“…A game-theoretic approach to R-sets, closely related to this work, is developed by Burgess in [8] where the following characterization is stated as a remark without a formal proof: (1) every set A ⊆ X belongs to a finite level of the hierarchy of R-sets if and only if it is of the form A = (K), for some set K ⊆ ω ω which is a Boolean combination of F σ sets, and (2) the levels of the hierarchy of R-sets are in correspondence with the levels of the difference hierarchy (see [13, §22.E]) of F σ sets. The operation is the so-called game quantifier (see [13, §20.D] and [6,7,12,18]). Admittedly, our characterization of R-sets in terms of game tree languages W i,k , can be considered as a modern variant of the result of Burgess.…”