My third category is "too much mana too easily," cards that generate mana at a higher ratio than is healthy. This is different than "too much mana too quickly." It still takes some investment for this third category, but instead of mana getting additive, it gets multiplicative. The banner-carrying card here is Cabal Coffers. Cabal Coffers leads to fifteen mana on turn 7 tomfoolery, and I'd prefer to see games develop a little more slowly. Once the game turns the corner on turn 10 or so, I'm fine with it creating 50 or 100 mana, but that explosion in the mid-turns leads to yawn-inspiring games.
The card that doesn't go here is Gaea's Cradle. I think the situations that it creates too much mana too early or too much too easily at the wrong time are rare enough to not be a cause of concern. Yes, it can happen, but there has to be a sufficient star alignment for it to come about that I don't see it as worrisome (ditto for Serra's Sanctum). I'm not quite sure about Nykthos, Shrine to Nyx. I suspect that it's in the Gaea's Cradle category, although closer to Cabal Coffers. I'd want to see more evidence of brokenness, so I wouldn't make it one of my forced ten.
I agree. For Cradle to power out an game-winning Tooth & Nail entwine before Turn 4, you need to have spent turn 1 and 2 on something like Magus of the Candelabra and a card that creates 3 other creatures. And you have to have both Cradle and T&N. Speaking of God-hands, that's 5 cards + 2 land necessary by Turn 3. So essentially, it's not happening before Turn 4. After Turn 4, that's more than enough time for Control in this format to set up with a 3cmc counter or removal. And at that point, it really doesn't matter what type of Control it is, either. It just needs high enough answer density to have 1 or more counterspells or removal cards < 3cmc by Turn 4. That is not hard.
Sol Ring and Mana Crypt don't rely on setup, and can turn a really high number of hands into game-breakers without trying. And whatever they do, they are always broken before Turn 4.