Kata

Great kata, howewer I really dont see the point of last test case :

# should work if (x1,y1) = (x2,y2) = (x3,y3)
Test.assert_equals(power_law([1,120], [1,120], 1), 120)

Does anyone see any interest, or any physical or practical meaning to that requirement, given that if (x1,y1) = (x2,y2), there is an infinity of solutions except on the given point...
That forces the user to construct a special case for that very situation (except for lechevalier clever hack but then it would silently yeald a wrong result when x1=x2!=x3...) and makes the solution ugly and bloated IMHO.
Would'nt it be more natural to give two distinct points or throw an exception if x1==x2?

hey everyone: I decided to change up the kata so that it gives you two points in the distribution and you have to find a third. I think this is a better way for people to understand what a power law distribution is, but it also makes the kata harder. My apologies to anyone/everyone who already solved this kata and whose solutions are now invalid, sorry about that. Any feedback/suggestions much appreciated as always!

I'm thinking of changing this kata so that instead of giving you the points held by position n and the relationship between position n and position 2n, it just gives you [position, points] stats for two arbitrary positions and has you infer the correct power law relationship from there (then finally prove you have done so by returning the points held by another arbitrary position). It adds a little extra challenge (will make the solution more than twice as long I think), but I think it better captures what power laws are about. Any thoughts?

hey guys -- I can imagine that this kata might run into problems with someone coming up with an equally-correct solution that is just a little different from mine, and which as a result is marked wrong by the checker (e.g. if they do rounding at a slightly different moment and come out with an answer that is 1 point different that would be marked as wrong). Could a more experienced coder help me figure out how to fix the tests so that this doesn't happen? (Or let me know if this isn't actually an issue)? Some possibilities I can think of are rounding the random decimals down to 2d.p. (currently I'm just using Math.random() as-is), or using some kind of leeway on the assertEquals statement (not sure how to do that but I know it exists). Any help/advice much appreciated; I've got no experience with these kinds of issues unfortunately so any fix I did right now would be kind of arbitrary.