1. Introduction

EPC calculates the result of expressions, in terms of variables whose values are uncertain. This is done by repeated calculation of the expression using variable-values which are generated using a random number generator whose mean and standard-deviation match the values specified for the variable.

If you don't have time to read this whole manual, just try running epc < file, where `file' is a file containing the two lines

$x = 1 +- 0.1; calculate 2 * $x;

where by the first line you mean that the most likely value for the variable named $x is 1, but that there is a 1 in 20 chance that it lies outside the range 0.9 to 1.1.

and you'll see the output from the calculation, which will looking something like the following:

$x = 1 +- 0.1; calculate 2*$x; Result without perturbation: 2 Result with symmetric tails: 2.0011 +- 0.199435 Result with asymmetric tails: 2.0011 (1.80269 to 2.20156)

where the first two lines are echos of the input, and the final four lines are the output of a monte-carlo calculation (5. How Epc works) based on the indicated formula.

The first line of results is the un-perturbed value, which is useful as a check on your entry of the formula, since it's likely that you've already evaluated that value on your calculator already.

The second line of results is the estimate of the formula's value, taking the error bar on it to be symmetric. This is meaningful for linear formulae or for nonlinear formulae in which the input variables are relatively well-known.

The third line of results is the estimate of the formula's value, but not assuming symmetric tails of the distribution. This is the most meaningful result for highly nonlinear cases.

You would be well-advised to compare the last two lines of results. If you need to learn more about the distribution of calculated results, run epc with the commandline switch -v, and it will print out a summary of the probability-density function of the results 4. Commandline Options.

In this case, as expected, the result is approximate 2, and the error bar is approximately 0.2, i.e. double the input error bar for this linear formula.