Overlay Positive Hedging EA

Thanks for taking the time to go through it all, the thread is quickly growing which is great. See below:

Okay, please read this post carefully, as I have dug into everybody's results and comments and am summarizing and then asking questions...
Dirty brown, on your first account detail post on this EA, it shows trades going both ways, buying EU, selling GU, AND...selling EU while buying GU, which is different than the other account detail posts and different than what Charvo was saying about the EA only selling EU, buying GU...and the results on that account summery as well as Paulss's first two days were incredibly good, and I mean unbelievable! Then on your last account summery you posted, closed trades profits were good, but there are some WEIRD open trades, looks like some where only single sell EU pair trades, which didn't appear to have their GU partner opened with them...
can you point these out to me pls?

also the account summery is difficult to digest because the order of the trades listed seem out of order, at least of time of opening trades...comparing it to the first account summery, which showed each trade opened both pairs at the same time...whatever changed there could be what all the growing DD and on closed trades are about???


One thing to keep in mind for the statement posted for the EA that's been running for a bout a week, that was the one with the 12 hr power outage on Monday so if you notice weird things on here that may explain it. I probably shouldln't have posted it, cause it will not be fair to the EA. I did make 1 change to the settings, after you recommended to burningmunke to change leveldistance to 150 from 100. I did the same... sorry!! I know you said I shouldn't, and that may also be confusing the interpretation.




Also, when you started the EA, did it immediately open that first trade? If not, how long did it take?

Yes, the EA seems to always take trades just as soon as its turned on.

I don't see that it opened trades as soon as others were closed, so I am not so sure about Charvo's theory on an initial test trade. Also, very important! If you count the number of trades that are still open, it's very obvious it's way more than the five "baskets" or layers that the EA should be allowing, that might explain some DD, too. We need to find out if this means it opens only five layers from one given original entry point, then may have another "original entry point" opened with it's own five layers??? If that's true, it needs to be limited to only one original entry point, with desired layers from that.

Digging further into this question. Let me try to summarize some of the code after spending some time reading it (realizing i'm no mql programmer but do have a programming background)..

This statement evaluates the Buy/sell, sellbuy preset variable:

Inserted Code

  if( Buy1Sell2 )  //boolean, so if =true, evaluates below code:
 
if( GetBuyOrderCount( Pair1, MagicNo ) <= 0 )
         SendBuy( B1S2, Pair1, ScaledLots, 1, MagicNo );
      if( GetBuyOrderCount( Pair1, MagicNo ) == 1 && GetSellOrderCount( Pair2, MagicNo ) <= 0 ) 
         SendSell( B1S2, Pair2, BaseLots, 1, MagicNo );

This says if there are no current open buy/sell orders..make them, this is proof of what I said that it immediately opens trades regardless of anything else besides no other trades being open. I see this as a major flaw, at the very least users first starting the EA need to be very aware of current market conditions.

Inserted Code

if( GetFloatingPip( BUY1SELL2 ) <= ( -LevelDistance ) && GetBuyOrderCount( Pair1, MagicNo ) < MaximumLevel )

This seems to check if current floating profit/loss is less than or equal to the pre-defined level distance AND the count of buy orders is less than (NOT less than or equal to) than the predefined 5, then do the below code. the fact that < is used instead of <= suggests to me first trades it places are level 1..
So in fact this means there is only 4 levels opened after the initial (bogus) trade that is done when the EA is turned on........problem you think?

Inserted Code

      {
         LstBuyTicket = GetLastBuyTicket( Pair1, MagicNo );
         LstBuyLevel = GetLastBuyLevelNbr( Pair1, LstBuyTicket, MagicNo );
         if( GetBuyOrderCount( Pair1, MagicNo ) <= LstBuyLevel ) 
            SendBuy( B1S2, Pair1, ScaledLots, LstBuyLevel + 1, MagicNo );
         if( GetBuyOrderCount( Pair1, MagicNo ) == LstBuyLevel + 1 && GetSellOrderCount( Pair2, MagicNo ) <= LstBuyLevel ) 
            SendSell( B1S2, Pair2, BaseLots, LstBuyLevel + 1, MagicNo );
      if( UseTakeProfitByPip == TRUE && GetFloatingPip( BUY1SELL2 ) >= TakeProfit ) 
      {
         LstBuyTicket  = GetLastBuyTicket( Pair1, MagicNo );
         LstSellTicket = GetLastSellTicket( Pair2, MagicNo );
         if( LstBuyTicket > 0 && LstSellTicket > 0 ) 
            CloseLastBuy( LstBuyTicket, Pair1 );
      }
      if( UseTakeProfitByUSD == TRUE && GetFloatingUSD( BUY1SELL2 ) >= TakeProfitInUSD ) 
      {
         LstBuyTicket  = GetLastBuyTicket( Pair1, MagicNo );
         LstSellTicket = GetLastSellTicket( Pair2, MagicNo );
         if( LstBuyTicket > 0 && LstSellTicket > 0 ) 
            CloseLastBuy( LstBuyTicket, Pair1 );
      }
   }

