I don't understand why this is more useful than just creating an object and settings its prototype to `Error`.

Anyone care to explain?

Well, in that case you'd lose the stack trace. If you just set the object's prototype to Error.prototype, and you add additional properties to the object's prototype you'd be modifying the Error's prototype as well.

The aim of this helper is to create a custom object which has the same functionality as a traditional Error object.