Most importantly, don’t have an answer already in mind. GMing Dogs is a different thing from playing it. Your job as the GM is to present an interesting social situation and provoke the players into judging it. You don’t want to hobble their judgments by arguing with them about what’s right and wrong, nor by creating situations where right and wrong are obvious. You want to hear your players’ opinions, not to present your own.-- Dogs in the Vineyard, page 124
There are a number of interesting parallels here to what high-quality D&D play is for me. In a sandbox game, you don't want to hobble the choices of players in terms of where they should go or how they should approach a problem, nor create situations where there is a clear optimal solution. There should always be interesting trade-offs, whether those trade-offs are spending more time to be more careful or taking one route rather than another.
I've yet to play Dogs, and it's always hard to get a sense of how an RPG will play just by reading it, but this looks to me like the social hazards are somewhat isomorphic to the D&D dungeon. Not in a plot sense, as is sometimes discussed (where dungeon-like event flow charts are created), but more like a set of potential connections between various NPCs and their interests.
In D&D, making a physical choice is almost always what engages the risk/reward system, whereas in Dogs the act of judgment engages the risk/reward system. Judgment is more abstract, which is perhaps why the stake setting system is required. The same kind of thing is going on in D&D, it's just that what is at stake is usually more obvious, and so requires less mechanical centrality.