I have no clue how to program in MT. (I wish I had a VB6 API...)
So I have just manually checked the system by looking at 8 years of EURUSD.
1 day time frame.
Only one bad trade
1] Calculate a 10 day simple moving average. -> SMA(10)
You can use any combination of OHLC. It doesnt really matter.
For this example we use the current intraday price (close)
2] Plot SMA(10)*1.02 --> upper band (+2%)
3] Plot SMA(10)*0.98 --> lower band (-2%)
Of course you can tweak the values a bit.
4] If the price crosses above the upper band:
a] Short.
b] Set a target of 30 pips below the upper band.
Often we can earn more then 30 pips by moving the upper band up or
use a trailing stop to take profit. An endless variation is possible.
We use a very conservative 30 pips to prove the system is good or bad.
5] If the profit taget isn't reached at the close we hold the trade 'till the next day.
a] Now we wait what happens first:
- Profit target is reached. --> close trade
- Price opens/crosses above the upper band then Short.
b] Set a new profit target of 30 pips below the prices you got your positions at.
c] Wait the remainder of the day if the target is hit. No new positions.
6] Target not hit then goto 6
----------
Yep, your position can grow.
But when you test you will see: (some rough values)
40% of the trades complete in 1 day.
40% of the trades take 3-4 days
19% complete in 5 days.
1% take longer than 5 days. That are usually the loosing trades too.
Cons:
- Position can get large.
- Only trades ~12 a year.
Pro:
- System looks very well.
Note: if the position grows you profit grows 30 pips for every extra position taken.
How to improve?
We can discuss it in detail if someone cares about this system.
Only a programmed system can *prove* if this system is good or bad.
At the cost of a few trades less the profitibility can at least double.
It will even get more reliable. Less loosing trades. An most likely the losses get smaller.
Now I just hope there is some kind soul in here with speedy programming fingers
