The problem is there's no way you can ever achieve this, except to make a game that is very close to devoid of (or completely devoid of) any meaningful mechanical choices at all. Consider BECMI. Basic is very close to this. You randomly roll 6 stats, and pick a class, then roll hit points. You have literally one choice, and then you get to buy equipment, which you could consider more of a part of play vs chargen really. It is a perfectly fine game, but even in AD&D where only a couple things were added (race AND class) you instantly have some optimization.
Once you have a system which has several categories of choices and a large number of valid combinations of such you really can't insure that all of the millions of possible combinations of those choices are all equal. You could simply introduce nothing but new options that are clearly inferior in some way and will NEVER equal the existing options, or you WILL end up with some degree of 'creep'.
The best you can really do is make sure that there's nothing obviously and egregiously out of whack about a new option. Put it in the 'middle of the pack' as close as you can and don't worry about it, and if the concept demands something a bit nice, or might not really be super incredible and remain true to the concept, well, its better to have a good concept.
So the game IS going to have outliers in terms of optimization. The worst ones on either side probably need to be quashed, and there's nothing wrong with bargle0 saying "hey, this can be closer to the ideal baseline with this tweak". He's probably right about that. The ultimate point is though, given that we play an option rich game like 4e, the game will be too complex to practically perfect every option. If the number of options available strikes someone as excessive then they need to think about whether 4e does what they want. You can certainly tell WotC you'd like them to make you a different game of course, but beyond that there are still plenty of options.