I spent the last two weeks visiting family and friends up north and it got me thinking about designing explicitly for an expected or actual player(s). Should a DM design world and rule settings that cater to a specific play style or type of player?
There's a number of ways to approach this question. One perspective would say that a DM should not really design towards a specific behavior. Rather, their own design style would attract and retain players that are ideally suited for such a game. Sadly, this is just a nonsensical intrusion of efficient market garbage into gaming. This outlook makes a wacky assumption that people are primarily playing the game that is most appealing to them at all times, and are switching games and groups whenever something more ideal comes along. Isn't it obvious that this is simply not an accurate picture of reality? There are a whole horde of social aspects to the game that easily trump any cold, objective analysis of "is this the best game world for me to be playing in?" So maybe we have to confront this overall question directly.
Ok, you do take an active stance, constantly asking "what type of play or player does this design choice attract or repel?" I would argue that this is foolish on a number of levels. How is a DM really evaluating what is driving a particular person to play in their game? Even more difficult, how do you typify a stranger and their desires? It is a common mantra through many service industries that says customers don't really know what they want.
In a general sense, you can control this selection quality with broad brushstrokes, such as using a modern D&D ruleset to attract players who like tactical combat as opposed to using Dogs in the Vineyard. Yet, when it comes down to individual decisions, can you really attract more cerebral, or more social, or more narrative (or whatever vague qualifier you want to use) players by tweaking rules and creating settings? If you create a land where everyone speaks in riddles, that is just boring design and even players who initially show interest will just tire of the headache and find ways around/through the issue.
If you have people who can all agree on the genre and ruleset, that is probably the most control you can exercise through design. If you really want to encourage or discourage a certain behavior, the ultimate method is through rewards and punishment. Give experience and accolades for combat while browbeating anyone who stops to roleplay and that will tailor a group to a certain playstyle better than using Shakespearean names for all of the NPCs.
So where does this leave us in relation to our original question? The best answer I can come up with is to design for the DM. Using settings, rules, and themes that are of the personal interest to the DM would be the best method to ensure a "good" game. Good in this sense, or any sense really, is highly subjective, but catering to a DMs interests means the overall product will be more carefully designed. There will be less chance of burnout on the DM's end and much more personal investment in the game. This is the best hook to attract players. Offer them a robust game, constant sessions, and an engaged DM and the players will share the passion.
In the end, it is the road I am taking.