How this works. Two's complement stores a negative number as
2n − |value|, which in practice means inverting every bit of the
magnitude and adding one. The top bit ends up carrying the sign, and — the reason the encoding
won — ordinary binary addition then works unchanged for both signs, so hardware needs no separate
subtractor. The range is asymmetric:
−2n−1 to
2n−1−1, because zero occupies one of the positive slots.
The value is outside the signed range for this width, so it has wrapped — the same silent overflow a narrower integer type would give you.