Pip Accumulator EA

[Ferenc86]

My backtest on EURUSD M5 is horrible, blow account, lost money.. Why? The settings are default.

[halcyonn]

soundfx..Thanks for this great contribution. As you say just what 11 needs to change mentality. Hopefully we can all work to make this a decent EA.

I started forward test on an ECN broker today, but error opening order : invalid stops. Is it possible to make this ECN compatible?

 

me too...same request!

[halcyonn]

My backtest on EURUSD M5 is horrible, blow account, lost money.. Why? The settings are default.

 

from what i have read and what i see in the backtest the number of trades taken in forward testing is 10 times more numerous than in backtesting...even if not ten times its multiple times more than in backtesting...the difference may lay there

[soundfx]

Guys,

 

Please ensure that you read everything thoroughly then we don't end up repeating the same old stuff as usual...

 

Note my comment in the first post:

 

Because the probability indicator which the EA uses operates over multiple timeframes, I don't think we can get accurate backtests, hence the EA will need to be forward tested.

 

Besides, the comment from Ferenc86 is a joke to be honest:

 

My backtest on EURUSD M5 is horrible, blow account, lost money.. Why? The settings are default.

 

Ferenc86 - Who told you that this was a "holy grail" EA that you could just backtest with default settings and life will be great? ;) Did I say that I've tested the EA for months on end with default settings and it works great? No, all I've said is that these default settings are an initial test and I've only run the EA for a day in forward testing. However the basis for the EA is sound in that the entries are based on a system being used successfully by manual traders and the risk/reward is 1:1, ensuring that we never get the gut-wrenching drawdowns seen by so many rubbish EAs these days which have 300 pip stops and 10 pip targets lol.

 

Guys, I hope I don't have to keep repeating this...

 

1. This EA can't be backtested with 100% accuracy. The reason for this is because the entry signals come from the master probability and that is in turn derived from real-time MTF (multiple timeframe) indicator values. If you read around on Google or wherever you'll start to see the problems which are encountered with MTF systems.

 

2. The EA is intended to be a "work in progress" and a collaborative effort by II members to create an EA which actually works and doesn't blow your account (for a change lol).

 

3. Everyone is invited to test, share their ideas for improvement, make code changes to the EA etc. I'm happy to help out as much as I can, though my prime interest is in manual systems and not EA's, so I won't be making loads of code changes - I expect everyone to help out on this. Maybe the idea will just fizzle out and everyone will revert back to what they're normally doing lol - we'll have to see.

 

4. The fundamental entry points come from a system which is currently being traded with good success by manual traders in the manual trading thread, so we immediately have an "edge".

 

5. I earn money in the markets from manual trading and not from running EA's. However I know that the manual system which gives rise to the signals in the EA is a good one. So...with a solid foundation like that we've all got a fighting chance of creating a decent EA with plenty of tweaking :)

[soundfx]

Ok - last post for tonight :)

 

You EA guys need to get out of this cycle of backtesting the same rubbish EA's all the time (without even thinking about the logic and how to improve it) and look at how to create an EA which can emulate a manual trader as much as possible.

 

You'll need to get your thinking caps on for this and actually look at price moving and indicators as if you were manually trading lol.

 

Let's take some great comments from psaini1973 in the manual trading thread who describes how his trades this system manually...I'll add in comments as to what we can think about in terms of the EA:

 

1) Trade only one pair. i consider EUR/USD as the Major of Major Pair. (Sometimes i trade EUR/JPY also if its in correlation with EUR/USD, but only after Japanese mkt close)

 

This goes without saying really that any EA worth its salt should be able to run and create profits on EURUSD.

2) Fix trade timing... i start my trading from London open to London close, i will only carry forward a trade after london close (but will close as soon as possible), will not take a fresh trade after london close.

 

Only trading in peak trading hours is a good idea because that's when the biggest moves typically are (i.e. UK and UK/US overlap from around 8am GMT to 4pm GMT). The Asian market is too unpredictable - sometimes price trends like crazy, though in my estimation this is only around 20% of the time on EURUSD, the rest of the time we end up in a range of around 50-70 pips.

 

3) Never enter in the middle of a signal given by the system, always wait for a fresh signal... if u missed 5 Min signal then look 4 fresh signal in 15 Min....

 

When the EA is first started up we need to ensure that it doesn't just jump into the first trade it sees.

 

4) After a signal is given dont jump in to it , look for recent high n lows and accordingly place ur trade.

