Personally, I have never found much of a reason to use the Y combinator. I typically use the method you used in the second example; I declare a named function within the scope of the anonymous one. I find it to be a readability win.
But I agree with you that Y combinators are not generally a good approach for working programmers because there are clearer ways of accomplishing the same thing. However I've heard that in some functional languages they can be very useful as a pattern to compile language constructs to.
a = (function(x) {return x == 0 ? 1 : x * arguments.callee(x - 1)})(8);
builder = function(x){ return function(n){ return x(x)(n); }}
factorial = function(f){ return function(n){ return n==0?1:n*f(f)(n-1);} }
alert(builder(factorial)(8))
r = (function(x){ return function(n){ return x(x)(n); } } (function(f){ return function(n){ return n==0?1:n*f(f)(n-1);} }))(8);