Talk About Network

Google


Register and Login
Nick
Password
Register create new account Sign up is FREE and you can post replies, new topics, bookmark posts and more!
Recover lost password


Gaming > Abstract (perfect information, pure strategy) > Draws in Crossw...
Latest [ Topics | Posts ] Archive Post A New Topic Post a Reply
<< Topic < Post Post 1 of 1 Topic 652 of 669
Post > Topic >>

Draws in Crossway?

by "marksteere@[EMAIL PROTECTED] " <marksteere@[EMAIL PROTECTED] > May 10, 2008 at 11:09 AM

The possibility of draws in Crossway hasn't been proved or disproved.

Anthony Kam (antkam@[EMAIL PROTECTED]
) has discovered some interesting
formations in Crossway, =93eyes=94 if I may borrow a term from Go.  There
is a formation with a single hole in which neither player can place a
stone, and a formation with a double hole (two adjacent unoccupied
spaces) in which only one of the players can place a stone.

H V .
V . H
=2E H V


H H V H V
V H . V V
V V . H V
V H V H V
H H V H H


It=92s Anthony=92s believe that a draw cannot occur in Crossway, but it
has yet to be proven either way.  The presence of the eyes certainly
would complicate any proof.

Here is an analysis of Crossway by Anthony Kam:

=93. . . there should be no game position where BOTH players cannot move
(but the game has not been won). =85I've worked on this for a few days
and still could not prove it completely. The best I could do was to
reduce it to a different (and I hope more intuitively obvious)
hypothesis below.

-----------------------------

Hypothesis: In any game board situation where H has not won, it is
possible to find a path P with these properties:
(1) P consists of a sequence of positions (spots) x1, x2, ... xN such
that successive positions are connected horizontally or vertically
(but not diagonally).
(2) Each position of P is empty or occupied by a V stone.
(3) P touches the top edge and the bottom edge.

"Proof" of Hypothesis: Consider the set of H stones which can be
traced (via a sequence of horizontal, vertical or diagonal
connections) back to the left edge. Call this set X. Consider the set
of the rest of the H stones and call that set Y. By definition, there
is no stone in X which is a horizontal/vertical/diagonal connection
away from any stone in Y. Then, "intuitively" there should be a "thick
barrier" P which separates X from Y. This barrier can contain V stones
or empty spots. Moreover, since X stones and Y stones dont even touch
diagonally, it should be possible to find such a barrier which
satisfies condition (1) above. I know this "proof" is not a rigorous,
but it's the best I can do for now.

---------------------------------

Theorem: Assuming the Hypothesis above is correct, then Crossway
cannot end in a state where neither player can move (and neither has
won).

Proof: It is sufficient to prove that if H has not won, then either H
has a move, or V has a move, or V has won. So, consider a board where
H has not won. By Hypothesis above, we can find a path P with
properties (1), (2) and (3).

If P consists of all V stones, then V has won.

If P consists of some empty spots, then H and/or V will have a move,
unless ALL empty spots are of the following form:

VH*
H?V
*VH


Here, ? is the hole where neither can play. (* =3D doesnt matter) This
kind of hole (plus reflections/rotations) is the only situation in
Crossway where neither player can play to an empty spot.

By assumption, P includes the ? hole. By property (1), P also includes
the V stones horizontally and vertically connected to the ? hole. Now
consider a path P' which is the same as P, but excludes all such
unplayable ? holes. This new path P' now has diagonal connections
(near where the ? holes used to be), but it now consists of only V
stones (no holes), and still satisfies condition (3), i.e. it touches
both the top edge and the bottom edge. Therefore, V has won the game
via this path P'.

I hope some day, someone can come up with a more complete and elegant
proof!=94

And a follow up comment from David J. Bush:

=93Maybe you can prove that this path P exists with a "robot explorer"
like in the proof that Hex games are decisive. Call HL the set of all
H stones connected orthogonally or diagonally to the left edge (you
call this set X.) Start your robot in the top row, the rightmost cell
which is empty or V and is adjacent to an HL cell on the left. If no
HL cell exists in the top row, start in the leftmost cell of the top
row. The robot has two possible states: drop seeking, or left seeking.
Start the robot in the drop seeking state. From there, follow this
algorithm:

If the robot is drop seeking
Then (If the cell below is HL
Then move one cell to the right, stay in drop seeking mode
Else the cell below must be vacant or V. move down one cell and switch
to left seeking.)
Else the robot is left seeking.
If the cell to the left is HL or the robot is on the left edge
Then stay put and switch to drop seeking
Else move to the left and stay left seeking.

Repeat this process until you reach the bottom row. The path followed
may have some extraneous left branches, but should include the desired
path P. It shouldn't be difficult to prove the robot should always
reach the bottom.=94
 




 1 Posts in Topic:
Draws in Crossway?
"marksteere@[EMAIL P  2008-05-10 11:09:57 

Post A Reply:
  Go here to Signup

AddThis Feed Button


About - Advertising - Contact - Frequently Asked Questions - Privacy Policy - Terms of Use - Signup

Contact
tan12V112 Thu Jul 24 7:14:22 CDT 2008.