 

We could make better use of ZigZag here in possibly only trading HH or LL breakouts after we get an arrow and probability is lined up - something as simple as this could be the secret to huge gains ;) Alternatively, we could create our own S/R levels or use and indicator for that or even use Pivot/Camarilla Lines etc. - all can be looked at to optimise the trades that the EA takes.

 

5) I personally never trade during or b4 the imp. news, only after it is over and its effect has subsidized.

 

There's an indicator kicking around called FFCAL which reads the forex factory calendar. It would be sensible for us to include a call to this indicator in the EA and prevent all trades within x minutes before and y minutes after the "Red" new items have been released - x and y can be parameters allowing folks to test and find the best values.

 

6) Discipline and patience is the key to success ( which i lack both of them but i m trying to attain both of them)

 

This in my view is the ONLY advantage that an EA has over a manual trader :)

 

7) Each day b4 starting ur MT4 platform check FF for imp. news and act accordingly.

 

Our "news module" can handle this...

 

8) This system is not Holygrail.. losses will happen, its the cost that we pay for trading forex,

 

Beware of trying to over-optimise the system - this typically happens when EAs are simple to backtest, and folks just keep on tweaking the parameters over the period they're running the backtest for until it give profitable results. Of course when you run it in live it collapses in a big heap lol. We will get losses and with 1:1 risk:reward we need to ensure that our trade accuracy is quite high to give us a good return. That's not the end of the story though because you can also look into using this system essentially as a "trend follower", in this case it may be possible to have risk:reward 1:2 or even 1:3 which means that we can lose 50% or more of our trades and still be in profit!

 

9) Entering a trade at right time is imp. but its the Exit that will make us rich or poor

 

Currently the EA default settings are totally basic. However there are various options for deciding when to exit. We could look into trailing stops (which in my view are a double-edged sword and rarely work as well as you think they're going to), we could consider scaling out of good trades or even scaling into good trades (whilst moving stops on earlier portions of the trade to breakeven etc.). We could always exit when we're close to a significant previous high/low resistance/support area on a higher level timeframe etc.

 

10) Be honest with yourself, share ur all results whether in loss or in profit, many forex experts are here to help u with ur trading. they will point out wat mistake u r doing.

 

The same applies with the EA, we need to see good and bad tests. Running the EA and just saying "the results are rubbish" helps nobody. Running the EA and saying "the results are rubbish"..." but I've discovered that these probability signals if used with SecretSuperIndi.mq4, the results are stupendous is totally different :)

 

Right now mkts are in a pathetic range. look at EUR/USD for the last 15-20 days , its going no where, directionless. if u could make some amount of money in this range bound mkt using this system, then think how much money u will make using this system when the mkt comes out of this range and follow a clear direction.

 

Trying to work out whether a market is trending or ranging has been the downfall of stacks of EAs in the past. Perhaps we could look at lowering targets/stops in ranging markets and increasing them with scale-in, trailing stop etc. options in trending markets...

And the way to determine if a market is ranging or trending. Just find a couple of MA's for the timeframe we're trading and look at what they do in ranging and trending markets. MAs which cross more than twice in the last say 20 bars could well be a perfect example of a ranging maket - just experiment with this stuff and see what you can find.

[vitamin]

hi sound fx,

try ichimoku cloud or any ma for filtering the signals..

[soundfx]

Hi vitamin,

 

try ichimoku cloud or any ma for filtering the signals..

 

Thanks for the suggestion, though exactly how am I (or anyone else) expected to do this lol ?

 

Ichimuko cloud has a zillion different ways of being interpreted. As for MA's well, they can cross and uncross a zillion times too - which MAs are the best? SMA, EMA. WMA etc. never mind the periods which you can use, or the timeframes which we should be looking at which work best...

 

I could list thousands of indicators and say "try this" without justification. I'm looking for a lot more on this thread.

 

As I mentioned before, I'm looking for everyone to pull together and "do their bit" to help develop an astoundingly successful EA. Unfortunately as a lot will be disapppointed to find out, this will require a lot of work and a lot of research and thinking on everyone's part.

 

I started this thread going, not with the intention of everyone coming up and spending a few seconds of their time saying things like "why don't you add this", " change this to make it work for me" etc. But with the intention of waking everyone up to the fact that here we have an opportunity to do something different - and we all have to play our part.

 

We have a system which works when manually traded and has good foundations behind it which will stand the test of time. To convert this into a successful EA needs a lot of research, testing and thinking.

