This article describes a way to split a square into three non-overlapping regions that all have the same border. It is very counterintuitive that this is possible. If you imagine a line on a square, you can define two parts of the square that share this line as a border – and how could there possible be three?

Hope that helps a little bit.