion.

At the moment I am coming up with an EA framework that will let me include and remove different indicators so I can experiment with weighting them and finding good "poker hands" for trading. Essentially what I could do is continue to add to it, including more indicators until it has a plethora to choose from to execute trades... and once trained over actual historic data, it will have a good idea of how to play the market (I hope).

 

With any luck the EA will still be relatively fast, but I could see the whole thing becoming incredibly complex. That's OK if it becomes a "holy grail" of sorts though.

 

The indicator that tells the clock time of trading session looks interesting... and I'm thinking of adding it. Obviously, some pairs are traded more in certain time periods, so it could be an interesting filter... or not. Maybe it doesn't matter. Food for thought though. I'm also wondering about how to include data like the distance from high, distance from low, and distance from open... and how these variables could affect trading.

 

Yep, the 5/8 settings on the EMA crossover could be tweaked back to the default 4/8 settings if desired. Perhaps these values are worth optimizing.

[xxl205]

Re: Thoughts about an EA / Can we put this up from scratch ?

 

Yo Rio,

 

thanks a lot ! My goodness, you're talking about so many things here, but I think I've set you on fire for this one :D

 

Yes indeed, if we can frame a rudimentary running thing, we can optimize the settings for the "turning points". Surfing those waves with the least loss of pips on turning around the trades would be a traders dream, wouldn't it ? :) But we will not leave the main idea of being an always-in-EA ? We put all the indicators together for getting together the turning points ... but in perfection I would let that thing simply go zigzag for the next turnaround to maximise profits ... THAT would be the uttermost trader's dream ... and mine fullfilled ^^

 

I also give you the choice of a name for it ... I personally thought of something like "Dreamwaver", "Dreamsurfer", "Waves of Fortune" or whatever, I'll leave it up to you as it's 95% your baby ^^

[Rio]

Re: Thoughts about an EA / Can we put this up from scratch ?

 

I don't know about the being "on fire" part. I more or less see it as a continuation of what I've been doing with the Steinitz Fractal EA. Similar concept but taking on things I learnt from doing that (incomplete) system. Come to think of it, I could simply reinsert the indicators from that one into the mix and see what happens... but let's just get THIS system done first and then we can think about more tweaks... like adding MTF Trix indicators and what-not.

 

As for being an "always-in" EA. You are suggesting that (more or less) we follow the route of the EMA arrows. I think that's good up to a point, but during consolidation we could have a few issues with that. This is where having a score-based system could help. We could also differentiate between high-risk trades and low-risk trades and adjust the money management and risk to suit... or martingale when we lose... so that trades which don't sit well with the indicators could still be taken, while low risk trades could trade with more risk (say when all the indicators line up in one direction).

 

That's a hell of a lot to code, and then I don't know if it will even work. Oh well, Rome wasn't built in a day either.

 

As for naming the result... I was thinking along the lines of "Free Market Cowboy"... it plays poker hands of indicators and rides trends into the sunset or somesuch. I also have a name in mind if I ever create a holy grail.

I'll keep dreaming in the meantime. :hammer:

[xxl205]

Re: Thoughts about an EA / Can we put this up from scratch ?

 

I'll dream with you, as all of us do :)

 

Agree with the name, too :D and a cowboy's horse ride is as bumpy as the market sometimes !

 

So, I'll let you code, no pressure from my side here ... IF this comes out to work accordingly as it was planned ... well, a life long to enjoy left ;)

[Rio]

Re: Thoughts about an EA / Can we put this up from scratch ?

 

I'm making some progress on this. Bear with me, it's not easy going.

 

How am I going to use the RSI divergence indicator into the mix though? I don't fully understand the role in plays in making trading decisions.

[Rio]

Re: Thoughts about an EA / Can we put this up from scratch ?

 

OK, I've finally put something together here. Please note that this is still very much in a "beta" state, and should not be used live at all. In it's current state, it is not insanely profitable until it gets madly optimized -- and even then it may not bear any fruit.... but here it is in its current form.

 

UPDATE: It's moved further down the thread

 

The works by tallying up the positive indications it gets to enter a trade, and then takes it when the "score" goes over a certain threshold. At the moment, the different "weights" of the indicators are all set to 1, and the threshold for trades is set to anywhere above 6. Meanwhile there is a TradeWithdraw threshold set to a certain value, which given the EA a signal to exit the trade when it sees things going the wrong way.

 

(For example, the indicators give a score of +6, and since indicators 1-6 are positive, the bonus multiplier is added, putting the score at 7 and triggering a trade. When the score goes to less than zero and a certain value in the negative (as determined by the TradeWithdraw value) -- the trade is closed out.)