 

The funny thing is that if this thread takes off, it will cover many pages with test results, improvements etc. and those who sit back, do nothing and just download the most recent EA without contributing and expect to make a fortune are sure to be disapppointed lol.

 

It's the participation in a project like this which gives you the awareness of what this game is all about to you personally. Some EA enthusiasts may well discover patterns in price which they can easily trade with pending orders, make their fortune and may never run an EA again. Others may take what is discussed here openly and develop their own secret money making EA (never to be seen again in public lol), some will steal the ideas, fake a backtest and sell the EA for $97 a time :). What I'm hoping is that for a change is everyone can pull themselves out of their fear of "giving away secrets" (there are no secrets in this game, it's all in the mind), and also sheer laziness in expecting someone else to do all the work for you and work together. The people who don't contribute are missing out on the development and will miss out on the specific usage/settings/pairs/timeframes of the EA which are best to use for them personally. The whole project is a voyage of discovery for all involved and those who don't participate have "missed the boat" - and even though they may obtain the boat at the end of it, though they've missed the amazing scenery along the journey which is what gives the boat it's value ;)

 

It's your choice...I'm offering you what is essentially the closest to a "holy grail" system most of you will ever see. Are you going to say "Nah...this is too much hassle, I actually have to do some work on this..." and continue backtesting and losing money in live on rubbish EA's for the next 10 years ? Or, are you going to make that huge step and actually get involved in some serious trading systems development research and testing?

Edited August 29, 2011 by soundfx

[conglo]

What a bunch of bull.

 

Over way too many years I've seen this way too many times.

 

 

Add this. Add that. Change this to that, and change that other thing while you're at it.

 

 

The original rules work, so stick to them and just tweak them a little.

 

Stop trying to turn this into something it's not.

 

This kind of BS eventually begins with every new decent method and you (all) always end up turning it into a LOSER.

 

 

Sorry, soundfx, you're doing the yeoman's work here but while the "committee" idea sounds good in principle it does not seem to work well in practice.

Edited August 29, 2011 by conglo

[soundfx]

Hi conglo,

 

I agree, it was an ambitious experiment which I don't think human nature (or at least that on II) is ready for yet lol.

 

Let's leave things for a while to brew and see what happens ;)

[conglo]

I'm with you. =D>

 

 

Hi conglo,

 

I agree, it was an ambitious experiment which I don't think human nature (or at least that on II) is ready for yet lol.

 

Let's leave things for a while to brew and see what happens ;)

Edited August 29, 2011 by conglo

[Diesel 10]

I like the idea of entering on a fractal breakout (after a signal), although I am not adept at programming to be able to incorporate this concept into the EA. Still the concept appears sound because it requires the Market to prove itself prior to entry.

[conglo]

I agree (imagine that..!) @-)

 

Fractals have enourmous potential.

 

I like the idea of entering on a fractal breakout (after a signal), although I am not adept at programming to be able to incorporate this concept into the EA. Still the concept appears sound because it requires the Market to prove itself prior to entry.

[musketeer]

Sorry mograst, I thought ZigZag was standard...

 

Here it is:

 

http://www.multiupload.com/ROCRB2MKMC

 

Hi fx4a,

 

Yes it is, we just need to ensure that orders are placed without stop or target, then afterwards modify that order with stop and target.

 

I'm still working on enhancing indicators for the manual system in the other thread, so I'm hoping someone else will be able to "take up the reins" on this one and look at including ECN compatibility and possibly handling MM like the DSMM spreadsheet.

 

zigzag is :)

 

here is the ecn/stp compatible - too early morning's task, later will be modify to get slippage in the account...

 

http://www.mirrorcreator.com/files/1WGLSNKG/

[soundfx]

Many thanks for entering into the the spirit of the thread musketeer :)

[soundfx]

For anyone watching the EA right now...

 

We're currrently in a buy trade because we have a green arrow and probability is > 20, however look at where the target is...it's above the resistance line where price has already bounced back four times already and on each bounce the move down was more than 13 pips, so the odds are that this trade will be a loser.

 

Coding in logic to ensure that we don't prematurely trade potential breakouts from resistance or support levels like this will filter out a good few bad trades. Price may suddenly break out after all those attempts, the resistance level should in theory be weaker - but who knows? we're not trading to guess, but trading to have an edge and the only way we can be sure that a breakout is likely to have occured is when a candle has closed with a higher high which isn't a spike.

 

Update...Well that trade worked out lol. Let's see what happens to the next one. This is part of the work we all need to do -actually monitoring the EA and its trades and identifying patterns where the EA makes good or bad decisions.

