Ronald L. Rivest held 2011 IACR Distinguished Lecture

Abstract

We model security situations where an Attacker can
"take control" at any time, but where the Defender can
also "re-take control" whenever he wishes, but with the
interesting property that a player finds out when or whether
the other player has moved only by moving himself.  This
models many security situations such as theft of passwords or
of secret crypto keys, and computer system compromise.

The resulting two-person, non-zero-sum continuous time game
with imperfect monitoring we call "Flipit", and study some
of its properties.  (This study is joint work with Ari Juels,
Alina Oprea, and Marten van Dijk.)