Bugs: aprates

For low counts, aprates uses Poisson statistics, but for moderately large n or m (100's), this takes a very long time. If pdf=alternate, the tool uses the max_counts parameter to determine when to take an alternate path and use Gaussian statistics.

The quantity (rn-m)/(rf-g), which is the solution for net counts S in the set of simultanaeous linear equations given in the help file, is used to determine the transition between Bayesian and Gaussian approaches for the max_counts parameter.

(The "approximately" is due to how aprates works. It creates a grid of probabilities around the expected mean and evaluates the Poisson distribution. So if, for example, n=295 it might [depending on the other parameters] need to go out to 310 ; the values greater than 300 would be artificially truncated.)

If n or m are greater than 300, max_counts must be set low enough to use the Gaussian approximation to avoid this bug.