MGH System (Martingale Grid & Hedging) | Page 49 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Simply "grid" system, no hedge, no martingale 9 replies

ECN && STP && Scalping && Hedging - Broker Review - Oct 2009 24 replies

Martingale, Reverse Martingale, Modified Martingale, Maths 1 reply

Simple grid manual system without martingale 8 replies

Grid, Burst Grid, Grid Risk and Grid Profit 1 reply

Attachments: MGH System (Martingale Grid & Hedging)
Exit Attachments

MGH System (Martingale Grid & Hedging)

  • #961
  • Nov 25, 2014 12:03am Nov 25, 2014 12:03am
Quoting Paracelsus
{quote} I have to assume that there was something wrong with the EA used for this test, as according to the original algorithm there should have been only 3 (or 2) open trades in one direction. On the screenshot 19 trades are closed at once...
The test is good, the EA worked properly on the test.
This was the v4. 2(.1) version, I think.

In this version if we reduce the number of max trades, the EA can't open enough positions to be able to close everything in porfit.
  • #962
  • Nov 25, 2014 12:08am Nov 25, 2014 12:08am
Quoting vaper
Test run: EUR/USD on M15 from 2014-07-07 to 2014-11-08 with basic settings in MGH EA_v4.2 where 2 settings were changed BaseLot = 1.00 Basket_TakeProfit = 1000 Here are the biggest moves so far .. it shows how wrong it is to stack buys or sells on the same grid-line. (no offence, just analysis)
No, the probelm is not "the grid on the same line". What's more: it is a very good idea, much more better than the other (previous) versions!
The previous versions weren't able to close its previous trades when the price moved for a long time in one direction. The balance was grown but the equity was reduced.
  • #963
  • Nov 25, 2014 8:06pm Nov 25, 2014 8:06pm
Test run: EUR/USD on M15 from 2014-07-07 to 2014-11-08 with basic settings in MGH EA_v4.2 where 2 settings were changed:

  1. BaseLot = 1.00
  2. Basket_TakeProfit = 250


Attached Image (click to enlarge)
Click to Enlarge Name: Test4_4-mnd_std-settings_Lots1-00_BasketProfit250_report.png Size: 93 KB



With these settings the test passed with profit in 4 months = 10.4 %.
Profit factor = 1.11 is too low because of all those extra trades on the same lines.
cheers
vaper
  • #964
  • Nov 26, 2014 6:30pm Nov 26, 2014 6:30pm
hi all,

apologies but can i check regarding the original strategy posted by kfx?

how do you handle a situation where the inside grid is triggered, but it did not complete the course upwards to clear the initial outside grid. thus resulting to something similiar as below:

3 x 1 buy
3 x 2 sell
3 x 4 buy

and subsequently, it just continue to trend downwards.. with that scenario, u would then be opening 0.8 sell orders. but the 3 x 1 and 3 x 4, would be a total of 1.5 buy in the opposite side.

also, in the same scenario, when would u restart inside grid again?.. based on the lowest 0.4 buy order and highest 0.8 sell order?

or should i implement a stop out for this unfortunate set up since it is likely to put the account into losses rather than profit?

thanks for your time in replying this!
  • #965
  • Nov 27, 2014 6:08am Nov 27, 2014 6:08am
@belgeran

Hi , for the very original method you could read from post #1 to here. @Kfx was doing a good job together with all who helped along until .. well, you will read that.
He did have some scenario pics (don't know the posts) where you can study his intentions and of course in post #1 .

Personally I am testing all the time and cannot understand why so many trades are opened on the same grid line over & over again. This cannot be the correct way. So, if you have good idea's about how to change the progression of the baskets and the trades in it, we (all who liked Kfx methods) can try to change that, maybe even with the help of @SwingMan who is a seasoned coder.

My idea is that every batch (basket) should go into a separate full hedge when going the wrong way after a certain number of trade entry's. That way every basket becomes an object in itself with a object-DD. Next basket would start with a slight higher lot size good enough to take one or more hedged ones along to profit-all or zero-all. If you have seen my tests then you noticed that trades go to many hundreds of lots which is absurd.

