Hi everyone, and welcome to another Dandelion Void devlog. Today we are talking about improvements we’ve made to the looting and scavenging experience in Dandelion Void!
The Balancing Act
When a game is early in development like ours, we have to be careful how much time we devote to game balance. We could spend hours perfectly tuning a set of 10 weapons, but a lot of that work would be rendered invalid when we add 10 new ones. In general, it’s natural for early builds to have some balance issues.
With all this said, when we notice balance issues significantly affecting our internal playtests, it does become worthwhile to do a quick smoothing-over pass. In particular, we recently identified some issues with our current build’s looting experience:
Gameplay slowdowns: Any time players got to a storage area, they’d spend an inordinate amount of time going through every container with a fine-toothed comb. This significantly slowed down gameplay
Progression pacing issues: Many players were finding some of the game’s better weapons and food within the first few minutes of their demo!
Repetition of play: Players were overly incentivized to visit the same map areas every playthrough
Lots of shelves means lots of loot to go through -- for better and for worse
Even though Dandelion Void is in an early state and we plan on significantly increasing the pool of available items, it felt worth it to do a quick balance pass on our looting experience.
It didn’t take long to find some elegant, fast-to-implement solutions! Our main interventions were:
Container differentiation: We strictly limited the quality of loot from our most common containers, and created new “premium” containers with better loot
Decayed Container Spawns: We introduced a “decayed spawn” system that adds texture through randomization of premium containers
Let’s break each of these out into a little more detail:
Container differentiation
In the game’s pre-alpha state, we have a few types of containers that are extremely common: the industrial footlocker, and the metal shelf:
Metal Shelves (left) and industrial footlockers (right) are currently some of our most common containers
Since these containers are themed as generic storage locations, they were given very expansive loot tables (High-quality items are underlined)
Shelves and footlocker possible loot BEFORE rebalance:
Weapons: hammers, metal pipes, utility knives, etc.
Food: Chip-Chow, atp bars, Chow-Wow
Beverage: Empty water bottles, filled water bottles
Cleaning: Sponges
Healing: Bandages
Junk: All junk item types (plates, vases, bolts, calculators, etc.)
In short, almost every type of item, including high-quality variants, could be found in the most common type of container. This created a homogenous looting experience where the best strategy was to go from closet to closet, combing through each shelf and footlocker until you have all of your needs met.
The natural solution to this issue was container differentiation: by adding new types of lootable containers and limiting the contents of any given one, we can add a lot more texture to the looting experience.
Using a whiteboarding program called Miro, we did an exercise where we listed each type of container (footlockers, desks, etc.) and each category of loot (weapons, food, books, etc.). We then assigned different loot categories to each container, with the quality of loot within that category corresponding to a red/yellow/green color key.
Note that this was just a brainstorming exercise -- information above will not be fully accurate to loot distribution in the final game
Our goal was to limit most containers to just 1-4 categories -- the fewer, the better. We also introduced a number of new "premium" containers like vending machines, wall first aid kits, and tool chests, each of which is limited to very specific loot categories.
With container differentiation in mind, this is what shelves and footlockers are left with:
Shelves and footlocker possible loot AFTER rebalance:
Weapons: metal pipes (footlockers only)
Cleaning: Sponges (shelves only)
"Beverages": low-quality water containers like vases, etc. (Shelves only)
Junk: Split between the two (footlockers have industrial stuff like bolts, shelves have domestic things like plates)
Note that categories like food, clothes, and healing have been completely removed from the loot tables, and the only weapon available is the weak metal pipe item. There are still some useful items to be found in generic containers for players who are just starting, but there's no longer a strategic obligation to go through every single one.
A toolchest like this one is a great place to find boxcutter knives, hammers, and other weapons
Using specialized containers for higher-quality items adds pizazz to the looting experience; when you enter a room and see a toolchest, you’ll feel a surge of excitement at the prospect of better weapons. This also allows players to loot more strategically. Hungry? Look for a vending machine!
Returning to the overall issues we listed above, container differentiation significantly helps out with our gameplay slowdowns and progression pacing issues. But because the locations of our high-quality containers are manually placed in the map, it actually increased our repetition of play issue. This is where the decayed container spawns fix comes in!
Decayed container spawns
Those following this blog will remember that Dandelion Void uses a hand-authored map with some random elements on top. Specifically, all of the walls and furniture are placed manually by level designers, while plant growth, grime, and enemy spawns use random noise algorithms.
This meant that once you knew where a given vending machine is placed in the map, you could head right there if you ever got hungry. This is especially pronounced because our current test map is much smaller than the final one will be. This is a big repetition of play issue!
Long-term we want to add more in-depth randomization systems for individual rooms in the map, but that will take a while to set up. In the meantime, we’ve added a much simpler form of map randomization: decayed spawns!
The middle 2 vending machines in this test map spawned just fine; the left and rightmost ones spawned as their tipped-over “decayed” variants
Decayed spawns is a new system that allows furniture items to be randomly replaced in the map with another specified entity. As shown above, vending machines now have a 40% chance of spawning as a “pillaged” vending machine
We’ve created a decayed variant for every "premium" container. Corpses can now spawn as scattered remains, first aid kits have a pillaged version, etc. Decayed variants either contain junk loot, or no loot at all! 
Left: a full corpse like this will typically have some useful loot on it
Right: the corpse’s decayed variant is just a pile of scattered remains – better luck next time!
Originally we just had the “decayed” variants be the same as the resource the container would drop if the player destroys. For instance, a destroyed vending machine drops wood scrap, so it’s decayed version would also be wood scrap. This makes logical sense and was less work for us to implement as developers.
But we found that there was a real emotional punch to having a specific decayed variant for each container. If you stumble across a pile of wood scrap you don’t really think twice, but if you see a cracked-open vending machine on its side you think “dang, I almost found some food!” These moments of negative surprise make it all the more exciting when you do find a useful container.
With this new system, your favorite toolchest or medical kit might not be there the next time you play! Even a very simple layer of randomness like this helps a lot with our repetition of play issue and creates fun little story moments.
Low-effort, high impact
Container differentiation and decayed container spawns are both what we call “low-effort, high impact” changes. They didn’t take very long for us to implement, but they significantly improved the gameplay of our demo! In between more labor-intensive tasks like creating new content or designing new game mechanics, we’re always on the lookout for this type of low-hanging fruit (and loot!)
That’s all for today. Everybody please take care, and have a great week!
-- Robin and the Manzanita Interactive team