Why 0.6+0.3+0.1-1 is -1.11022302462516e-16, not 0 ? why?? A serials bug??

Question

denny 2018-6-20

回答： cdarling 2025-6-4

0.6+0.3+0.1-1

denny 2019-9-9

编辑：denny 2025-2-26

This is a bug? this is bucase the numeric calculation in a computer.

You can use a fraction to calculate it.

Answer 1

cdarling 2025-6-4

As a numerical calculation software, MATLAB stores data as double data type by default.

IEEE-754 is the standard for double data type.

This may better explained by the following code:

>> num2hex(0.1)
ans =
    '3fb999999999999a'
>> num2hex(0.1+0.1+0.1)
ans =
    '3fd3333333333334'
>> num2hex(0.3)
ans =
    '3fd3333333333333'

If you need to calculate the theoretical value, you may try Symbolic Math Toolbox:

sym('0.6')+sym('0.3')+sym('0.1')-sym('1')