 

Currently we're entering trades continuously provided the signals line up. Should we be instead waiting for the last candle high/low at least to broken before entering again in the same direction?

Edited August 29, 2011 by soundfx

[musketeer]

here it is with small changes.... cosmetic only except counting slippage ...

 

1.Changed magic No., it shows symbol and ending with TF, ex.: 52119405 (5M TF)

2.Changed comment, added the requested price (one may track slippage), ex.: [email protected]

3.The slippage in modifying orders is counted, so the SL and TP must be at exact distance from opened price

4.The name is II_PIPA_EA-1_01.mq4 for easy tracking where I mess the things :)

 

http://www.mirrorcreator.com/files/HO3JRIIE/

http://content.screencast.com/users/myonlinecont/folders/Jing/media/c69e0d33-3768-48ed-82ff-66b6005f52cb/2011-08-29_0835.png

Edited August 29, 2011 by musketeer
picture worths....

[musketeer]

For anyone watching the EA right now...

 

We're currrently in a buy trade because we have a green arrow and probability is > 20, however look at where the target is...it's above the resistance line where price has already bounced back four times already and on each bounce the move down was more than 13 pips, so the odds are that this trade will be a loser.

 

Coding in logic to ensure that we don't prematurely trade potential breakouts from resistance or support levels like this will filter out a good few bad trades. Price may suddenly break out after all those attempts, the resistance level should in theory be weaker - but who knows? we're not trading to guess, but trading to have an edge and the only way we can be sure that a breakout is likely to have occured is when a candle has closed with a higher high which isn't a spike.

 

Update...Well that trade worked out lol. Let's see what happens to the next one. This is part of the work we all need to do -actually monitoring the EA and its trades and identifying patterns where the EA makes good or bad decisions.

 

Currently we're entering trades continuously provided the signals line up. Should we be instead waiting for the last candle high/low at least to broken before entering again in the same direction?

yes, absolutely

[PIPHORE]

For anyone watching the EA right now...

 

Currently we're entering trades continuously provided the signals line up. Should we be instead waiting for the last candle high/low at least to broken before entering again in the same direction?

 

I like the idea, it give better chance of success if the price break last candle (or last fractal) high/low. A break of a pip or two is good enough to enter a trade....imo

Edited August 29, 2011 by PIPHORE

[musketeer]

current market... is not so typical, since the london's bank holiday and seasonable low liquidity it is too manipulative

[soundfx]

Yep, I agree. I'm planning on watching and sitting on my hands rather than trading today unless things move a bit in the US session (though I've come unstuck trading holiday days before). Hence, if days like these are bad for manual trading - we should put them in our calendar and turn the EA off on those days.

 

Having said that this EA seems to holding its own quite well today :)

[musketeer]

yes, it is true :)

am running it on eu, gu, gj, uchf M1 to M60

let every TF/pair gets at least 40 trades it will worth statistically... the only pair that does not 13/13 is gj with 30/40

[jshear]

Hi Musketeer and Sound FX both thanks for your great work. I have on suggestion so far can a trade time setting be added so the ea can be set to trade only during london and NY hours.

 

Thanks

 

Jeff

[musketeer]

Hi Musketeer and Sound FX both thanks for your great work. I have on suggestion so far can a trade time setting be added so the ea can be set to trade only during london and NY hours.

 

Thanks

 

Jeff

 

Jeff, Soundfx already added it, just one time filter, but can be tested. Also, the analyzing sites are good enough to give such info even with 24 hrs trading. Also, one may do same via excel. Yes, I think these are the best times for it, but more data/statistics must show the truth.

[scorpion]

Hello, SoundFX , Musketeer, after reading some of the comments published

I would add that indeed you are right, and that many projects that have potential, they stay on the road.

New participants are incorporated into the thread, and do not devote enough time to read the thread from the start,

and want to be all done over, better even if you may gain a lot, with a DD low

And with that went on to describe some observations, just my point of view:

Option 1 .- When the operation starts, stop opening new orders if the signal does not meet a number of parameters XX, for example: opening only if the signal is positive and the indicator 5/10/15m probability still indicates <or - 20

Option 2 .- The EA simply open an operation when it receives the signal, once closed a new signal should expect fresh

 

I hope the explanation and the translation is understandable

[musketeer]

off topic....

 

Errors

 

The following errors occurred with your submission

 

 

• soundfx has exceeded their stored private messages quota and cannot accept further messages until they clear some space.
  

:) can you leave some space for messages?

 

Thanks!

 

:)