In the May 2013 issue of the magazine, Sklansky brings up the odds of rolling two dice and seeing a four appear. He mentions the need to subtract the odds of a four appearing for both dice, i.e. 1/36. He then says that to figure out the odds of one four appearing, 1/6 + 1/6 - 1/36 = 11/36.

When I write out the possibility tree, it appears as though there are actually 10 possibilities rather than 11. The tree shows six columns of fours and six rows of fours. So before even considering whether to include the "4,4" square, we have to make sure we do not count that square twice since it appears in both the 4 row and 4 column. When we do subtract "4,4," we end up with 10 possibilities of rolling a 4 rather than 11.

This is the way it appears to me. Am I missing something in the math, or am I misinterpreting the semantics of the problem?

Never mind---this was a semantics error on my part. He was referring to "at least one four" rather than just one four appearing. Dumb oversight.