File talk:Comparison standard deviations.svg

This picture has 4 major problems.

• First, and most severe, it's shifted by 5 units. The "bins" of 10 units are graphed at the leading edge instead of at the midpoint. E.g., the samples falling between 90 and 100 are shown at the x=90 position instead of the x=95 position where they belong.
• Second, the max and min bins (for the red sample) are handled differently. The max bin is graphed as a vertical line from its count, down to the origin; the min bin is graphed as a diagonal line from its count, over to the left neighbor bin's origin. The diagonal line method should be used in both places.
• The samples do not have the mean and S.D. indicated. The RNG produces values with the given statistical properties, so the resultant samples represent a subset of values from a larger (infinite) population with the indicated mean and S.D. The samples shown have something close to, but not exactly, the indicated properties. This is handled by shifting and inflating/compressing the output data to match the desired values.
• Finally, the x-axis is truncated at zero. I think the x-axis should include most or all data points, but it doesn't. And the fact that it truncates at 0, but not at the high end, gives the false impression that there is something special about 0, which is not the case when discussing the standard deviation and mean of a sample. — Preceding unsigned comment was added by 100.16.231.141 (talk) 04:35, 9 May 2018 (UTC)