If anybody started the EA this last Monday between 10am and 3pm CST, just know the pairs were perfectly correlated, and I mean on top of each other that whole time, with no gaps at all...slow trading day, but I say all this because, that probably would have been a good time to have started the EA, on correlation...anybody who happened to do this at that time, please post your results since then, because it will greatly help us understand this EA!


I am not advocating that anybody try making EA trade EG, because it is directional, and as much as it is debated, there are pip size, scale size, and other differences, plus it's directional, not "hedged"...that last post was just a suggestion for back testing only! Yes, I have had some success with EG, but just look at a daily chart to see how much you would have lost holding a buy over last weekend! The Euro isn't coming back, at least not for a while.


1 min TF, hedging EU and GU, or similar correlated pairs has a successful, proven track record with many, many traders, they are using the Vitrite technique, of course.


Exactly, we can't lose sight of the simplicity.. and I'm afraid that is what's happening. As a parallel path to the current efforts on this thread, I am going to open a thread in the Platform Tech forum on calculating correlation with a free add on library to python called NumPy. I found an example in a book that gives the exact code to calculate correlation, covariance and all kinds of other neat stuff. I've mentioned previously the difficulty of performing true statistics in mql. Using Numpy, we can conceivably let python do the correlation and spit a number to the EA (via the 7bit Python -> MT interface) which can then be used as a primary trade filter, by calculating a correlation delta. When the gap starts to close, we should notice the delta of correlation over the last X bars to change by a certain percent. When the change hits a predefined threshold, a trade could be opened.

To me, this is a much more accurate way to quantitatively filter entries than just looking at pip distances. I started: http://www.forexfactory.com/showthread.php?t=361703



If we can get EA to replicate PaulSS's first two days to continue on............

Burningmunke has 14+%, now....he would probably have beaten my 19% gain for the week, with a small increase in lot size!....I don't understand his unclosed trades, either.


If your curious why a successful manual trader like me wants this EA to work?
I don't have much time to spend in front of a screen, either...plus, there are always 3-6 trades that I miss out on while asleep.

Me too, my job drains the life out of me .. don't exactly want to stare at a screen more than I already do. Hopefully this will make us rich someday, it will just take some effort and we will do it.

Thanks for your reply, dirtybrown, and first the 15 pip spread was just a suggestion, sorry, didn't mean to make you think you needed to do anything either way, I don't think that or the power problems matter as far as what I'm finding, which is:1. If you look at your first posted account summary, at the top the first trade (which means both pairs) was profitable, within ~3 hours, then next trade opened in the same minute (musta been moving 10 pips fast) but in the opposite direction, buying EU, selling GU, which means the EA wasn't doing what We thought, just selling (effectively, not literally) the EG only. It closed a lot later, but in profit.
2. The EA had four trades going, (8 actual trades, of course) before it closed the first one, which then it immediately opened another trade, all eventually profitable.
3. Looking at bottom of same account summary, there are 6 trades still open, that's more than five! I would blame power outage, but your other account summaries also have more than 5 trades still open, so EA isn't sticking to 5 levels, unless it's maybe 5 levels added to "bogus" trade.

I'm very thankful for your effort, as well as everybody trying to work this!

Now, we may have some contradictions here, but please, anybody running this EA, don't stop, if you have good profit, while it may seem illogical, the P/L doesn't lie!!!
Maybe for DD control, you could lower the max open trades some, but quick note on this...if your VPS can handle multiple instances, try to keep the EA you already have running, going as is, and make your mods to run on a different instance, so you can compare the old with the mod, doesn't hurt, and very educational...the programmer, medisoft, on baby pips was trying 6 versions of his own EA at same time and we were all surprised when the least likely criteria was the best, (and the simplest, funny, huh?) so don't throw away the free ATM machine!


