Classic RPG Dungeon Design

Attrition

Dungeons are based on attrition. The gameplay loop of an early RPG dungeon is not to barrel through them. A dense dungeon is going to have branches and loot which make finding your way less than trivial. On top of that, the encounters whittle your HP down, forcing you to expend items or other resources (MP) to keep it up. There eventually comes a point you are in a danger zone without reliable means to heal, and must either chance going a little further or turning back now to not get wiped on the way back. Status effects that persist between battles may also weigh on this, such as the classic antidote to restore poison, lest you take damage every few steps.

Dungeons are also not level-by-level scenarios. They are not necessarily done in sequence. They can be left and returned to, or visited out of order. Chests that had items are now opened, equipment you obtained stays with you, and experience gains from combat remains. Your player knowledge is also greater, as you may have identified dead ends, or felt out the encounters to see what is most threatening. Your knowledge means you can restock items, fully restore resources, and try more efficient routes.

This is plain and obvious design, but it bears stressing because designing dungeons with these fundamentals in mind allows for more mindful dungeons. The core of a classic turn-based RPG is resource management, both in and out of combat.

Field Utility and Mysteries

Related to dungeon attrition, an under-utilized feature in the classic RPGs was bringing more out-of-combat utility to the table. Dragon Quest offer some of this. Locked doors required keys to open. Keys occupy an inventory slot. Thus, the player must decide if and how many keys to bring with them to a dungeon. The more they bring, the more they can unlock as they discover them. Otherwise, they must take note and return with a key later. This can also be used to change the layout of the dungeon. A player can be rewarded for opening a locked door by opening a shortcut, making later returns quicker, or giving access to a treasure trove with useful equipment. Additionally, this can all be applied to skills, should there be a skill that interacts with world tiles.

This extends further by having puzzles involving using items or skills on specific items. A dungeon can have context-specific spots that use items. This can be hinted at with dialogue and text dumps. But it should also be potentially solvable on its own, especially if there are no text hints in-game. The partial obscurity of secrets also adds to the density of the world, making even simple maze dungeons feel larger and mysterious. Most notably, this allows for a group of players to call out something they may have stumbled into, should a few players running through the game.

Crop and Claw supports this level of interaction. Many of the specifics will be kept shuttered until release, though. Secrets wouldn’t be very secretive if the developers just posted everything.

…But The Bag Was Full.

Some folks are not a big fan of limited inventory in classic RPGs. There’s a difference in how one approaches a dungeon in early Dragon Quest compared to Final Fantasy. There are some middleground cases. The Gen 1 Pokemon games had a 20 item limit, but items can stack up to 99 without glitches. But I want to focus on the polarizing opposites for this.

In Final Fantasy, by the point you reach endgame, it is often easy to purchase a large quantity of healing items due to the low gil costs and few gil sinks. 99 of the lowest tier potions isn’t very efficient in combat compared to strong magical healing. But it can heal a lot of damage out of combat. This greatly reduces the attrition of a dungeon, and removes a lot of the dungeon’s intimidation factor. In order to restore it, dungeon difficulty would have to be cranked up to expect the player to have an entire shop’s inventory going in. If the game’s design is intended to be easier, this isn’t a big deal.

In Dragon Quest and Mother, dungeon difficulty is partially influenced by the fact you have a tight inventory limit going in. Your inventory is limited in quantity, and items do not stack. Suppose a player enter with four medical herbs, one antidote, and no healing magic. You can safely assure that you have up to 120HP, and they can restore poison once. You can assume they will either cure poison immediately or allow it to persist, since enemies can’t reapply poison if you have it. You can also distribute loot with the assumption of them having a few slots free, should the player find it.

As a designer, you have an easier time controlling the player’s ability to resist a dungeon’s attrition. Combat can be toned to throw lower damage numbers, as the gradual decline cannot be so easily undone. Inventory slots are a premium. Additionally, if an enemy’s regular attack is chunking three-fourths of a character’s HP, it’s easier to gauge that a region’s encounter is still too strong for the player, as they cannot so freely restore HP with items. They must use limited use skills which costs another resource instead, such as MP or spell slots, which could have been used for utility or offense instead. Once those resources are depleted, the dungeon’s attrition has done its job and the player now must decide whether to leave or chance a little longer.

The main issue with limited inventory stems from when a game’s inventory is too small for the amount of items that cannot be removed. This is possible to remedy through an item storage, or by just designing to not have so many key items in the first place. Crop and Claw will remove equipped items from the inventory, lest four slots always just be occupied by that. Mother didn’t do this, and you spend a long time with only Ninten in the party. This leaves the small per-character inventory even smaller.

It is also notable that in some games, such as Mother and Dragon Quest, your inventory expands with new characters in the party. Final Fantasy does not need this system, since it is lighter on item restrictions and all items are shared across the party. This is an interesting dynamic since allies, which already make combat more manageable and interesting, become even more valuable since more items can be carried. This also, however, means that healing must be shared across other party members, and thus the expanded inventory also means items may have to be spent as well.

This is the reason that we use the Dragon Quest inventory approach in Crop and Claw. Party management is not currently relevant in our first entry, however it is planned for sequels, when the game’s framework has been solidified and is ready to be expanded upon.

Foundation of Simplicity

All the main points I bring up are relatively simple concepts. Once the basic code infrastructure is in place, it’s not even difficult to add the specific edge cases. In Crop and Claw’s case, it is designed akin to an NES age JRPG. These kinds of games have done all of these things to some degree. Mother even had entities that responded if Ninten casts Telepathy on them, which expands on the Talk/Check distinction. These could all be done by a modern RPG made on the NES in theory, limited only by the 1MB ROM size and the WRAM limit for individual event complexity. Should Crop and Claw enact this design framework correctly, then the player that likes this type of game should enjoy it, and it allows us a solid framework to build more complex projects on in the future.