MGH System (Martingale Grid & Hedging) | Page 48 | 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)

  • #941
  • Nov 21, 2014 4:31pm Nov 21, 2014 4:31pm
Quoting Paracelsus
{quote} The issue with two non-synchronous EA is that at any point they may align in the same direction which might double the DD. So once started together, the EA which reaches TP first would have to wait with its next "Batch1/Trade1" entry until the other one reaches its Global TP, i.e. when it can escape from the Basket. Then, and only then they could open new simultaneous bi-directional "Batch1/Trade1" trades according to Mastersenseo's idea.
I think you're right Paracelsus , they would both go their own way if they don't wait for one another.

It is possible to make as many EA's in ONE file as you want but I imagine its a hard job to do !!
  • #942
  • Nov 21, 2014 4:45pm Nov 21, 2014 4:45pm
Test run: EUR/USD from 2014-07-07 -- 2014-11-08 with basic settings in MGH EA_v4.2 with only one setting changed: BaseLot=1.35

Attached Image (click to enlarge)
Click to Enlarge Name: TEST_4-maand_std-settings_Lots1-35.png Size: 23 KB


@Paracelsus: would you run that same period too on your EA so we can see the differences? tnx

Remarks on this test:

  1. largest lots used = 166.05 lots
  2. there are more baskets then single (or double) trades
  3. the baskets close at zero profit/zero loss (as intended but much more baskets then normal trades)
  4. close-all works fine in a test or demo but will never work like that on a live account (closing a trade takes time!)

General: what is an EMA doing here in a Grid system ?
Some stuff to think about. cheers,
vaper

edit: after close-all the new 1st trade is opened immediately. This will not work on live account without issues (depends on broker)

  • #943
  • Nov 21, 2014 4:55pm Nov 21, 2014 4:55pm
Quoting Mastersenseo
... Swingman, I have a suggestion for your nice EA. As stated several times before, the benefits of the system comes from the first batch and next batches are to facilitate the escape. . . . we can add the EA to two different EURUSD charts, one to follow the trend and the second to enter against the trend. With this, we will ensure that at least one of them will be right into the selection of the first batch. I hope that this will increase the overall profitability while not increasing the risk significantly.
Excellent idea Mastersenseo!
Everything happens on the same chart and a hedging account is needed for this system. So, it must be easy to open both Buy & Sell every time at the 1st and 2nd trade after a close all.
  • #944
  • Nov 21, 2014 4:58pm Nov 21, 2014 4:58pm
Quoting SwingMan
I just noticed that after a restart the EA is not working properly ... Please be careful and do not trade live. The cause was an unexpected restart of the Metatrader server ...
Hello SwingMan, if you want to memorize the crucial info you can use a file or just lines.
Draw the lines far away from from real-time and put info in their title.

Happy Trading
  • #945
  • Nov 21, 2014 5:41pm Nov 21, 2014 5:41pm
Quoting cescof
{quote} What's happen if we put one Ea only long and other only short in two different charts with 2 magic number?
I haven't test it, but the structur of the EA will not help out on this strategie. This Ea is briliantly working, really almost excellent ,but needs to improve on MM and exits, the need of bigger capital to trade over longer period because of the DD, big lot sizes. Focus focus hokkus pokus
" Accepting losses is the most important single investment devise "
  • #946
  • Edited 10:14am Nov 22, 2014 8:53am | Edited 10:14am
Quoting vaper
Test run: EUR/USD from 2014-07-07 -- 2014-11-08 with basic settings in MGH EA_v4.2 with only one setting changed: BaseLot=1.35 {image} @Paracelsus: would you run that same period too on your EA so we can see the differences? tnx Remarks on this test: largest lots used = 166.05 lots there are more baskets then single (or double) trades the baskets close at zero profit/zero loss (as intended but much more baskets then normal trades) close-all works fine in a test or demo but will never work like that on a live account (closing a trade takes time!)...
@vaper

Thanks for the - little surprising - results.

