But apparently that’s not the case here, it seems they estimate the answer is correct by approximations from smaller cases.

For many problems, checking whether an answer is correct is way easier than finding the answer in the first place.