It's actually wrong. The author is rounding.

1/3 = .33 repeating. .99 repeating gets close to 1, but never actually becomes 1

He's using the same assumption that 1.68 can be rounded up to 2. 1.68 does not equal 2