# Why SIVP Is Worst Case Problem?

I just started to study lattice cryptography.

I'm now studying worst-case to average-case reduction for SIS.

In previous question, "worst means any and average means random".

And I wonder why Shortest Independent Vectors Problem(SIVP) is worst case problem.

And, does SIVP means that

"For 'any' given lattice basis B, it is hard to find n linearly independent short vectors in L(B)"?

The definitions that I saw didn't have the word 'any'.

Does the word 'any' is involved in the sentence even though it is not explicitly mentioned? I'm guessing that SIVP is worst-case problem because of 'any'.

If it is correct, does

"For given 'randomly' chosen lattice basis B, it is hard to find n linearly independent short vectors in L(B)"

is average-case problem?

• Yes, you are correct. To solve SIVP, an algorithm must work for any given input basis $B$. One can also formulate an average-case variant of SIVP, where the input basis is generated at random according to some probability distribution. – Chris Peikert Dec 17 '15 at 15:38