Where does a bell curve miss?
Where do recorded counts depart from a fitted bell curve? This view’s reference export has not yet been verified against v4; its diagnostics need validation or replacement.
This is a check of a convenient approximation, not a test of whether a seed is “normal.” Use exact frequencies for rarity. Skewness and kurtosis describe aspects of shape but do not establish a particular tail probability. Data: published bake
Diagnostic summary
Which metrics depart most in this group?
RMS is the square root of the average squared gap over the evaluated quantiles. It is not an average error over every individual seed. Tail RMS uses only the outer 10.00% at each end. These are descriptive magnitudes, not pass/fail tests.
Inspect the 200 evaluated quantiles
These probes run from the 0.25th through the 99.75th percentile. They do not measure the most extreme 0.25% at either end.