The most surprising finding was that "there were more baskets then single (or double) trades", especially that EMA200 related Batch1/Trade1 direction choice was intended to get (at least slightly) more single/double trades (i.e. "home run" TP from Batch1 trades) than Baskets. Based on your results EMA related initial direction choice was clearly counterproductive. Wow...! Who would have expected that...?

I have just downloaded tick data via TickStory and I am about to do my very first 99% backtest, so at this moment I cannot confirm the above results.

Also, I have just asked my programmer to remove the EMA related direction choice and to modify my own version of MGH EA to become BI-DIRECTIONAL, so this way the 20 pips or 50 pips TP would be GUARANTEED with each new pair of trades. In other words: One entry direction will profit 20 or 50 pips, while the other entry direction will get out at BE (or 5 USD profit to cover commissions) from the Basket.

Also, I have added UltraGlobal_TP to the Settings, as I am planning to use the EA on multiple (at least 10 loosely correlated pairs) simultaneously for diversification and risk reduction. So if one or two pairs are in "trouble" the other pairs can "sacrifice" their floating profits to rescue the "troubled" pairs which are struggling to get out of a Batch3/Batch4 Basket, for example. So if and when floating profit on the account reaches UltraGlobal_TP, all the trades on all pairs are closed at once and the EA resets itself on all pairs. I am planning to do more extensive backtesting on this new version.

So from this new version of (bi-directional and diversified) MGH EA I expect increased safety and at least DOUBLE profitability.
  • #947
  • Nov 22, 2014 10:37am Nov 22, 2014 10:37am
Hi,

The bi-directional idea sounds very interesting. However, even though it does guarantee a TP of Batch1, it also guarantees at least a Batch1/Trade2 on the opposite trade, and very likely a hedging Batch2. Sometimes i find that most of my profits come from surfing the Batch1/Trade1 trades on a strong trend. However, I am very eager to try and implement this. The very ideal scenario of this would be a TP of Batch1/Trade1 in the buy direction, while the hedging sell trade hits Batch1/Trade2, and then retraces back to TP.


At the moment I have been employing MGH with discretionary entries and exits, and I've been up 80% for the week, on a $3000 account with 0.1 lot size (risky, i know). I don't believe in 100% automation. Despite what many say, the market is definitely not random, and i believe it is the collective psychology of all of its participants. Many times, price can only go one way, and it would be stupid to have a majority of your hedged lot size in the opposite direction. The longer the price runs, the more reason it has to come retrace. Take UJ for example, it broke through 118, and started to grind to a retrace. I wouldn't keep piling up buys anymore, but instead probably pick the tops and bottom of the channel.

Another thing i believe is important is volatility. It allows us to get out of the hedge faster, and decreases the chances of a range which blows up our martingale. One suggestion would be a time-based entry, such as London Open for GBP/EUR pairs (maybe 2am-6am EST), and NY Open for USD pairs. (7am-12pm maybe?), and Asia as well.
  • #948
  • Nov 22, 2014 10:55am Nov 22, 2014 10:55am
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=1.35
  2. Basket_TakeProfit=500

Testing means to me finding its boundaries ans see what they reveal. Technically it all works perfect and I put my hat off to SwingMan for his coding capabilities !!

This test went into crash a couple of days before the end of test period. Here are the overall Balance movements :

Attached Image (click to enlarge)
Click to Enlarge Name: Test2_4-maand_std-settings_Lots1-35_BasketProfit500.png Size: 22 KB



Many batches close sooner then in previous test which makes me wonder how the decision to close-all is organized. Is there anywhere a flow-chart of this EA?

