SUDOKU is an interesting combinatorial structure embedded within a Latin Square. It has been
gaining popularity as combinatorial puzzle. Website provides plenty of examples of such
SUDOKUS. It is well known that Latin Square Designs [LSDs] go beyond CRDS and RCBDS
in eliminating three external sources of variation in the experimental units [eu's] in an ANOVA
set-up. We examine the possibility of viewing SUDOKUS as experimental designs, going one
step beyond LSDs. It is contemplated that one additional [fourth] component of variation can be
suitably accommodated in a SUDOKU, though by sacrificing orthogonality. A detailed statistical
analysis is provided along with the underlying ANOVA Table for such designs based on
SUDOKUS. It is envisaged that the SUDOKUS will provide extra dimension of utility as
experimental designs.