KNIGHT COVERINGS
FOR LARGE CHESSBOARDS
Frank Rubin
November 9, 2000
      
The problem of how to cover a chessboard with the smallest
number of knights has fascinated both mathematicians and chess players
for years. The problem is to place the smallest number of knights
so that every square on the board is covered.
That is, every square on the board is either occupied by a knight
or attacked by a knight.
      
The optimal solutions for boards of sizes 3x3 to 10x10, plus 12x12 and 13x13
have been known since the 19th century. The best solution for 11x11 was
found in 1973 by Bernard Lemaire, and Davis found the best solution for 14x14
in 1977.
      
On this site we present solutions for boards as large as 50x50.
The solutions for boards 20x20 through 26x26 were first discovered
in Oct. and Nov. 2000 by Frank Rubin (the webmaster of this site).
It is likely that some or all of these coverings are optimal,
but for boards larger than 10x10 proving optimality is
a much more difficult problem than finding a good cover.
      
As of August 2005, the solutions for boards up through 20x20
have been verified as optimal by Lee Morgenstern using exact methods.
He has also developed optimality proofs for square boards up to 13x13,
and for several rectangular boards.
      
To view the solutions click on any of the links in the box below.
Anderson H. Jackson, Roy P. Pargas, "Solutions to the NxN
Knights Covering Problem," J. Recr. Math. 23(1991),
pp 255-267.
David C. Fisher, "On the NxN Knight Cover Problem,"
Ars Combinatoria 69(2003), pp 255-274.
Bernard Lemaire, "Knights Covers on NxN Chessboards,"
J. Recr. Math. 31(2003), pp 87-99.
Frank Rubin, "Improved Knight Coverings,"
Ars Combinatoria 69(2003), pp 185-196.
Frank Rubin, "Knight Covers for the 50x50 Chessboard,"
Mathfest 2004, Providence RI.
Frank Rubin, "A Family of Efficient Knight Covering Patterns,"
J. Recr. Math. 33(2004), pp 165-175.
Send us an email for any comments or results about knight coverings.
Be sure to change the $ to an @ in our email address.