SWB Martingale EA

I have ordersend error 131 by hedge. The whole trade closes only when Hedge TP is reached..
so when the price goes against hedge order but in favor with "original countertrend trade" it wont close and error 131 is in the journal..

I dint change hedge settings except start hedge level (6)
and orignal multiplier is at 2
TP in money is 10

Any thoughts?

Thank you anyway it looks promising. but I think It cant be used on nano accounts (1pip = 1 cent) because almost every broker has limitation of maximum opening lots... So it can be used on regular account but there youll need huge capital... Which I do not have..

Which error is that? ... "Not enough money?" I don't know, I don't work with the error codes a lot, although I do look at the journal from to time.

Yes, this is part of the problem ... or, is the main problem. The lot sizes multiply so quickly. This is why I set multiplier to 1.2 in the first instance above. It seems fairly effective when hedge_alt is set to true. But still, like you say, it probably requires more startup capital. Although it might help if you tried extending the range, to say 100, or something like that. You can also try reducing the lot size to .05 and tp_in_money to 3.00.

OrderSend Error 131 is a very popular problem that is usually encountered when testing MT4 expert advisors. What causes this error? It’s called ERR_INVALID_TRADE_VOLUME in the MT4 code. That means that your expert advisor is trying to send an order with invalid trade volume. On the absolute majority of the MT4 brokers setting some EA to open an order 0.123 lots will generate this error. But sometimes it’s generated when the EA, created for mini or micro accounts, is used on the standard account. If you stumble on OrderSend Error 131 during your testing, you can quickly find out the wrong settings of your EA — find the standard init() function inside your EA’s code and insert these lines of code there:

Print(MarketInfo(Symbol(), MODE_LOTSIZE));
Print(MarketInfo(Symbol(), MODE_MINLOT));
Print(MarketInfo(Symbol(), MODE_LOTSTEP));
Print(MarketInfo(Symbol(), MODE_MAXLOT));

The first line will give you the information regarding how many units one lot holds when you trade in this account (100000 would mean a standard-sized lot). Remember, that in your expert advisor’s log this line will be first starting from down to up, not vice versa. The second line will tell you the minimum amount of lots you can trade (this is the most usual error; you’ll probably just need to fix the amount of lots your EA trades from 0.1 to 1). The third one will give the minimum step for the trade volume in lots. The fourth line will tell you the maximum amount of lots that your EA can trade.
For example, demo account at FXOpen generates this info when I insert those lines into the code:

2008.07.10 15:13:37MACD Sample EURUSD, H1: 10000
2008.07.10 15:13:37MACD Sample EURUSD, H1: 0.01
2008.07.10 15:13:37MACD Sample EURUSD, H1: 0.01
2008.07.10 15:13:37MACD Sample EURUSD, H1: 100000

That means that 1 lot is 100,000 units (a standard size), minimum trade volume is 0.01 lot (so, one can trade starting from $10 on 1 position in a dollar-based currency pair), minimum trade volume step is also 0.01 lot (one can trade 0.33, 0.4 or 1.25 lot volumes, but can’t send orders with 0.333 lot size) and the maximum volume one can use to open a position is 10,000 lots.
You can incorporate the MarketInfo() function at a more complex level into your EA, so it could automatically check the allowed values and correct its settings. But if you don’t want to code much, you can just use the code above to find out the right values and correct the settings manually.

as i read this i think my testing platform would cause the problem (Alpari UK)

will try.. Thanx

Wow! How someone can keep track of all this stuff is beyond me? I tend to pick up bits and pieces as I go, while focusing on those things I'm more familiar which, for the most gets the job done. There's almost always something that comes up that's got me scratching my head though.

It could be. I'm using IBFX 4 digit mini account for most of my testing. So yes, micro lots are still very much in the picture here.

Attached is a script which returns details about your broker account & does not require writing code.

The script outputs a .csv file to directory Experts/Files which can be read by a text editor or opened by a spreadsheet program.

Hope you find it useful ...

I just got through running test with settings below. Started on 6/17/2009 (mini account limitations) and ran until 5/6/2010 when the big crashed occurred. It looked like it was about ready to pull out, but crashed at level 15. The account balance was up to $975 before it crashed. There were also three instances between December and February where it got up between L9 and L12.


start_lot=0.05;
range=40;
level=20;
multiplier=1.2;
tp_in_money=3.5;
hedge=true;
hedge_alt=true;
hedge_start=2;
h_lot_factor=0.35;
stop_loss_by_level=7;
sl_increment_money=2.50;
break_even_set=true;
initial account balance=500;

Oh, I had to modify EA, because it wasn't calling out correct parameters for sl_increment_money variable. I specified an integer instead of a decimal value, so I couldn't break it down at .50 cent level.

