This one is driving me nuts. I’m trying to code a very simple strategy:
“If close within 10% of range – buy on open next day. Then sell on first profitable opening.”
This is what I got so far:
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
varrange=high-low
varinrange=low+(varrange/10)
//Check if yesterday closed within 10%
IFclose>lowANDclose<varinrangeANDintrade=0THEN
varbuy=1
intrade=1
price=open
ELSE
varbuy=0
ENDIF
//Check if open above paid price
IFopen>priceANDintrade=1THEN
varsell=1
intrade=0
price=0
ELSE
varsell=0
ENDIF
// Conditions to enter long positions
IFNOTLongOnMarketANDvarbuy=1THEN
BUY50CONTRACTSATMARKET
ENDIF
// Conditions to exit long positions
IfLongOnMarketANDvarsell=1THEN
SELLATMARKET
ENDIF
1
I got as far as to enter correctly, but I cant exit without at least a days delay. I seems that PRT doesnt work “in the moment”, it looks at an entire bar and takes action the next day. Is there a way to exit as it happens? In this case I want to get out as soon as we open with a win.
You are right, each condition are tested only once at each new bar open. This Prorealtime engine behaviour will be soon changed to a whole new way to deal with it : each condition will be tested on every timeframe bars. From the one you are trading to the smaller one below it.
What you can do here with your strategy is to simply test if the price goes up or down at each new bar. So if you are long and price has raised up, exit your position and vice-versa.
I’m not following, how will that help me? Wouldnt I still get a one day delay with PRT’s behavior?
Even if price goes up, it could still close below. I need to confirm that open above entry before I could do anything and exit immediately once that happens… Perhaps its just not possible without those engine changes…
It will help your strategy not to wait an entire day (if you are trading on a daily basis) to see what happened. The backtest engine will loop all timeframes from the daily one to the 1 minute one to test each of your strategy conditions.
While your strategy is running on a Daily timeframe (1 bar = 1 day = 24 hours), on backtest, all your conditions will be tested like this :
Backtest of the X bar on Daily TF -> conditions tested on each 1 hour bar -> conditions spotted on the Nth bar of the day -> conditions tests on each minute of the Nth bar previously spotted -> Found the minute where the conditions occured -> Execute SL/TL/TRAILING or anything else on the prices of the bar -> Go on next Y bar on Daily TF -> and so on ..
That I understand, what I dont get is how the actual code for that would look like. I mean what does the code looks like where I check if the price is above entry and exit in the same bar?
Your code don’t have to be adapted to this future improvement of the backtest engine. Actually, if you want to deal with your orders, you can browse all the instructions available in the online prorealtime documentation, Probacktest specific, here : http://www.prorealcode.com/documentation/category/probacktest/
Also, if you want to store your price entry whenever you enter the market, you just have to create new variables and store the open price. Then, make conditional loop for anything you want around this data.
Nicolas I feel that there is a bit of a language barrier here 🙂
I know that there is online documentation, i’ve read it, but i’m still stuck. Could you please provide an actual code example that implements what you are saying. That will allow me to study your solution in a different way.
I have just came across your code above, (the one you helped Jerry with), I’m no programmer but I thought it would have been simple to modify the same code for just sell trades, couldn’t be further from the truth. What my missing? Its not a massive code, could you share the same code for just sell trades? That would be great if you could. I’m in the process of learning the coding language and the platform most of my trading for the last 7 years is through cmc markets but i’m looking at joining I.G Index to get access to the prorealtime platform.
.
.