The first results did tell us a story. And when introduced to correlation/stat arb, I was pretty skeptical, myself, until I tried it and, well,...
1. No losses
2. Profit speak for themselves!

I had the EA attached on both EURUSD and GBPUSD and had 19 trades open on Friday...

Firstly, I would like to thank you Dirty Brown introduce us this great EA then we have this opportunity to use it. This EA is great but maybe need some improvement or maybe we don't know it well.
This is my new test using demo account:
Broker: Hotforex
VPS: CNS (very fast connection with Hotforex)
TF: M1
Set:
Basic_Setting=----------------- Basic Setting -----------------
Lot=0.05000000
LayerDistance=150
UseTakeProfitByPip=1
TakeProfit=80
UseTakeProfitByUSD=0
TakeProfitInUSD=100.00000000
MaximumLayer=5
Pair_Setting=----------------- Pairs Setting -----------------
Pair1=GBPUSD
Pair2=EURUSD
Note1=Couple 1: Buy Pair1 Sell Pair2
Note2=Couple 2: Sell Pair1 Buy Pair2
TradeCouple1=1
TradeCouple2=1
ADR_Setting=------------------ ADR Setting ------------------
ADR_Day=365
UseManual_ADR_Ratio=1
ADR_Ratio=0.80000000
MagicNo=90000

you can view the statement http://www.myfxbook.com/members/paul...exm1cns/304067

Generally, the result is very good, I deposit $500 and have 24% profit in 2 days, but I found some strange trade and question.
1. Support by the EA rules, each trade are open parallel by 2, one is sell and one is buy at the same time. But I found some trade are open parallel by 3 trade, you can check it at the open tab and history tab.
2. I'm not sure the TakeProfit=80 is 80pip or 8pip
3. I would like to learn more about DD, then I can control enough fund on for the margin level when I open a real account.

But some of this trade eventually are profit.


Another testing already 1 week http://www.myfxbook.com/members/paul...dging-1/298913
Set file almost same above one.
Lot=0.02
LayerDistance=100
MaximumLayer=10
TF: M5

In both testing, profit is great but we need to learn more in control DD.
I believe this is a great EA and concept is good and safe.

Please Advise.

Would you mind if I pm you? I want to ask you a few questions about your manual trading, and I don't want to mess the EA thread with it. Thanks.

Anybody is welcome to PM me anytime!

Great results, Paulss's!

The only reason for the DD is the same thing we are trying to figure out:
You have 45 open trades, which is 22 (and 1/2?) "baskets" ...umm we need to figure out why that isn't being limited to 5 (10 total trades), like the EA says.
If somebody can figure out how to limit the total number of open trades, that's really the only recommended change from me, because your profits don't lie...it would be a bummer, though, if this EA got stuck on bad open trades maxed out at 5, for a long time...maybe this is meant to be used with smaller lot size and the open trades waited out until they eventually profit?
Nobody can afford for that to go on, I mean the longer the EA runs, the more open trades it leaves, making DD increase, maybe those are orphans, meant to have been closed at some time?

Thanks, yeah actually I've been using sigtrader for some time now its great. I use a computationally intense VSA template on the dollar index when manually trading any major, it works quite nicely.

As I post before that the EA sometime open 3 trade (not 2 trade), today, I found the reason is I open 2 chart (EU and GU) and run EA on both chart, I think this is the reason why it open 3 trade. Now I open 2 chart(EU &GU) and run EA in EU chart only. Up to now the trade are in normal.

By market close on Friday I still had some positions open for a DD of approximately 7% - and I decided to leave it to see what happens today.

Well, I'm happy to say that the EA managed to close out 3 levels of profitable orders and now I'm up almost 16%. I manually closed level 1 + 2 becuase I didnt think that the market would go much further in recouping those - so I closed them at -17 total. lol.

http://www.myfxbook.com/members/burn.../oph-ea/302314

On another note - an idea!

Not that it is groundbreaking in any way - but I had an idea on using this EA more effectively. It would require some work on the trader's side thou - making it not totally automated.

What I was thinking is that we manually check for the gap's. Once we see a gap of our predefined, desired size we launch the EA - but, with the correct settings, eg. GU has a gap of 20 pips above EU - set EA to place a short position on GU and long position on EU and hedge from there accordingly, lets say 5 levels. We either set TP's on every level as it currently stands, or if we want to stick to the babypips.com manual version of the system we only set a TP on the first level and once that gets hit close all orders (1-5) for profit. That way, we know that the gap has been effectively closed as it takes profit at the basic level.