ETA: Updated to include Break Even mod referred to below.

Okay, I re-ran the test with better results changing variables in bold below. Of course it didn't yield nearly as much money (less than $400), but did manage to squeeze by, especially around may 19th, when it reached level 12 with range set at 60 (over 700 points).


start_lot=0.05;
range=60;
level=20;
multiplier=1.25;
tp_in_money=3.5;
hedge=true;
hedge_alt=true;
hedge_start=2;
h_lot_factor=0.35;
stop_loss_by_level=7;
sl_increment_money=3.0;
break_even_set=true;
initial account balance=500;

Just to let you konw, I updated the break even feature so it's now customizable, and you can specify it to start at whatever level you want. So, if you set it to level 6, and the stop loss to level 10, that means levels 6 thru 9 will close out at the break even point, before the stop loss kicks in at level 10. I think it's a good idea, since the point is to get out of the trade at your earliest convenience. Also, if you are hedging, it would probably be better to set it at least one or two levels after the hedge. That way you will realize any hedge profits above and beyond the target profit.

Here is the latest test I ran with the settings below. It was doing pretty well until it crashed on May 6th at L10 but, I had also gotten more aggressive with my settings. I had about $620.00 in profit ($1,120.00 total) before it crashed.


start_lot=0.05;
range=40;
level=20;
multiplier=1.4;
tp_in_money=5.0;
hedge=true;
hedge_alt=true;
hedge_start=3;
h_lot_factor=0.4;
stop_loss_by_level=10;
sl_increment_money=4.0;
break_even_level=6;
initial account balance=500;





Oh, I've included the update in the post above.

Hopefully May 6th won't happen again - it was not a 'normal' market move - see attached for details, so I guess if you can get past May 6th you are certainly on the right track.

my forward test:

start_lot=1;
range=25;
level=10;
multiplier=1.2;
tp_in_money=50.0;
initial account balance=10,000;
period : 6 june 2010 till 20 june 2010
profit 86.94% (8 694)
max level 5 = 2.0 lot

suggested money management: use only 20% of investmentcapital

f.e. when you have 50k use only 10K for trading and 40K in a non traded subaccount.

recalculate every week the balance and transfer 80% of profit to subaccount.

Can't figure out why you had level 12 on 19 may.
Didn't you had a gap in your data from the broker.

I tried backtests on ALPARI and MIG and both has gaps, so backtests are useless on that way.

There were gaps of 3-4 month's with over 1 000 pips difference.

Gaps? Really? Although I noticed there was a gap in April, that kind of messed things up. It looked like it occurred over the weekend though, so I figured it had something to do with broker opening times. If so, there's not much you can do about it. I didn't notice any gaps around may 19th though.

No, I haven't noticed anything like this. But you're right, if these things happen, then backtesting isn't very useful for testing an EA's performance.

Indeed the data must be compleet otherwise a backtest is not reliable.
3 years ago the data of ALPARI was the best in the industry, but now I see they have gaps also.
Anyone knows an MT broker with complete data for backtesting?

What software are you guys using for back testing??
thnx, Ron

Guys,
Anyone working on a trailing stop for this EA???
thnx, Ron

Im sorry but trailing stop at this kind of martingale has no point.. but you can use Steve Hopwoods MPTP for trailing all opening trades...

Ive been playing with the ECs SWB 4.1.0.3.2_h.. I backtested it on GU and EU from 8/jan 2001 to 30/Mar 2010 with no crash - 90%Modelling quality!

It was backtested under regular account condition so 0.1lot=1USD with 200.000 USD start capital (on nano lots would be only 2000 USD)

settings were
lot start 0.1
Range 50 (GU), 40 (EU)
Multiplier 3 - this is big point
TP 20
BB false
Stoch False (to have as much trades as possibile)
RSI 21, 45/55
TF 15min
allow trade on friday True
Hedge false

GU made 445.197 USD net for 10 years
EU made 385.948 USD net for 10 years

The deepest level for both pairs were L7 (14x by GU, 16x by EU in 10year period) which is in term of opening lots almost equal for L10 with multiplier 2. The big point is that with Multiplier 3 Youll need less points to hit the target by retrace than by multiplier 2.

Multiplier 3 will blow your account faster But in term of probability it is "safer" than multiplier 2.

I cant upload the statement since Im running new tests.. it takes about 24hours!!!