This is documentation for Mathematica 5, which was
based on an earlier version of the Wolfram Language.
View current documentation (Version 11.1)

Documentation / Mathematica / Built-in Functions / Numerical Computation / Numerical Precision /

Further Examples: SetAccuracy

When an expression contains machine-precision real numbers Mathematica tries to do all computations with machine precision.

In[1]:=

Out[1]=

This causes a loss of accuracy. One of the summands had an accuracy of but the result has a much lower accuracy.

In[2]:=

Out[2]=

In[3]:=

Out[3]=

By raising the accuracy of each subexpression, you can increase the accuracy of the overall result.

In[4]:=

Out[4]=

Adding two numbers of the same accuracy gives a result which is less accurate by one binary bit.

In[5]:=

Out[5]=