Unlike D&D, where the DM's judgment is encouraged, in magic there are actually wrong and right answers. Many old cards actually break when the complete rules change and just don't work as intended, and in a tournament with a high rules-level enforcement these 'broken' rulings are enforced. Also, there's a specific hierarchy within magic judges, from the level 1s (you need to be damn sharp just to get to this, the lowest rank) all the way up to the level 4 (or is it 5s?) and finally the actual 'Rules Manager.', who's an R&D employee. I'd be surprised if Magic customer service didn't have a decent way to escalate magic questions, because that's exactly how it should work.
Grey areas in the rules are meant to be determined by the DM in a D&D game. I think there really is a correct answer for many very tricky rules questions, but that answer is 'It's up the the DM.'
Background: Before I landed a programing gig I spent 4 years in phone & E-Mail tech support, including trainer/supervisor/technical escalation agent. I've supported everything from sprinkler systems to software to internet service.
These people are customer service, first and foremost. I'm sure good history in the CS field will land you a job there a lot easier than being a huge RPG nut who knows every rule. And that job is customer satisfaction. Most of their customers don't go to the internet and debate the answer they got with other customers, most will just accept it. So as long as they don't get an outright wrong (i.e. gamebreaking) answer, things will usually work out fine.
We're not asking common questions when we debate things like the intricacies of stealth with full-page flowcharts. Heck, we don't care, because we know how opportunity attacks work. I bet if you send them a test of 10 'easy' questions three times worded three ways you get very nearly the same response. Because with any e-mail work in this field you (should) have an experienced agent identify the top X questions and then come up with pre-written responses. You do that for your call drivers, too, constantly training things that generate the most calls.
So, all and all, I would guess their CS does their job well 95% of the time. I'm not perverse enough to test the waters. Having a better chain of escalation would be very nice; but this is D&D. I'd rather they focus efforts on getting the rules just to not contradict each other and balanced than defining grey areas that I can make a judgment call on when I'm actually playing.
(And why haven't they offered Hypersmurf a position as the equivalent to the Magic Rules Manager, speaking of removing internal contradictions?!?)
