You have to remember that this is still in beta. Does any 1 game really need 3 million beta testers?
I think you really start diluting comments that will help you advance the game and start caving to a mass of people looking for a complete game.
Markus had sold somewhere between 800,000 and 850,000 copies before it went into Beta. Does a game really need 800,000 or so Alpha testers? No.
However considering how the game has ultimately developed, it's best to ignore the naming conventions entirely, considering that -- bugs aside -- it's still a very enjoyable playing experience. "Beta" is effectively a discount. "Alpha" was an even larger discount, and the promise that all future updates would forever be free.
Notch has been in very close contact with the players of the game throughout development, and has a keen sense of what this mass of people is looking for.
If you can show that Mojang does TDD, 'story' card walls, velocity tracking or kanban pipelines or any of the concrete "developing practices" (to use your term) pushed by the agile consultant folks (vs some motherhood "values" that can be retrospectively applied to any successful project try say Google+ instead of MineCraft and you could call it an "agile" success story too) then you'd have more of a case.[1]
Otherwise you could label any successful project anywhere "agile".
[1] I am not saying Mojang doesn't do "agile". Maybe he does - in which case a comment that explained which concrete practices he found useful would be great. I am just tired of agilists latching on to any and all successful projects and calling them "agile" to the point where "agile" becomes a synonym for "successful".
I'm just speculating of course, as are we all. Just saying that from my perspective, they don't look like a team using agile practices. They don't look like a team using any kind of standard industry practices at all, really.
Disclaimer: I love Minecraft, play it all the time.
Honestly, I think this could be a great business model for indie game designers.
'Beta' in this situation is a modesty thing, I think.
I imagine most people, when the game is finally "finished", will respond something: "oh... that's cool ... so, what's ~this~ update changed?" It'll be little different to them from the many other updates that have occurred.
For Minecraft, I think, Beta is just a word. (though it does affect payment options, which I'm sure some people will be very sad from)
It's obviously not a "beta" version of the product in historical terms. It's a product that's being publicly used, with all of its imperfections on show, which is continuously being developed.
These are paying customers.
My fear is this is because of the 3 million beta users and grand views of the masses.
I guess my main point is that 3 million users might not be the best for a beta game. It might work in the end, but I think Minecraft might get bogged down with 3 million users making suggestions to a non-complete game.