Gertje: a hedged grid trader

[Gertje]

One thing that bothers me is the initial hedge. No matter what happens next, the first two trades will at all times hedge each other, so their contribution to the TP target will be null, apart from the spread that we pay for them. I guess placing just one iniital trade (randomly or defining the direction according to some parameter - RSI, last candle, other) would make more sense.

My $0.02

I'm playing with the concept since Steve gave us the 6pip Dipsum EA with the currency ring. That didn't work for me, so I took out 1 pair and found it more useful. After watching it trade, I had the brainfart to try it with 1 pair, and here we are.

You NEED the first trade to be hedged as in case you take only 1 direction, and price moves against you, when the first stacked trade appears you're already x-pips in the hole that needs to be recovered. With the hedge, you're at -2*spread so the 1st stack trade can make the profit in no time.

I'm trading this strategy on 7 pairs with small spread, with 0.01 lot per 200,- of balance with €1,- profit per 0.01 lot [equals appr. 10 pips]. Since this strategy trades a whole lot of times, the best practice is to go with a broker where one can get some sort of refund. The profit from the trades give several % a day, the rebate can be even bigger.

I use a stacking distance of 2*spread +1 [17pips for a 3pip spread], so normally the 3th or 4th trade can deliver the profit.

The main catch will as always be the downdraft. If things are going wrong, and one ends up with say 5 longs en 5 shorts, the DD will be 5 * 5 * stackingdistance so in my case 5 * 5 * 17 = 425 pips

Key will be to choose the stackingdistance with care, to avoid a cumulation of open trades.

Some results:
http://www.mt4i.com/users/gertje [live]
http://www.mt4i.com/users/gertje-instaforex2 [live, watch the refunds!]
http://www.mt4i.com/users/gertje-got_money_fx [demo] This one has a indi on the Empty4 stating the earned cashback = $2.200,- over the period of trading

Any recommendations are very welcome of course.

Best,
Gertjan

[K-Lander]

One thing that bothers me is the initial hedge. No matter what happens next, the first two trades will at all times hedge each other, so their contribution to the TP target will be null, apart from the spread that we pay for them. I guess placing just one iniital trade (randomly or defining the direction according to some parameter - RSI, last candle, other) would make more sense.

My $0.02

I'm playing with the concept since Steve gave us the 6pip Dipsum EA with the currency ring. That didn't work for me, so I took out 1 pair and found it more useful. After watching it trade, I had the brainfart to try it with 1 pair, and here we are.

You NEED the first trade to be hedged as in case you take only 1 direction, and price moves against you, when the first stacked trade appears you're already x-pips in the hole that needs to be recovered. With the hedge, you're at -2*spread so the 1st stack trade can make the profit in no time.


Gertjan

Yes, you´re right about the single trade. However, the two trades still don´t make any contribution. Have you tried placing the grid without any initial trades and let the market pick the pending orders as it moves? Sorry if I´m missing something here, but I cannot see how the two iniital hedged trades can have any influence either in a loss or win scenario, other than the spread we pay and the margin they eat.

K

[digipl]

12 Hours on a demo account....
.-2.000$
.-1:500 leverage
.-8 pairs with less spread
.-0.1 lots

Profit: 23,62$
Gain: 1,18%
Monthly: 48.63%
Drawdown: 28%
Equity: 1601.16$
Best pair: GBPCHF almost 50% of the gain

Not bad but now 45 open trades....

We'll see tomorrow.....

[SteveHopwood]

Latest update in post 1. This one in response to this pm from Alpenkorps:

"Just a question,
As you said the EA will rinse and repeat, so the EA will open another set of trades once first set reach TP? If thats it, then is it possible to make a option so the EA will not trade again if TP achieved?

The idea is simple, starting the bot manually only when market is trending or about to make a trend shift (possibly). This EA almost matches my 4h strategy, I usually start the process on OB/OS zones, mostly after a pinbar or engulfing."

So, if StopTradingAfterBasketClosure is set to true, then the EA will send no further trades until it is restarted. Bear in mind that this can be an accidental restarting following a platform reboot. I have left the variable name deliberately vague in case we decide to include a basket stop loss.

A's pm also provides a classic clue as to the use of this kind of EA.

[SteveHopwood]

By the by, someone earlier mentioned a red label FF news release.

Two or three years ago, when Scoobs first burst into our lives and in what feels like a previous incarnation, we tried to use Fcal to help avoid trading around news releases.

Basically, it does not work. FF news calender is unreliable: it mis-labels events; it misses important releases. Given what we now know but did not know then, that FF is owned by the crims this is not a surprise.

There is no way of coding for the release of a major news event that we know of at this point in time.

Not that they matter these days. Even that original source of terror, NFP, raises barely a flicker. 6 months from now, who knows? But for now, they are a non-event.

Which is not to say we should not try to catch them when news causes a half-way decent spike (the best we can hope for these days), but not in the 'normal' way of trying to jump onto the spike as it happens. Nope, when a ludicrous spike happens, watch it for a few minutes and then jump in hoping to catch the reversal.

This kind-of implies that this EA is not for always-in, unmonitored trading. Anyone surprised by this has not been reading the first posts in my threads properly. Or at all.

[Jollyroger]

Steve: I can't seem to get the latest update on the bot to load on a chart. It won't compile. The problem seems to be in the 1st two # include entries and I haven't a clue as to how to fix them. Is it me or the bot? Thanks
Roger

[SteveHopwood]

Steve: I can't seem to get the latest update on the bot to load on a chart. It won't compile. The problem seems to be in the 1st two # include entries and I haven't a clue as to how to fix them. Is it me or the bot? Thanks
Roger

Hehe. This is a classic. How the blue blazes should I know what is going on at your end? Or maybe you think I upload stuff that has not first compiled at my end? Or think I can read your mind? And psychically know what is happening on your machine?

Do my God-like powers know no end?

Get a grip JR. Close down the platform, delete the ea, redownload the bot, restart the platform, reinstall the platform, reinstall Windows, ritually sacrifice a few of Whip's chidkens. Usual stuff.



Or should that smilie be a sigh of despair? Anybody happen to have one handy?

[hiredwhip]

Steve: I can't seem to get the latest update on the bot to load on a chart. It won't compile. The problem seems to be in the 1st two # include entries and I haven't a clue as to how to fix them. Is it me or the bot? Thanks
Roger

JR,
I don't think you need any of my chickens......change your lot size to 0.01 or 0.10......Whatever your crooks min lot size is and then recompile.......

Have a great day

whip

[TehCount]

Basically, it does not work. FF news calender is unreliable: it mis-labels events; it misses important releases. Given what we now know but did not know then, that FF is owned by the crims this is not a surprise.

There is no way of coding for the release of a major news event that we know of at this point in time.

Perhaps we could use the dailyfx calendar. They have a CSV feed there, and the news is marked by importance:

http://www.dailyfx.com/files/Calendar-03-11-2012.csv

If an external program was to process/parse this information, could it be used in the EA? I should be able to create a Perl script (compiled to exe) that can return some useful values using this calendar CSV information. Something like "Is there high importance news in the next xx hours", then you could call it with the EA once an hour. When the EA gets a value of 1 it can disable trades. Let me know if that would help.

[neilrickaby]

this might be a noob question, but isn't this strategy more or less impossible to succeed over the long term, due to the cost of spread versus the cash profit target, i.e with every opening trade the breakeven point becomes more expensive - eventually creating a situation where it becomes impossible to breakeven?