In many countries around the world, there is a crisis in the teaching of mathematics and computer science. Governments have tried to address the problem by investing in computers in schools; when they should have invested in teaching computer science in schools. Formal methods bridge the boundary between computing and mathematics in a natural way. Through our experience of teaching algorithmic thinking in schools, young children have been observed using concepts such as refinement, proof, abstraction, complexity, non- determinism, equivalence, etc. . . in their own reasoning about problems. We argue that this ability needs to be better leveraged in order to improve both the teaching of mathematics but also to improve childrens' understanding of computer science as a discipline in its own right. We give concrete examples of the type of formal methods teaching that succeeds.