I have never used convolutions before, but I decided to take a look at them today. They seem just an application of conditional probability. Is there anything special about them beyond applying conditional probability and LOTP?
Are there problems (in the context of probability) that can be solved with convolutions but not with conditional probability?