Something wrong whenever I try with dataseries

[pstrusi]

Hello ninjas

this could be and should be a very easy situation to solve:

I'm working with very simple formulas based on Close[0], Close[1], median [0], median[1]....etc

I've developed an indicator to chart it with range bars charts, so I'm trying to backtest it..the problem is when I try to calculate something like this (example) :

if indicator[1]>0 && indicator[0]==0 and I have no results when actually there are a lot of cases that match my conditions, formulas...etc. I thought thet it could be the floating point issue but it's not.

I've tried change <>== but the results don't make any sense for the data

For even odd situation I tried to do a new strategy, same formula, but without non related variables, formulas and surprisingly come out with some results...really I don't get it

I have good enough experience backtesting, developing formulas but I'm stuck with this.

Thanks in advance

[NinjaTrader_RyanM1]

Hi pstrusi,

Here are a couple pointers that may help you identify what's happening.
1) Have the strategy add the exact same indicator for visualization. You can do this in the condition builder by checking the box "Add to chart" when defining your condition. To add by code, please see the following link:

http://www.ninjatrader.com/support/forum/showthread.php?t=3228

2) Add drawing objects to your conditions so you can easily spot when it's true. See this post for a couple examples:

http://www.ninjatrader.com/support/forum/showthread.php?t=3419

If your indicator values look good and your condition shows true on the chart, next step is to see how the strategy respond to your order commands. For this, add TraceOrders = true; to Initialize() method, and then view output under Tools > Output window. This post can help further with TraceOrders usage.

http://www.ninjatrader.com/support/forum/showthread.php?t=3627

[pstrusi]

I'll run different trials with your suggestions...i'll let u know hot it goes. thanks a lot

[pstrusi]

By the way:
Are there any suggestions, precautions...etc to have in mind whenever you work with dataseries, running a lot of comparations among x[0], x[1], x[2]...etc ?

Under what conditions won't these comparations work?

[pstrusi]

Here you go a little but clear example what I deal with:

Below you can find two conditions under same variables, timeframe...etc

This one forks just fine:
if (Close[0]-Median[0]>0)
{
// do whatever
}

BUT IF I ADD THIS:

if (Close[0]-Median[0]> Close[2]-Median[2])
{
// do whatever
}

Despite there are a lot of cases I have non result, why?

[pstrusi]

Even more odd yet, if I add this simple condition:

if (Close[2]-Median[2]> 0)
{
// whatever
}

I have non results

What and how should I set conditions in order to be able to work with Close[2], Median[2], High[2]...etc?

[NinjaTrader_RyanM1]

By the way:
Are there any suggestions, precautions...etc to have in mind whenever you work with dataseries, running a lot of comparations among x[0], x[1], x[2]...etc ?

Yes, please have a look at this post:

http://www.ninjatrader.com/support/forum/showthread.php?t=3170

If this is what you're running into, you'll see index out of range errors in the log tab of the control center.

[pstrusi]

Yep, that was the reason, thanks a lot

[koganam]

Even more odd yet, if I add this simple condition:

if (Close[2]-Median[2]> 0)
{
// whatever
}

I have non results

What and how should I set conditions in order to be able to work with Close[2], Median[2], High[2]...etc?

 
if (CurrentBar < 2) return;

at the start of OnBarUpdate(), to escape the exception condition?

[pstrusi]

Yep Koganan, it was solved yesterday. I was working with an old code and I forgot to change barsrequired to more than 1.

Thanks