Got it, and solved. Thanks a lot!

Examples of k-primes:
k = 2 -> 4, 6, 9, 10, 14, 15, 21, 22, …
k = 3 -> 8, 12, 18, 20, 27, 28, 30, …
k = 5 -> 32, 48, 72, 80, 108, 112, …

I don't get what a k-prime is, honestly. If it is 1-inclusive, isn't 6 a 3-prime, and isn't 8 a 2-prime? If it is not 1-inclusive, aren't both 4 and 8 1-primes?
4: prime factors -> (1,) 2
6: prime factors -> (1,) 2, 3
8: prime factors -> (1,) 2

What does counted with multiplicity mean?

Did you finally debug it? Because I have the exact same issue with Python.
I pass all sample tests, though.

HIT! 5
HIT! 15
HIT! 2
HIT! 10
HIT! 4

The above is my result.

Alien [17] with [3] health
Alien [19] with [15] health
Alien [21] with [2] health
Alien [23] with [10] health
Alien [24] with [4] health
: 36 should equal 34

This is the reference result. So far, I've mostly overshot by 2 points, and always at the first alien to hit. :/
EDIT: My bad. I didn't read the hit-sequence instruction properly.

Why the hell is this a 3kyu? This is 5-6.

I can't see my solution in the solutions list, even though I can see the others. :|

That was definitely fun. Spent a whole day.

Pretty straightforward algorithmically. But it was tedious and the debugging process was long. 2 kyu!
I think you can make it more challenging by testing more invalid expressions.

Thanks for this!

My bad... thanks!

-(-2+3^2)^(5-3)*2 this expression expects -98.0.
Tell me if I am missing something here:

-(-2+3^2)^(5-3)*2 
-> -(-2+9)^(2)*2
-> -7^2*2
-> 49*2
-> 98

Any even-power of a negative number would be positive.

Solved it!

After a bit of hair-pulling, the answer turned out to be quite easy.
Now, to do it in Java!

if (size > 2 && size%2 == 0) { answer[size/2][size/2-1] = 1; }

The test cases are wrong for all even sizes > 2. You should use the above code just before you return the answer.
This is not a spoiler - this is an essential piece of code to finish an annoyingly buggy kata, even if your code is correct.

if (size > 2 && size%2 == 0) { answer[size/2][size/2-1] = 1; }

The test cases are wrong for all even sizes > 2. You should use the above code just before you return the answer.