feature, more variation when using random ranges about dummyjs

[paulcollett]

One approach could be to reduce the possible outputted variations to 4 or 5.

Instead of:

<div data-dummy="100,300"></div> // 200 possible outcomes

5 possible outcomes:

<div data-dummy="100,300"></div> // 5 possible outcomes

100 // low
150 // low+(high-low)*0.25
200 // low+(high-low)*0.50
250 // low+(high-low)*0.75
300 // high

Or, use some kind of algo with preference for low + high + middle? Called "inverted bimodal distribution"? https://en.wikipedia.org/wiki/Multimodal_distribution