Agreed. Wish I'd thought of dropping the obvious unnecessary keyword! I started at about 500 chars!

Good kata - thanks. I'm surprised I'm not seeing anybody else using the stochastic acceptance method of selection given it's simplicity and performance. Mind you, I needed 200 iterations on a population of 200 to get reliable results. The paper about the method does suggest that the probabilities are the same as roulette wheel selection, and it's so easy I couldn't have implemented it wrongly, perhaps it's not as good. Anybody that knows about these things have an opinion on it?

+1 for downvote.

I don't think it's enough to just add a constructive comment to a bad solution that has a tonne of upvotes. I doubt if a large number of users even read the comments.

I've only been a member for a day and already noticed the problem described by others here.

I like the idea of requiring a comment for a downvote, and I like the idea of requiring one for an upvote too. I find the solutions disappointingly devoid of discussion, this may help.

Have to agree that this solution and others like it are not good practice. There is more to a good solution than getting the correct result - speed and usage of computational and memory resources are important, along with readability, maintainability, and so on.

Without wishing to be negative, I encourage others tempted to upvote this solution (and the author) to review other solutions where these factors have been taken into consideration.