 

At the moment, the multipliers for the weights of the indicators are uneven, and the thresholds are as well -- so essentially this thing will not work until all the multipliers and thresholds are optimized for performance.

Still, I managed to get some profitability out of it, but nothing phenomenal.... yet. Optimization is very much needed. Ensure that money management is switched off when you do this.

 

Also note that there is a MoveBreakeven option which will move the stop to breakeven when the opposite side of the bollinger band is hit. Generally I found it not very useful in backtesting.

 

As for things missing/not working.... sometimes the ECN compatible setting of TP and SL after opening a position fails to work. Not sure why yet.

Also, there could be more score bonuses made available... but again, the more you add, the more optimizations needed, the more complicated it gets....

In addition during backtesting.... sometimes the signal to enter a trade comes too late, and the trade misses and flings the other way... not sure what to do about this yet.

 

You can drop the EA on any timeframe of any pair (you can control the timeframes it trades on with the internal settings.) ... but you will need the indicators in the indicators folder. The EA does not use all of them (for example, the huge RSI div, and the huge FerruFX indicator are not used). These could be added, somehow, but already we have a hell of a lot of indicators in the mix....

 

As for other future ideas... I could add higher timeframe confirmation as well, which would be an interesting idea.

Another idea I have is to apply proper trade closing rules, rather than rely on a score threshold to close, like the trade entry indications. Also a Break Even and Trailing Stop would probably be a good idea as well.

 

in any case, if you guys are interested you can help out by optimizing the multipiers over different pairs/settings. Hopefully we can find something useful out of this. It will need a lot of "visual backtesting" to see where it could be vastly improved, because I'm sure it can be.

 

As it stands, it is hard to make an EA when the trade entry rules are not precise. Maybe this will be more useful when optimized, but i'm going to have to set up my PC to run for a week or so to find out I think.

[saendee]

Re: Thoughts about an EA / Can we put this up from scratch ?

 

:shand: Rio is the template & indicators for FMCOWBOY EA or 2 systems :-?

[Rio]

Re: Thoughts about an EA / Can we put this up from scratch ?

 

Both actually. The EA only uses the BBandsStop, BarTrend, EMA crossover, and Flattrand indicators... so far. So you will need them in your indicators folder in order for it to work.

 

... but you have the other indicators and the template there which will be useful if you'd like to see some visual backtesting on it... or just use the manual indicator system.

 

UPDATE: Hang on a bit longer everyone.... I'm going to add the breakeven and trailing stop features... won't be a sec and I'll update the EA a bit.

 

UPDATE: Here we go again. The Breakeven and trailing stop features were hastily added... may not even work. I'll do some testing with them regardless. :-bd

 

http://www.4shared.com/file/153131994/57a82a65/FMCOWBOY.html

[Guest dave_rimmer]

Re: Thoughts about an EA / Can we put this up from scratch ?

 

Hi

I have a trading system below which works great on all majors normally i use M5 but others just as good!

I understand that making an ea from manual indicators is difficult but if one could make one on my system I am sure it will be a winner.

Basically when colours line up make the appropriate buy/sell trade..also if one could make an ea to have TS when the flattrend macd goes yellow 'flat market' and stops and closes and reverses on colour 'trend change'

 

Hope someone can help

 

http://depositfiles.com/files/tnkry93ly

[petewright]

Re: Thoughts about an EA / Can we put this up from scratch ?

 

Dave,

 

Thanks for sharing,

 

your template doesn't include two of the indicators that you have attached? mainly the one showing the trend with the arrows, can you repost the correct template or tell the settings for these.

 

Thanks!

[Cyrillic]

Re: Thoughts about an EA / Can we put this up from scratch ?

 

Hi

I have a trading system below which works great on all majors normally i use M5 but others just as good!

I understand that making an ea from manual indicators is difficult but if one could make one on my system I am sure it will be a winner.

Basically when colours line up make the appropriate buy/sell trade..also if one could make an ea to have TS when the flattrend macd goes yellow 'flat market' and stops and closes and reverses on colour 'trend change'

 

Hope someone can help

 

http://depositfiles.com/files/tnkry93ly

 

I'm going to demo test the system over a significant period of time to see if it worth coding will get back to you...

 

By the way thank you for coding this system Rio it was a very interesting experiment with indicator rating, i played around with it for a while but couldn't optimize it enough (too many settings and indicators). As soon as i get a break from my own crap I'll give it another go.

[Cyrillic]

Re: Thoughts about an EA / Can we put this up from scratch ?

 

Dave what times do you trade this system in? Euro session and US Session? Also could you describe entry and exit conditions keeping in mind that machine does not have a feeling when to get in and get out it need precise numbers or precise functions. So an average SL and TP would be appretiated.