Instructions: Fill in a nominal value and error bar
for each input variable you intend to use (x, y, and/or z). Also fill
in a formula that maps the input to the output. Then press
<Enter> or click the Go button to calculate the nominal value
and error bars for the result. Explanations and examples can be found
in the documentation: here. That's the
simple version.

The fancy version provides additional statistics, plus graphs of
the distribution over results. It also has documentation: here.