There has been some miscommunication if it is believed that the Stable_games category means that the games are not much played! I have been opposed to this hierarchy from the beginning due to just this sort of problems.
I think this category should be reserved for genre defining roguelikes. I do not have any clear criterion for determining this (which is why I think trying to categorize this way is foolish and likely to just lead to bad feelings), but I do not believe ToME is there.
One approach to make this more systematic would be just to limit Major Roguelikes to those with USENET groups. This would remove Crawl, but if it allows us to move forward without needless infighting, I'm all for it.
Perhaps the category should be renamed to Influential Roguelikes, where the game's influence is defined by how much the game is being copied by other popular games and having other popular games derived from the game.
This would have a bias toward games with sources available and licensed to allow derived works, but that seems reasonable to me. The more open a game is, the more it CAN influence other game developers.
So, if we look at the tree that was just put on the Major Roguelikes page, everything with a child node influenced its children, and should count as an Influential Roguelikes. List the leaf nodes in the Stable section, including Adom, Crawl, SLASH'EM. ZAngband seems on the border to me, since it seems to have spawned as many variants as Angband has.