For my current project I intend to have an overworld, but find myself a little stuck for a design that fits my goals for it. In particular, I don't want this game to be a wide-open sandbox--I want to avoid both the issue of sandbox paralysis (what TVTropes calls the "quicksand box") and that of content creation on such a scale. On the other hand, I do want a degree of exploration and discovery--I want the player to be able to spot an interesting tower while travelling, and decide to visit it.
As mentioned, this is a overworld that I'm designing: the meat of the game is intended to be in the levels, with this overworld providing a sense of travel, an impression of a world outside the levels, and opportunities to discover new locations.
To these ends I've been thinking about having my overworld be composed of map locations: the player has a map, and can visit specific places in the area shown, rather than being able to trudge every centimetre of mountainside.
The problem then, is that of how to approach this.
I've thus far considered three options, each with their own advantages and problems:
- Each map location corresponds to a small "level" that the player can explore freely, as with standard levels.
On the postive side, this allows the player to discover secrets hidden around corners, or new vantages from which to spot potential locations to visit.
On the negative side, the boundaries--wherever they lie--might damage immersion by confronting the player with the limits of the "level". I've considered having the boundaries automatically produce a prompt to the player offering to "travel", taking the player either to the map or to the next location in that direction, but am not sure that this completely solves the problem.
-
Each map location corresponds to a fixed-position, free-rotation "node" (as seen in some of the Myst games, as I recall): the player can look around and interact with objects, but movement is limited to moving from node to node. The graph of nodes would be sufficiently dense that one could move in most (reasonable) directions and end up at a new node. In essence this replaces discovery by exploring a small region with discovery according to which nodes the player chooses to visit.
On the positive side, this allows for fully pre-rendered art (which has certain advantages over real-time 3D art), and can be relatively simple to set up.
On the negative side, this is quite apparently artificial--but, funnily enough, I've think that I've found that it can be less immersion-breaking than coming up against an invisible wall, as in option (1) above. Perhaps more problematic, I'm somewhat concerned about the amount of content that so dense a graph calls for, even allowing for variations in density according to the density of interesting features in an area.
(I've actually prototyped this idea; it seems to work well enough, but prompted me to notice the content-creation issue.)
-
This is essentially the same as number (2) above, but with a far less dense graph, and no free movement: the player would return to the map and click on a new node to travel to. This reduces the content-creation issue, but also reduces the element of exploration.
I'm currently somewhat leaning towards the third option--the second has the content issue, and the first has the "invisible boundaries" issue--but it pretty much cuts out the exploration element, which leaves me somewhat dissatisfied. That said, the levels should provide at least some exploration, and I could see some of them providing access to "hidden" map-locations...
So... What thoughts or advice do you have?