Fear of View: self-indulgent ruminations on the design

The original seed for the game was a comment by Michael Brough that his games are often built around a single mechanic taken from traditional roguelikes -- spells for 868-HACK, the identification minigame for Cinco Paus, and so on. Some years later, Broughlikes were on my mind thanks to 868-HACK, Magpie, and Seven Scrolls, and it occurred to me that "field of view", i.e. walls blocking vision, is an obvious unexplored candidate for a mechanic which could be built out into a game. Field of view actually wasn't a feature of the original Rogue, but it was already in Hack, and then most of the canon. In a good Broughlike, every turn involves a significant choice. Manipulating vision through the interaction of the positions of the player and walls easily produces intricate tactical consequences for each move. With a small grid, the basic idea of the dungeon persisting only while you keep it in vision seemed inevitable, and from then on I came down with a bad case of compulsive game design. Implementation was the only cure.

I put some thought into the structure of the game. Dividing a run into a series of rounds was something I was sure I wanted to copy, but I was also sure I wanted to do it differently. 868-HACK has no connection between the rounds in a run and simply sums the scores over a sequence of rounds until you die, with some randomised difficulty modifiers to make survival and scoring harder in later rounds. It works fairly well, but sometimes feels too much like flipping a coin and hoping for a run of heads. It also means that as you get better at the game, runs take longer. Cinco Paus has you working to power up by obtaining artifacts throughout a run, trying to outrun the steadily increasing difficulty so you can safely go for points. That works great, but it plays out over 50 rounds, which is just too much. Seven Scrolls has score essentially independent rounds, but the mechanics make playing for score tedious.

From 868-HACK, I wanted to take the flexible risk-reward tradeoff, allowing an experienced player to make fine judgements on when to take risks to obtain points more quickly. From Cinco Paus, I wanted the struggle to power up the player to keep pace with the increasingly deadly environment. From Seven Scrolls, I wanted the reasonably short run length applying even to experienced players. Thinking about this last aspect, I hit upon the idea of fixing the maximum number of points rather than the number of rounds, with runs which obtained the maximum points ordered with shorter runs higher on the leaderboard. That way, optimisation by skilled players would tend to decrease run length by finding risky opportunities to get points more quickly.

This need to make obtaining points difficult and risky led to the idea of objects being determined by where they are found, with score objects located such that it takes planning and hairy play to spawn them, compounding the opportunity cost of missing out on a directly useful item. This was also inspired by other Brough classics, particular Ora et Labora and P1 Select, where different grid locations have distinct properties. With this principle in place, having location-determined powers as persistent upgrades followed naturally.

The natural desire to have something special in the centre led to equipment as a secondary persistent upgrade; this is arguably redundant, but I wanted something which would really make each run feel unique, and the player obtains too many powers for them to be enough for that. It also felt natural to have passive abilities to complement the activated powers and the directional items.

Another aspect I struggled with was how to increase difficulty between rounds. 868 has its random rule-changing "upgrades", CP has each monster type gain random powers, and 7S has new monster types. I liked the CP approach of having persistent modifiers which develop over the course of a run, adding to the individuality of runs. The randomness of it can be frustrating though, so I decided to give the player some agency over the development by having the exit direction determine the new difficulty modifiers; thanks to the way the exit is spawned, this adds to the strategic consequences of vision manipulation. The details weren't obvious, and trying to communicate the mechanics to the player was the most difficult (and I fear least successful) part, but I'm reasonably happy with how it turned out. The idea of rolling creatures/walls from an N-sided die, with N fairly small, was included to make the mechanic drawable; by a happy accident, it gave a neat way to increase difficulty if you linger too long in a level, by decreasing N over time.