Ludum Dare #29 just started, and I’m jumpin’ in by blogging my first thoughts.  The theme is “Beneath the Surface”.

One thing that comes to mind is minesweeper, where information is hidden from the player, just beneath the surface.  Literally, given the metaphor.  (Am I allowed to use a metaphor as justification for calling something literal?  Probably not.)  Reminds me of a similar puzzle game I tried to make using mazes.  Your objective was to discover enough of the maze structure to find your way through to the goal.  The only information visible was how many walls each cell contained, or how many cells each cell could reach in two moves, or things like that.  Too bad that every variant I tried was either too hard (too much random guessing because information was straight up too incomplete), or too easy (it was simple to determine what was safe).

That was a while ago, I could try to revive those ideas and see if I can find new life in them.

One idea that I think is curious focuses on the reveal at the end of a successfully completed level.  It would be cool if the solution does not depend on a perfect understanding of the hidden information, but just on a rough mental model.  That way, when you beat a level, you can discover how correct your mental model was, or how lucky or off track you were.  I’ve been playing a lot of SpaceChem recently, so my brain is naturally going toward dynamic systems.

Obviously there needs to be some visible information propagated from the hidden system.  And it’s best if not all information is immediately propagated, but instead that the player has to incrementally reveal the information, and the better their mental model is of the hidden system, the more quickly they’ll be able to reveal the information.  But that kind of requires that scattered bits of information are available, not just information localized to a small portion of the hidden system.  Otherwise, it’s just a tedious slog of figuring out the first step, then the second, all the way to the last.  Instead, the big picture needs to be present in the visible information from the beginning, except it is really really fuzzy, and player choices can gradually make that information more focused.  Bad choices do little to improve the picture.  (Maybe they could even generate misleading information?)  Good choices reveal a lot.  Score could perhaps be the number of actions taken before a full solution could be realized.

An metaphor that is coming to mind is a disease that acts in somewhat unknown ways, but has complex symptoms.  The first goal is perhaps to treat the symptoms.  That would be the goal of the game, to understand the disease’s operation well enough to effectively combat the symptoms.  Maybe there could be a way to even attempt to cure the disease, but it’s costlier and riskier, and requires a more accurate mental model of the disease, but gives bonus points, or counts as a harder challenge.

One difficulty for gameplay is how to introduce people to the game, when the core game mechanics are hidden from sight.  Perhaps the way one treats the symptoms is by generating a virtual model of the disease and execute that model; symptom treatment is applied based on that virtual simulation.  That way, the player is working directly with the same mechanics that are simultaneously hidden.

Taking the disease metaphor further, there’s also the fact that diseases start small, and then grow in impact over time.  If you can catch it early, you can defeat it quickly, but you have less information to work with at that early stage.  Also, there might be some tradeoffs between managing symptoms and probing for additional information.  Some actions might leave symptoms to linger (or perhaps even accelerate the disease) but provide value information about the disease in the process.

<thinks for a bit, pulls out and doodles on a whiteboard>

Ahah, I think I have an idea.  A “disease” will simply be a single hidden object that moves around on a hexagonal grid.  Time is divided into discrete turns.  During each turn, the disease moves a fixed number of steps from its prior location to a new location.  Movement is entirely dictated by hidden fixed channels in cells.  Each cell has three distinct channels, each channel linking one edge of the cell to another edge.  (A hasty count suggests that there are only five arrangements possible, ignoring rotations; with rotations I count 15 distinct arrangements.  This feels good to me, simple components, but when combined as multiple cells, hopefully a fair amount of complexity.)  Each time the disease lands in a cell at the end of a turn, the location is revealed to the player, and the cell’s arrangement of cells is completely randomized.

The player needs to note the disease’s past locations and attempt to infer the arrangement of channels of the cells that it passed through, and eventually attempt to accurately predict the disease’s location before it moves there.  Some additional requirements might be included for winning, such as predicting the location multiple times (that is, give the disease hit points), or requiring that the destination’s channel arrangement be known (at least the one that the disease was actually traveling on).

The number of cells that the disease travels could be a major factor on the difficulty.  Short paths are easier to predict (too bad the disease destroys the channel arrangement; that could mess up a player’s attempt to figure out a local area).  Long paths are difficult because the disease’s final location could have been influenced by numerous cells.

Perhaps instead of guessing where the disease will land, the game will auto-calculate that guess given the model of channels that the player has already guessed.  For each of the channels along this guessed path that are in fact accurate, and that the disease actually passes through, the disease takes damage.  If the player has not guessed enough cells, then the game predicts as much as it can, and requires the player to fill in the missing segments of the guessed path, which would participate in causing disease damage.  Whether or not the amount of damage caused would be visible, I’m unsure.  Making it visible might be important both to let the player assess the accuracy of the guess beyond just what the disease’s final location is, and also provide a sense of progress (how close is the disease to being killed?).

As for erasing the destination cell’s channel arrangement, that could be controlled some by predicating it on how much damage the disease takes that turn.  If it was harmed a lot, it goes into panic mode and destroys the cell.  This could also be interpreted as side effects of the drugs used to attack the disease.  This would provide a counter-balance on the player.  The better the player is doing, the harder the disease makes it to finish it off.

Thinking about this design, I just realized that it might actually provide a type of game play where the player pretty much always wins.  The only question is how quickly.  Not sure what to make of that, but it could be particularly appealing to casuals.

Okay, time to dive into code, I think.  I’ll bother cleaning this post up later.

No Comments

Leave a comment