double numbers problem

Elsa Hejazian 2021.06.18 17:14

double NORMALIZE(double DOUBLE)
   {
    return round(DOUBLE/0.01)*0.01;
   }

VOLUME=MathPow(FACTOR,PositionsTotal()-2)*(MathPow(FACTOR,2)-1)*P.Volume();
Alert(VOLUME);
// result : 0.6333333333333334

Alert(NORMALIZE(VOLUME));
// result : 0.63

VOLUME=NORMALIZE(MathPow(FACTOR,PositionsTotal()-2)*(MathPow(FACTOR,2)-1)*P.Volume());

Alert(VOLUME);
// result : 0.6333333333333334     why not 0.63?

why last line doesnt work?

Fernando Carreiro 2021.06.18 17:28 #1

Elsa Hejazian: why last line doesnt work?

Forum on trading, automated trading systems and testing trading strategies

Problems with division especially when there's 7

William Roeder, 2021.06.15 14:43

Floating-point has an infinite number of decimals, it's you, not understanding floating-point and that some numbers can't be represented exactly. (like 1/10.s)
Double-precision floating-point format - Wikipedia

See also The == operand. - MQL4 programming forum 2013.06.07

If you want to see the correct number of digits, convert it to a string with the correct/wanted accuracy.
question about decima of marketinfo() - MQL4 programming forum 2016.05.18

How to get BID Can't round a number Color change in button

William Roeder 2021.06.18 17:37 #2

Elsa Hejazian: why last line doesnt work?

See also The == operand. - MQL4 programming forum 2013.06.07

If you want to see the correct number of digits, convert it to a string with the correct/wanted accuracy.
question about decima of marketinfo() - MQL4 programming forum 2016.05.18

Problems with division especially How to get BID Can't round a number

amrali 2021.06.20 08:29 #3

A double precision floating point number is just a binary approximation to the true real value. There is always this sort of roundoff errors. It is a drawbabk of the binary (base 2) floating point format, and you have to deal with it.

[WARNING CLOSED!] Any newbie NormalizeDouble not working Can't round a number

Dominik Egert 2021.06.20 08:58 #4

For precision. It so not a drawback by the binary representation.

It is the nature of real numbers.

As an example, try to think of the smallest possible number above 0.

As you can see, the "issue" is also within your brain. It is the natural limitation given by the universe we live in.

It has nothing to do with the representation.

The representation only defines the finite granularity, here limited to 64 bit. And therefore the result of the "rounding process" is based on the granularity of the underlying representation. In this case it's a Base2 representation. But the granularity is given with any representation.

There is no way out.

"New Neural" is an NormalizePrice() any suggestions on Metaquotes Language 5-Typecasting Question

amrali 2021.06.20 09:17 #5

0.63 is a finite decimal fraction. It can be represented exactly if MQL5 has support for decimal (base 10) floating point type. What I meant that using base 2 has drawbacks of not being able to represent the numbers the way humans do (base 10).

Of course we all know that 1/3 cannot be represented and this is not an issue in itself.

Price value Double, 6 = 6.00000 Subtraction different