No language stated + probably duplicate issue (no random tests...) + OP long time inactive).

Just want to second this -- ebbruns description of the issue is good, and I think either the description should be updated or those corner cases should be abolished from the random tests (personally I don't think that they add much)

As an 8kyu kata this is the right difficulty. Maybe you need to look at harder katas instead? ;-)

I was just going to write that, verbatim.

Try now ;)

That could be one way to make it more interesting. Or maybe there could be a % chance to hit and a % critical chance, and they would have to estimate the average damage per attack. A kata similar to this one that was significantly harder had two fighters harming each other in turn, and required a knockout to be declared.

I've thought about it.
And I open for suggestion.
Mb it wuould be better to think about character strength and some multiplier according to it?

Thanks. COUNT will just count the rows with a value in name, two people with the same name will still be both counted. I'd have to test this, but I think COUNT(DISTINCT name) would produce the error you are concerned about.

I think this is a little dangerous because I imagine there could be a situation with two people with the same name?
Still nice code!

Sorry about that! Fixed.

For anyone reading- turns out not using "var" means you're declaring a global object. I have since refactored the code to include "var" because it bothered me.

Just realized that I forgot "var" was a part of Javascript while writing this code. Interesting that the language works in such a way that I got away with it.