Can't you just point out which piece isn't random instead of being passive-agressive ?
I made extra changes (even if I don't see the point of having them) to get random state for the key.
So now :
generated keys are random, key count is random, values for keys (true or false) are random, keys switch true/false are randome, amount of read/write to key is random.

Is there something more ? Please pin-point...

My bad, I see waht you mean.
I overlooked the problem.
I did extra changes and it should be fixed now.

Should be fixed now.
Test.assertEquals is supposed tobe a strict comparison or am I missing something ?

https://www.codewars.com/docs/js-slash-coffeescript-test-reference

The sample tests are limited, but the real tests are random.
the rule is 0 or 1 can be true. (i will update the description)

The idea was to give more insight on what is called "variable by reference"

I intentionnaly didn't want to make it to simple to solve.
If i add this details to the description, you can write the code with almost no thinking

why adding +1 if you substract it after ?

Oh, I see!
That's odd (if I can use this word ;) )
That should be part of the description I guess.

I mean, if shee can count to 100, that obviously include 100. She does not stop to 99.

That also mean "Expected" and "instead got" are somehow reversed.

Am I missing something ?
price: 360000
haggle: 'medium' (30%)
It should work for random inputs too - Expected: '251999.99999999997', instead got: '252000'

But 30% of 360000 is 108000
So final price should be 360000 - 108000.

best kata ever!
Thanks @buttonupbub

the kata description is innacurate.

if str = 'abcdefg' , the result should be 'a**'
but it should be
if str = 'abcdefg' , the result should be 'a******'

Also, what do you mean by "try Reg" ?