Here happened the crash:

Attached Image (click to enlarge)
Click to Enlarge Name: Test2_CRASH_3.png Size: 18 KB


Remarks:

  1. more than one Buy on the same grid line -> not needed (same for Sells)
  2. no extra trades needed between top & bottom of a particular grid range once both were set before (B at top & S at bottom)
  3. lot sizes go very high, this cannot be the correct strategy
  4. where 1.35 goes to 840.78 it means that 0.01 lots would go to 6.23 (I think .. not sure)
  5. nice profits on the close-all

I have some ideas how to make this better but I would need a logic-flow-chart first.
One thing is for sure: always start a new series with trades #1 and #2 hedged at open.

cheers & happy trading
vaper

  • #949
  • Nov 22, 2014 4:16pm Nov 22, 2014 4:16pm
Hi Paracelsus,

@vaper has an interesting art to backtest..., but on two of his ideas I also used to think:
- more than one Buy on the same grid line -> not needed (same for Sells)
- no extra trades needed between top & bottom of a particular grid range once both were set before (B at top & S at bottom)

When I have time I will test a little change of the grid construction, as example:
- Batch1 is sell (level 100 and 110), Batch2 is buy (level 120 and 110)

- For a new sell Batch3, instead to take the same levels as Batch1 (100, 110), take a lower level (90, 100)
- For a new buy Batch4, instead to take the same levels as Batch2 (120, 110), take a higher level (130, 120)

I havent calculate the values, but I think that the shifted levels reduce the lots amount.
  • #950
  • Nov 22, 2014 4:27pm Nov 22, 2014 4:27pm
This is my backtest 2years with EUR/USD.... I've change grid range to 20 (only way to survive) ... DD is to high... but...hope for new development...
Thanks
Regards
Attached Image (click to enlarge)
Click to Enlarge Name: Strategy Tester MGH EA_v4.2.1.png Size: 197 KB
  • #951
  • Nov 22, 2014 5:25pm Nov 22, 2014 5:25pm
Quoting cescof
This is my backtest 2years with EUR/USD.... I've change grid range to 20 (only way to survive) ... DD is to high... but...hope for new development... Thanks Regards {image}
@cescof Hello, nice testing, tnx. Would you also show the graph please?

How different is it when you set Basket_TakeProfit = 15 or 50 or 100 ?
cheers
  • #952
  • Nov 22, 2014 5:33pm Nov 22, 2014 5:33pm
Here is an idea about the first batch only. It would need separate treatment then all others.

Attached Image (click to enlarge)
Click to Enlarge Name: Idee-1_start-hedged_bis.png Size: 29 KB


In this example the market is going up. Everything should be mirrored if the market moves the other way.

First batch opens with hedge-1 (B1 & S1) on start-grid and hedge-2 (B2 & S2) on next grid. Startlots is from the setting and the 2nd lots is double of the first. The method is wait until either both buys go in TP OR both sells go in TP. Then it is as noted in the picture.

Comments plz ?
cheers

edit: this way there is first 2 TP's and then either next 2 tp's if market turns before reaching next grid or there is a full hedge with the 2 sells and 1 buy. That buy would have lots = L1 + L2 to make a full hedge. From then on it doesn't make any difference where the market goes and how far away .. the DD value of the hedge will stay the same and is a good measure to work the next batches.
  • #953
  • Nov 23, 2014 2:42am Nov 23, 2014 2:42am
Quoting vaper
{quote} @cescof Hello, nice testing, tnx. Would you also show the graph please? How different is it when you set Basket_TakeProfit = 15 or 50 or 100 ? cheers
Hy... above the 2 charts.. Firs is with basket tp set to 50....and second is the original with basket tp set to 0
Attached Image(s) (click to enlarge)
Click to Enlarge Name: StrategyTester1.gif Size: 6 KB
Click to Enlarge Name: StrategyTester.gif Size: 7 KB
  • #954
  • Nov 23, 2014 11:16am Nov 23, 2014 11:16am
Quoting cescof
{quote} Hy... above the 2 charts.. Firs is with basket tp set to 50....and second is the original with basket tp set to 0 {image} {image}
@cescof thanks mate, it seems you have the same effects as I have: with a basket-tp > 0 some batches close sooner then with tp=0 and thus the spikes in the chart are smaller (less DD).

