A very simple cashcow

 

Well I hope it's a cashcow. Visual tests are very promising.

I use 1 day EURUSD.

Wait until the bar (a trading day) closes.

If High-Close<0.0015 then take a short trade with a 25 pip target.

If Close-Low<0.0015 then take a long trade with a 25 pip target.

We just lay in a market order on the todays bar.

Not yesterdays bar that we used to check if we take a trade.

It saves us the rollover costs.

You will get around 1 trade a week. That's 25-spread=~20 pips

It won't make you rich but $200 extra a week/contract ain't that bad....

Hopefully it's helpful to someone.

Now please be helpfull to me too :

I want to get this thing programmed as I want to backtest.

For this system to work speed is essential. You must lay in the order a few seconds from open. Not a few hours later when you get home or wake up.

I tried ZeroCode and the my own code (see below) but both give me not one single trade. Not even a loosing one.

============= only the shorting part ========

if (High[1]-Close[1])<0.0015 then

Begin

SetOrder(OP_BUY,Lots,ask,9999,stoploss,ask+30*Point,Red);

exit;

end;

 
TraderSeven:
Well I hope it's a cashcow. Visual tests are very promising. I use 1 day EURUSD. Wait until the bar (a trading day) closes. If High-Close<0.0015 then take a short trade with a 25 pip target. If Close-Low<0.0015 then take a long trade with a 25 pip target. We just lay in a market order on the todays bar. Not yesterdays bar that we used to check if we take a trade. It saves us the rollover costs. You will get around 1 trade a week. That's 25-spread=~20 pips It won't make you rich but $200 extra a week/contract ain't that bad.... Hopefully it's helpful to someone. Now please be helpfull to me too : I want to get this thing programmed as I want to backtest. For this system to work speed is essential. You must lay in the order a few seconds from open. Not a few hours later when you get home or wake up. I tried ZeroCode and the my own code (see below) but both give me not one single trade. Not even a loosing one. ============= only the shorting part ======== if (High[1]-Close[1])<0.0015 then Begin SetOrder(OP_BUY,Lots,ask,9999,stoploss,ask+30*Point,Red); exit; end;

For the ZeroCode I'm sure its probably just a spelling error or something. Check with scorpion by posting your ZeroCode logics, theres probably a tiny mistake that's missing out. Post it in the ZeroCode section, I'm sure you'll double your responses

 

Hey Trader7,

This system can be create by pure ZeroCode. First jump directly to system logics step without setting up any indicator. And then, setup below logics:

Buy Logics:

AND Close[1]-Low[1] < 0.0015

AND Volume <= 10

Sell Logics:

AND High[1]-Close[1] < 0.0015

AND Volume <= 10

In the last step, set Take Profit to 25. That's it! simple

 

It worked at first try. Thanks man!

But it blows my account in a matter of days....

Little time right now but while the code looks ok the signals/backtest look wrong.

Example:

EURUSD

1 day bar

15 June 2005

Close=1.2123

High= 1.2133

That's within the 15 pips range so the sell on the next bar is correct.

There is arond 45 pips to be 'earned' on that bar but still the backtest gives me a loosing trade.

What's wrong......?

 

Like Gazuz said somewhere, don't backtest, just go ahead to live testing for a few days to confirm.

 

Ok, will do so. And report back in a week.

I'm very new to MetaTrader.

Is backtesting bugged?

 

TraderSeven. MT backtesting uses bar data , so theoretically the onlytrustable backtestedare for expertst forking with ohlc pricec. This is why in many cases backtestin is not effective. In mt4 ( we all hope ) tick data will be used, and backtesting will be more reliable.

 

That's clear. A system can't use intraday data when there is only EOD data.

That's the case on all platforms.

But if you look at my example you will see that a trade was possible.

The backtest only needed the OHLC data.

Just waiting for some RT trades is impossible IMO. Around 1 trade a week is generated. And the minimum system test is at least 100 trades IMO.

I'm not that patient So I programmed the system in ProSuite.

No better then break even.

But examining the winning bars shows theorecically there was more profit in them.

And the loosing bars show that they theorectically could have be winners if I had taken a smaller profit target.

Trailing stops are the solutions here. But (as you) say for that intraday data is needed.

Is it possible (with some script) to use 1 minute bars as intraday data when testing an 'EOD' system?

Reason: