The first test versions of Impossible Cave, which were only shared privately, were builds  with the goal of demonstrating a technically playable gameplay loop. The first one of these was exported in July 2022, a little less than 3 months after pivoting hard off the initial project idea (more info here).

I made maps for 10 floors, plus the entrance room that reset all your levels and items before entering. Using randomly rolled variables, I put chests and enemies on each of the floors. The 10th floor contained a final boss you could somewhat reliably defeat, depending on the items and levels you got along the way. Nearly everything consisted of either default RPG Maker assets or sounds/art pulled from previous game jam projects, because I was just aiming for playability. However, I did make and import an enemy pic for the Mite, which game from a much older fictional universe of mine.

I started noting the limitations of the editing software (without plugins). Your one human character got accompanied by three "companions" that were intended to not be directly controllable. However, the friendly auto-battle was so basic, I decided it would be better to have their actions fully controllable too. I wanted one companion to have a healing skill for companions only, but I couldn't restrict certain single target ally skills from targeting specific people. I worked around this by changing stats around. I gave the companions much higher base HP than the human character, but they could not equip defensive armor. I then made the healing spell also come with a temporary penalty to defenses. This meant that even though you could target human characters with it, there was much more risk.

"Test2" was exported in August 2022. In this version, I polished the 10 floor cave. I added a variable for Points, something to track your score throughout the entire game, as well as spendable Points, which worked the same way except you could spend them and the number could go down. I added special chests that pick from a pool of rare items.

In a direct response to feedback, I added small versions of floors that you had a chance to enter instead of the "normal" version of that floor. At first, I imagined that a large pool of maps would be selectable for every floor, but I could not find an easy way to pick from a very large pool without duplicates.

 I added a few more self-made assets, including monitors I drew myself as map events.

"Test3" was exported in November 2022. This version had 10 more floors, for a total of 20. The first 10 floors and the boss were rebalanced around this. I added skills for the second aura, Diamonds. I gave the cave a bit of a lighting tint for atmosphere.

In a couple nights, I added a silly side room to the Lobby where you get to pull for 15 different colors, each with different amounts of rarity, each play cost Points, and duplicates did nothing, as the first form of a Points sink.

(By the release of Test3, I renamed this "RefractOrb" after doing some name research, juuuuust to be safe)

The end goal was to make this a 100 floor cave, and to make it a challenging and satisfying experience, and I would work towards that in future versions. The direct inspiration of the structure of the cave was:

Lufia 2's Ancient Cave

Lufia 2 was a very cool turn based RPG on the SNES with quite the story. It also contained an entire optional town with an entire optional dungeon completely unique from everything else: The Ancient Cave. Each time you enter it, your levels and (most) items from the main game are temporarily held until you exit, and you traverse through 99 floors with random rooms, treasures, and progressively tougher enemies (both ones from the main game and unique ones to the cave), to see how far you get.

This side mode was super addicting and fun, but, without going into too much detail here (maybe a future post?), there were some things I always wished were different about it. And instead of waiting for the game or remake that appeals more towards me that may or may not ever be a thing, I figured... why not try to do it myself? The plan super early on was just to see if I could mimic the AC's structure, but I couldn't implement everything I wanted (I so wanted to copy the really cool movement system, but it's just not possible unless I gain the knowledge to drastically edit the source code to do so). It's fine, though, because this way I could put my own creativity and visions on top of a solid base.

Once I understood the realistic limitations of what I was working with, I then started to form a plan of how difficult I wanted the cave to be. The inspiration for the challenge factor of the cave was:

Self-imposed Challenge Runs

When players run out of challenges imposed by their favorite games directly, they sometimes find ways to raise the longevity of it by creating their own challenge (speedruns, restricting yourself from doing "easier" things, etc). That's fine, but... 

I'm not sure how many others out there share my view, but it just feels so much more interesting when that tougher challenge is given by the game itself. And for the devoted, the challenge keeps scaling up more and more until you have a mode that only true masters of the game have any chance at it. If those challenges are only enforced by the player itself, and the game is not in a position to adjust itself around that (basically every game that is not receiving or can't receive updates), it can introduce issues that will forever persist. Speedrun goals of "get this really good time" can work great, but then the goal of "get the absolute best possible time in a single run" can get pretty frustrating if things you have no control over have an influence on your end time.

I watched a lot of interesting streams of old Pokemon games but under the self-imposed ruleset of a challenge to make it much more likely to lose and/or invoke a total restart. Even though the rule list was pretty comprehensive, there were still the occasional ambiguous edge cases that come down to personal judgement, or situations where the player accidentally did a "banned" thing or, due to those same edge cases, was forced into doing the "banned" thing, and then has to consider what is the appropriate response needed to keep the run as "qualifying", or just restart. If the game itself follows those rules, these awkward cases won't happen. The game will either allow it, or it won't. If it unintentionally allows it, the designer can then make the judgment call whether to change the game or not.

I believe all the content in a game should be able to be accessed by anyone. I believe mastery should be available to those who seek it, and should that be accepted, nothing should be held back. Any game I ever design will follow that code, including making something nearly (but not totally) impossible.

The next post will talk about version 0.1, the first "public" linked version on here, and my eternal fight between meeting my own impossible standards and putting something out there to share.