On the other hand when baskets are not closed earlier, they build up trade upon trade (see my remarks earlier) until the strategy goes into crash ! Its good to see that this is not depending on the broker or history data.

The system is simply trading too many times on the same grid-lines. Tnx for testing , cheers!
  • #955
  • Nov 23, 2014 1:38pm Nov 23, 2014 1:38pm
Quoting SwingMan
Hi Paracelsus, @vaper has an interesting art to backtest..., but on two of his ideas I also used to think: - more than one Buy on the same grid line -> not needed (same for Sells) - no extra trades needed between top & bottom of a particular grid range once both were set before (B at top & S at bottom) When I have time I will test a little change of the grid construction, as example: - Batch1 is sell (level 100 and 110), Batch2 is buy (level 120 and 110) - For a new sell Batch3, instead to take the same levels as Batch1 (100, 110), take a lower...
I think this is a very good solution to prevent batch accumulate when the price comes into range, especially at low volume days ...
  • #956
  • Nov 23, 2014 3:56pm Nov 23, 2014 3:56pm
Why not work on something that you know isnt guaranteed to fail someday? I encourage you to watch what price and price alone is doing. This has been the holy grail of trading for me. For example, if you see an uptrend in place and price keeps hitting a certain level with higher lows...you know with certainty buyers are trying to pile in and you are probably going to have a significant breakout very soon. Or is price making new highs and higher lows after a significant move down? Price really does tell everything and theres really no need to martingale if you can win 70-80% of your trades. Good luck.
  • #957
  • Nov 24, 2014 9:02am Nov 24, 2014 9:02am
Quoting tradingeasy
Why not work on something that you know isnt guaranteed to fail someday? I encourage you to watch what price and price alone is doing. This has been the holy grail of trading for me. For example, if you see an uptrend in place and price keeps hitting a certain level with higher lows...you know with certainty buyers are trying to pile in and you are probably going to have a significant breakout very soon. Or is price making new highs and higher lows after a significant move down? Price really does tell everything and theres really no need to martingale...
Hi, I'm interested. Where is the thread for that ?

Here we work out a GRID system using trades, countertrades, martingales, hedging etc .. etc..

Happy Trading,
cheers
  • #958
  • Edited 12:16pm Nov 24, 2014 12:05pm | Edited 12:16pm
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 = 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)

Attached Image (click to enlarge)
Click to Enlarge Name: Test3_4-maand_std-settings_Lots1-00_BasketProfit1000_bis.png Size: 44 KB


The largest buy is at 1,000 lots !! and largest sell at 622.8 lots !!
Details:

Attached Image (click to enlarge)
Click to Enlarge Name: Test3_4-maand_std-settings_Lots1-00_BasketProfit1000_bis_Values.png Size: 23 KB


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. Suggestion : trail the position once a certain positive P/L is reached.

Let's find out how we can solve this problem of grids with many trades.
Happy Trading
cheers

edit:
Total Position:

  1. Top line :
    5 BUY @ lots (2.80 + 10.80 + 54.70 + 276.80 + 1000.00) -> Equals BUY @ 1345.10 lots .
  2. Middle line :
    4 BUY @ lots (0.60 + 3.20 + 16.40 +83.00) -> Equals BUY @ 103.20 lots
    5 SELL @ lots (0.50 + 1.40 + 7.30 + 36.90 + 186.80) -> Equals SELL @ 232.90 lots
  3. Bottom line :
    5 SELL @ lots (1.00 + 4.80 + 24.30 + 123.00 + 622.80) -> Equals SELL @ 775.90 lots

vaper

  • #959
  • Nov 24, 2014 2:04pm Nov 24, 2014 2:04pm
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 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...
  • #960
  • Nov 24, 2014 6:29pm Nov 24, 2014 6:29pm
The point is that all those lots eat each other instead of making profits or even progress.
Good Night
vaper
2 traders viewing now