Math.Round method

[pstrusi]

Hi ninjas,

I've had a round off problem with a strategy cause the floating point issue. A good fellow here called Koganan suggested me this link that would allow me configuring a better format for me and thus eliminating my floating point bad signals problem.

http://msdn.microsoft.com/en-us/library/75ks3aby

Since I'm not an advance programmer, I'd like to know how to implement instructions within my code in order to be able to call this method; from NAMESPACE to following modules.

I work with EUR.USD with Interactivebrokers, so my input would be in 0.00005 ticksize; since I have to make calculations with average I'd like to work with 0.000000 format.

Thanks in advance

[pstrusi]

Ny the way, the function Round2TickSize is not good enough cause sometimes rounds to a wrong side, so that's why I need more digits. Obviously the problem come with comparing to Zero.

[NinjaTrader_Joydeep]

Hello pstrusi,
To assist you further can you give an example of what number you want to round and what exact result you want to have.

I look forward to assisting you further.

[pstrusi]

I expect to have number with maximum 6 basic digits after zero in order to compare fairly. In the earlier case I expect that number -0.000025 can be the same.
There are other cases when I get for example a number smaller, for example like -0.00000025 that I'd like to be Zero cause it has too many non meaningful digits.
I need just 6 meaningful digits after zero.

[NinjaTrader_Joydeep]

Hello pstrusi,
Thanks, but what do you expect the number -0.000025 to be (after the conversion). Are you trying to make it +0.000025?

I look forward to assisting you further.

[koganam]

Ny the way, the function Round2TickSize is not good enough cause sometimes rounds to a wrong side, so that's why I need more digits. Obviously the problem come with comparing to Zero.

So then, per the URL link I gave you the first time out, round to six digits, or however many you want.

RoundedValue = Math.Round(value, 6);

Does that not round to what you want?

[HaveGunsWillTravel]

I realize this post is old, but I had the same issue and came across it so figured I would post the best fix I found for others who happen here:

You can use this:

double rounded = f * Math.round(x/f);​

let the instrument TickSize = ts
So, create a double variable and populate it with the instrument TickSize somewhere, probably State.Configure would be my guess.

you want a price of something, maybe an indicator price for placing an order that will not be rejected, but lets use Close[0] as a simple example. You could populate any indicator call in its place for indicator values.

ts=TickSize

ts * Math.Round(Close[0] / tsf,0)

What is it doing...
1- It takes your price, however many decimals (e.g. .01, .25, .00001 whatever) and divides it by your ts.
2- it rounds that number to 0 decimal places.
3- it multiplies the rounded number by the ts to give you the original price rounded to the closest tick size.

This should be preferred since it is the most accurate and the strategy may be applied to any instrument without changing any decimal places in the code. Assuming you correctly set ts to the tick size.

This is especially useful for placing orders based on some derived price without getting rejected for invalid price.

[QuantKey_Bruce]

There's also double bar = Instrument.MasterInstrument.RoundToTickSize(foo);.

[HaveGunsWillTravel]

Yup, I just found

Instrument.MasterInstrument.RoundToTickSize(double price);

and was coming back to amend this post. But you beat me to it. Glad the information is out there now. It is not search friendly.

[koganam]

There's also double bar = Instrument.MasterInstrument.RoundToTickSize(foo);.

The original poster stated specifically that that method is not suitable for that which he wants to do. I do not see why not, but that is what he claims.