I agree, but I would rather WotC make the effort at balance, leaving me to engage in occasional GM intervention when something goes wrong, than change to an approach that requires me as GM take responsibility from the get-go.The problem or issue with magic items in 4E is that they are essentially a secondary leveling mechanic, where gold = secondary XP. The sheer variety of items and combinations available has made balancing and testing their different combinations very difficult for Wizards to successfully do
The bottom line is that I want my players to build their PCs. That's their job, not mine.
I don't think it's about discouraging a toolkit mentatlity. Or about Rule 0. Rather, it's about "whose toolkit" or "whose rules". The game is everyone's, not just the GM's. So the ruleset in play has to be something that everyone agrees on.In my opinion, the player entitlement issue, initially, came from several different factors:
<snip>
With 3.5, we did, in my opinion, start to see more of a change from WOTC. The removal of Rule 0 was a big one. Then, the designers moving away from encouraging the toolkit mentality (e.g., Andy Collins in Sibling Rivalry column on WOTC's site telling DMs to to try and say, "Yes" and make room for player requests for races, classes, etc.).
Now one way to get everyone's agreement is for the rules to say it's the GM's job, and then rely on the GM not to make decisions that will drive the players away. This is the traditional D&D approach.
The alternative is to publish rules that work pretty well as a whole, and rely upon the play group as a whole to make decisions about which elements they'll introduce into their own game (eg if everyone things warforged are silly, then no on will play one; if everyone things iron armbands are broken, then they'll agree to exclude them from their game).
In practice it's always likely to be the GM who initiates these sorts of discussions. But it is meaningless for the rules to try to specify that the GM has the final say, when of necessity it's a group thing.