Last episode a question about the (value + 1) & -1 confused me and I turned to my debugger for comfort, but without success.

Why does ((value + 15) & -15) align the value to a 16 bit address? I thought Casey used this macro in HH and mentioned it last night, or am I wrong?

But -15 in hexadecimal is 0xfffffffffffffff1, and I'd think it should be 0xFF...0 to work properly.

The results I get for the following values are not aligned to 16:
((0 + 15) & -15) = 1
((16 + 15) & -15) = 17
((14 + 15) & -15) = 17

Obviously I'm doing something wrong or don't remember the macro. What's wrong in my reasoning?

The macro for alignment is slightly different: (ptr + (alignment - 1)) & ~(alignment - 1).

For alignment = 16:
~(16 - 1) = ~(15) = ~(0x0000000f) = 0xfffffff0

So, your examples:
((0 + 15) & ~15) = 0
((16 + 15) & ~15) = 16
((14 + 15) & ~15) = 16

Oh thank you! ~15 = -16. :)