Does any of this makes sense? I hope so lol - not the best at explaining things.

The downside of this way is that we have to ofc look for the gaps with the best God given indicator at our disposal - our eyes. From there, once we apply the EA we can go play golf or something (whatever you like - where I'm from we surf so... ). The EA should at this point trade just like we would have traded it manually..

Anyhow, just my 2 cents. Maybe this inspires people smarter than me to come up with something even better..

lol

Hello..

I have experience using this Overlay Hedging EA in 6 months...

After I use this EA 3 months...slippage can kill your account because orphans orders and too much floating...

This is is good and best for make money...

But this need modifiaciton:-

1- Add function check orphans orders..If one coulpes have only one pairs, so must orphans orders must delete for balance hedging...

2- Add functon Close All after profit target reach..example...If 20% close all for reduce floating...

Thanks U....

the two last posts I think make good sense, this EA may work great if started with the best timing, and if we can modify it to just take only up to 5 trades max at one time, no orphans. I might work good enough this way to start it once over once per day, at best time seen on charts.

Hi KingHigh and all,

many thanks for your patience in doing such a boring job.

I just came into this thread and used your translated version. I noticed a little bug in your order retry routine for Sell order. A return(Ticket); is missing causing the sell cycle to be excecuted TradeMgmtRetry times.

That shoud be the fixed code.


Regards,

skipperxit

Yes, you are right, the EA will create a lot of orphan order after some trade. this trade cannot close because they all have big DD.
The EA will open 4 orders on each point
Couple 1 - buy GU & sell EU
Couple 2 - sell GU & buy EU
the bug is in the Take Profit and close order. i.e we set TP=8pip and the total sum of the 2 orders of couple 1 reach 8pip profit. This mean when each Couple fulfill 8pip take profit level (Total sum of 2 order in each couple reach 8pip) then the EA will close these 2 order (couple 1). The point is here, it left 2 order (Couple 2) and these are the DD creator in the future.
If the EA can calculate both 4 orders (2 couple), and the total sum of these 4 order reach the TP level, then closed both 4 order. this will not create the orphan.

The concept is 4 order open at the same time, and 4 order close at the same time.

Could someone modify the code?

not sure if i understand u correctly

let's say the EA trade both directions, and it has:

pair group 1:
0.1 lot buy eurusd;
0.08 lot sell gbpusd;

and pair group 2:
0.1 lot sell eurusd;
0.08 lot buy gbpusd;

now @ TP = 8 pips, the EA should close one group.

you meant(?) the EA is supposed to close all 4 positions.

but i observed, when pair group 1 is in profit of 8 (or more) pips, the pair group 2 was never above water, how can the EA close it? the EA should take a loss on pair group 2?

IMHO, the EA should never have opened those orphan trades, they are backwards, aren't they?

DirtyBrown,

Please update post 1 so that it is correct.

Thx, KingHigh

what does "orphan" trade mean exactly?

from my experiences ( i have 1000$ exited profit, and -1000$ floating DD now), the EA just closed a pair of eurusd and gbpusd, if this pair reaches profit target.

The EA will never touch a pair of eurusd and gbpusd if this pair has NOT reached the profit target. (the EA actually trades without a Stoploss!!! )

where's orphan come from? are you saying one single position was left behind instead of a pair of positions? even so, that's not a major problem/cause for the DD.

charvo:

I have seen that you post several times regarding bi-direction test of EA.

1. I am 100% agree with you that current success of EA test is due to EG went down. If I am not wrong, everybody has certain DD on last Friday due to EG grow up. But it was solved because EG went down today.

2. But you will not get good result if you test in bi-direction because both side of trades conflick. IMO, you have to set either
TradeCouple1 = true;
TradeCouple2 = false;
or
TradeCouple1 = false;
TradeCouple2 = true;

It means that you have to know the trend of EG for this EA. If there are method to know the trend of forex, most of EA will be success and better that this EA.

3. I do not think this EA can get constant profit in long term excpet the you complete change the hedging method.

if you look at Dirtybrown's last account statement, all those open trades are orphan trades, some are just single pair, others were opened and never came to TP, probably opened in wrong directions...

It may be true about EG going down, but dirtybrown's first account statement looks different than the rest in that it has trades going the other way that profited, too????