One major complication I came across when designing levels for Slide Knight was keeping the puzzles sufficiently challenging as the player progressed through the game. The original sliding puzzle rooms in JRPGs were always a very small part of the game as a whole, meant to confound players for a few minutes at most. I discovered early on that after a player had spent more than a few minutes thinking about the dynamics of sliding floor puzzles, they quickly began to deconstruct the mechanics to a degree that trivialized the task getting from one place to another on a sliding floor.

The core of this problem boiled down to the number of options a player can choose from at any given time. With a typical situation in an ice-floor puzzle having only three possible options (the four cardinal directions, minus one blocked off by a wall), it becomes pretty easy for a player to collapse the decision tree once they can identify some of the fundamental patterns that occur in this type of puzzle.

In order to remedy the issue, I took two main approaches:

• Adding new mechanics to keep the player on their toes: Blocks that would change state when you collided with them or passed by them, teleporters, objects that required some amount of "runway" to interact with, etc. The more mechanics I introduced, the longer it would take for a player to deconstruct all the patterns that could emerge and reduce the solution of a new problem to rote application of those patterns.
• Complex level design: As the levels progress, layouts rapidly increase in complexity (in hindsight, probably too rapidly). In order to produce depth in a game with such few decisions to make at any given time, I compensated by making the total number of decisions required to solve any individual level relatively high.

Early playtesting for Slide Knight showed fairly disappointing results -- Most playtesters didn't really enjoy the puzzles I had designed. I made several changes to the design, from thematics to onboarding to tweaking individual levels before ultimately coming to the conclusion that complex sliding floor puzzles are simply not the kind of thing that most people enjoy.

The core design of Slide Knight involves tapping into a very specific kind of thinking -- one that involves juggling a lot of possibilities at once and reading fairly deep into a decision-tree. This is a very niche type of mental exercise that many people seek to avoid, finding it taxing and stressful. While this conclusion was fairly disappointing, I found it constitued an important lesson into human psychology and helped refine my approach to future puzzles.

At one point I considered simplifying the overall direction of the game, designing levels to achieve a kind of distracted zen instead of the more intense brain-straining puzzling that I had originally intended. Ultimately I chose to maintain the original vision for Slide Knight simply because it was the kind of game I wanted to make.

Fun fact: Slide Knight was originally not Knight themed at all, instead taking a more minimalist vector aesthetic -- the theme was added after-the-fact in an attempt to keep playtesters more engaged.