# Probability of $7$ white chessboard squares without neighbours

If we start with a $2\times 30$ chessboard and we remove $15$ black squares, how can I find the probability that we get $7$ white squares which don't have any neighbours? (A white square will have no neighbours if all of the black squares adjoining it are removed).

• I assume that the black squares are selected uniformly, but are you looking for the probability that at least 7 white squares are isolated, or that exactly 7 white squares are isolated? – Peter Taylor Jun 12 '13 at 9:19
• Excactly 7 white squares are isolated.... – user82063 Jun 12 '13 at 9:31
• I would start with a smaller problem. Solving it might give you an idea of how to solve the original, or give you a feel for how hard the original is. – Gerry Myerson Jun 12 '13 at 9:40

If we label the columns where we delete the black square with an x and the ones where we don't with a . then we get a 30-character string containing 15 x and 15 .
Prepend and postpend x to handle the boundary cases (so that the end white squares are isolated respectively by a leading and trailing xxx)
Then each maximal non-overlapping substring of the form x$^{2+n}$ corresponds to $n$ isolated white squares.
1. Partitions of 7. There are 15 partitions of 7, of which 4 would require more than 15 deleted squares (e.g. $7 = 1 + 1 + 1 + 2 + 2$ requires three runs of three and two runs of four, for a total of 17 deleted squares), so that leaves 11 cases to handle. They might be grouped with some thought.