Chomp is a two-player impartial game played on a board in the form of a rectangle, usually pictured
as a chocolate bar with one poisoned corner cell. Number the rows from top to bottom
and the columns from left to right, starting at 1, with the poisoned cell at . A player chooses a remaining cell
and removes it together with every
remaining cell
satisfying
and
. The player who removes the poisoned cell loses. The
rectangular version was analyzed by Gale (1974).
Every initial board with more than one cell admits a winning opening. To see this, consider the move removing only the cell from an
board. If this move is winning, the assertion follows.
Otherwise, the second player has a winning reply. Making that reply directly on the
original board produces the same position, since any move also removes
. The first player can therefore make a winning opening.
This argument establishes existence without identifying a winning opening (Gale 1974).
For a one-row board of width ,
the unique winning opening is
,
leaving only the poisoned cell. For a two-row board, the positions from which the
player to move loses have row lengths
with
. Consequently, the unique winning opening on a
board is
. On an
square board with
, the unique winning opening is
. The first player then answers each move
by
(Brouwer et al. 2005).
Gale (1974) reported that all
boards with
had exactly one winning opening and asked whether winning openings were unique in
general. The general assertion is false, with an
board providing a counterexample (Brouwer et al.
2005). Sheiner (2026) proved that every
board, for positive
integer
,
has exactly one winning opening, settling the three-row case of Gale's question.
This result concerns initial boards and does not assert uniqueness for arbitrary
partially eaten positions.
In the three-row case, write a position as , where
are the remaining row lengths. No winning
opening on
is made in the first row. An opening in the second row produces
, while one in the third row produces
, with
and
,
respectively. The widths admitting second-row openings begin 1, 3, 4, 6, 8, 10, 11,
13, 15, 16, 18, 20, ... (OEIS A029900), and
those admitting third-row openings begin 2, 5, 7, 9, 12, 14, 17, 19, 22, 23, 26,
29, ... (OEIS A029901). Sheiner (2026) proved
that these two sets are disjoint and together contain every positive
integer. The first sequence is indexed by
in the losing positions
and is not necessarily increasing (Brouwer et al.
2005).