1.js

> var name = 'World!';
undefined
> (function () {
...     if (typeof name === 'undefined') {
.....         console.log('Goodbye ' + name);
.....     } else {
.....         console.log('Hello ' + name);
.....     }
... })();
Hello World!
undefined

2.js

> var name = 'World!';
undefined
> (function () {
...     if (typeof name === 'undefined') {
.....         var name = 'Jack';
.....         console.log('Goodbye ' + name);
.....     } else {
.....         console.log('Hello ' + name);
.....     }
... })();
Goodbye Jack
undefined

3.js

> var name = 'World!';
undefined
> (function () {
...     if (typeof name === 'undefined') {
.....         console.log('Goodbye ' + name);
.....     } else {
.....         var name = 'Jack';
.....         console.log('Hello ' + name);
.....     }
... })();
Goodbye undefined
undefined

Good one ;) But here comes the Mr. Know-It-All explanation:
JavaScript defines it's locally scoped variables (defined with var) always at the top of a block. In examples 2 and 3 an undefined name variable gets defined after the (function(){ part. And because of this it always ends in the Goodbye part.

@tschortsch Yes, that's the right explanation, but no excuse for such a batshit crazy language design :)