Your idea ?
Happy Trading, cheers
  • #966
  • Nov 27, 2014 9:25am Nov 27, 2014 9:25am
hi vaper,

thanks for the reply. his post #1 didn't have that scenario covered, but nonetheless.... i have coded an EA that runs specifically to his trading methodology. now back testing it.

in regards to your idea of not having so many trades opening on the same grid lines.... in kfx's original method, his hedging orders will cover the same grid line and it happens in my EA as well. I tried swingman's EA but i felt it had differed from kfx's original method so i gave it a miss.

my feel, is that kfx's method works, but i may need to reduce the baselot to increase account survival chances and probably diversify to trade other pairs at the same time. it remains to be seen if that would actually work as well, since a large base is needed for a martingale hedging account. just testing on eurusd takes 50% drawdown.
  • #967
  • Nov 27, 2014 12:45pm Nov 27, 2014 12:45pm
Hi belgeran , good work man. I 'd like to do some tests too, can you send the ea plz?

And look at my former post http://www.forexfactory.com/showthre...12#post7893112 where I suggest to open the first basket with hedged trades.

Happy Testing!
  • #968
  • Nov 27, 2014 6:57pm Nov 27, 2014 6:57pm
Quoting vaper
Test run: EUR/USD on M15 from 2014-07-07 to 2014-11-08 with basic settings in MGH EA_v4.2 where 2 settings were changed BaseLot = 1.00 Basket_TakeProfit = 1000 Here are the biggest moves so far .. it shows how wrong it is to stack buys or sells on the same grid-line. (no offence, just analysis) {image} The largest buy is at 1,000 lots !! and largest sell at 622.8 lots !! Details: {image} What I could observe is that many times the position goes into profit, long before the goal is reached. In this test the running P/L went to +300 to +500 easily....
I have always wondered about the propriety of re-doing grid levels once they are filled, which would result in weight being placed at one particular level. Naturally, in a strong trend, this won't make much difference in the long run, but in the short run could result in large drawdown if price corrects and moves against your position.
Fireworks are fun ... as long as you don't blow your fingers off.
  • #969
  • Nov 28, 2014 1:51am Nov 28, 2014 1:51am
My testing MGH, with 100e acount, is working good..
Attached Image (click to enlarge)
Click to Enlarge Name: Dibujo.JPG Size: 59 KB
  • #970
  • Nov 28, 2014 11:08pm Nov 28, 2014 11:08pm
it is good actually, until a very strong trend hits it all the way. e.g. the USD/JPY jump from 106 up to 113 etc.

the weakness of this strategy lies in the point where it doesn't ride on winning trades but hold onto losing trades... as long as market does some form of consolidation, you will probably come out unscathed, but if the market rides strongly, u will be carrying bigger losing trades compared to the winning trades.

my testing of the EA in EUR/USD and USD/JPY, both resulted in account bankrupt.
saying that, i was doing a 0.1 baselot on a 10k account.... using 0.01 works fine though.... hits up to 128x.
  • #971
  • Dec 4, 2014 3:36am Dec 4, 2014 3:36am
Hy ... is this amazing project stopped? No any news no any update ... what's wrong with it?
My forward test continue ... Here the link
Kind regards
https://www.myfxbook.com/portfolio/mgh/1082277
  • #972
  • Dec 4, 2014 12:31pm Dec 4, 2014 12:31pm
Quoting cescof
My forward test continue ...
I also live ...

