I was wondering that suppose if we have a to find the probability of a random variable having a discrete distribution using Central Limit Theorem, and that distribution originally is neither binomial nor Poisson..... Then should we apply continuity correction or not? For eg. Suppose we have S being a random variable of total payments received on a portfolio of 200 bonds and we know that the payments are either 10 or 20 with probability 0.01 & 0.99 respectively and we want the probability of S>2040. So here we may use CLT. Thus should we use continuity correction? If not then can you please tell in which cases only to use continuity correction? Hoping for an early response. Exams ahead ! Thank you
