Theories of equilibrium selection in non-cooperative games, as well as the
notion of risk dominance, depend heavily on the so-called linear tracing pr
ocedure. This is the first paper to give direct, simple proofs of the feasi
bility of the linear tracing procedure. The first proof utilizes a result t
hat is related to Kakutani's fixed point theorem and that is an extension o
f Browder's fixed point theorem. The second proof shows that it is even pos
sible to avoid the use of correspondences.