Today was a very interesting day, and the EA has perhaps done everything right. But I have to analyze the orders because not even on M1 we can see what was traded.
Attached Image(s) (click to enlarge)
Click to Enlarge Name: EURUSDM15.png Size: 25 KB
Click to Enlarge Name: EURUSDM1.png Size: 24 KB
  • #973
  • Dec 4, 2014 2:22pm Dec 4, 2014 2:22pm
Quoting SwingMan
{quote} I also live ... Today was a very interesting day, and the EA has perhaps done everything right. But I have to analyze the orders because not even on M1 we can see what was traded. {image} {image}
MGH works great on trends and when there is a lot of volatility ... the problem comes when it comes in a range of 30 to 40 pips ... the other day ruined my account 100e demonstration ...: p
  • #974
  • Dec 4, 2014 3:41pm Dec 4, 2014 3:41pm
Quoting SwingMan
{quote} I also live ... Today was a very interesting day, and the EA has perhaps done everything right. But I have to analyze the orders because not even on M1 we can see what was traded. {image} {image}
Technically perfect but those chain-trades will not be accepted on a real account . . the market was way too volatile on that event. And if some trades would be accepted it would be with large slippage.

Q: suppose these entries on a grid-line : Sell @ 56-lots and Buy @ 32 lots and a new trade is to be entered. Is it not possible to close those 2 mentioned for 30 lots both sides? The difference of 24 lots would be the same but with Sell @ 26 lots and Buy @ 2 lots. (all numbers are examples)
Could it be calculated just before a new trade on the same line would be entered?
If so, then the lost spread costs could be taken on board of the new trade.

just wondering if that can be done
  • #975
  • Dec 4, 2014 4:24pm Dec 4, 2014 4:24pm
Quoting vaper
Technically perfect but those chain-trades will not be accepted on a real account . . the market was way too volatile on that event.
I wrote that the trades were live. One can see in my profile.
Because of the rapid high volatility, there are many strange orders against orders in normal volatility.
  • #976
  • Dec 4, 2014 6:07pm Dec 4, 2014 6:07pm
Quoting SwingMan
{quote} I wrote that the trades were live. One can see in my profile. Because of the rapid high volatility, there are many strange orders against orders in normal volatility.
sorry SwingMan, I must have missed that! (live on a demo or on a real ?)

Those strange trades are exactly what i meant: some are taken by the server, some are not. And if opened it would be with large slippage. This could be solved by opening only on next candle (always).
But I understand that then sometimes the grid level will be missed, hence no trade at all. It all depends on how it's programmed .. missed a trade -> just go on and wait till next .. or keep trying until trade is opened etc.. there are a lot of logical options to work around unexpected super-volatility.

Since it is an automatic approach, every possible obstacle must be dealt with.

Tnx again for making me aware of missing the live part, cheers
  • #977
  • Edited 9:12pm Jan 4, 2015 8:58pm | Edited 9:12pm
Very good idea and work. Thank.

I ran EA on strategy tester. 10 pips grid, EA got stop out on 19/11/2014, may be Fed minute news on Wed, due to 10000 deposit. I change to 50000. EA passed. So, 19/11/2014, M15, would be a outlier case.

I ran 20 pips grid, EA got trouble on 4/9/2014 to 17/9/2014, a long period of small range, last sell lots is 27.68. It is too big from 0.1 lot to 27.68 lot. It is ok for demo running.

I ran 30 pips grid, EA work ok, but, profit is lower that 20 pips one.
Attached Image (click to enlarge)
Click to Enlarge Name: MGH_eur-usd-19-11-2014.jpg Size: 110 KB
  • #978
  • Jan 4, 2015 9:16pm Jan 4, 2015 9:16pm
Quoting vaper
The point is that all those lots eat each other instead of making profits or even progress. Good Night vaper
Yes, this is a issue. But, a change to buy at lower or below ma line, to sell at higher or above ma line, positive gearing up may unfold.
  • #979
  • Feb 16, 2015 3:16pm Feb 16, 2015 3:16pm
Hi everyone,
Is there a version for 5 digits broker?
Thanks
  • #980
  • Mar 5, 2015 3:47am Mar 5, 2015 3:47am
Hi traders... Can anyone share with us the current result of this ea... and what is the minimum deposit amount for this ea :-)
1 trader viewing now