We introduce a model of random chain complex over the reals. The main question we are interested in is: what does the homology of such a random chain complex typically look like? We show that under various hypothesis, homology tries to be as small as possible. This is joint work in progress with Ayat Ababneh.