The design of Arena Guardians
This game's design was inspired by Teamfight Tactics (TFT), an auto-battler game. I first analyzed why this type of game is popular:
A highly core mechanism is the merge system: three one-star units merge into one two-star, and three two-star units merge into one three-star. This is similar to the simple mobile game, 2048. Players easily gain a sense of satisfaction and constantly feel like they are just about to achieve the next goal.
Low mechanical skill requirement, but deep decision-making: Players must decide whether to save gold for interest or refresh the shop. Will the next refresh give me what I want? Should I pivot my composition? The constant shop refreshing also plays on the psychological element of a "gacha game."
Low failure cost: The game is not a one-and-done battle; players can quickly recover their mood after a setback.
The presence of a Synergy/Trait system: This adds complexity and richness to the game, making it random yet manageable.
Our game was created for a one-week game jam with the theme "Risk for the Biscuit." I quickly decided on the auto-battler format, but due to time constraints, I had to simplify the synergy system. Also, since character design and animation are time-consuming, we only had enough time to design three types of heroes and one type of enemy. Therefore, the traditional auto-battler star-up system didn't make sense.
I decided to keep the three-into-one merge mechanism but allowed for an infinite increase in hero level. Enemy levels and stats would scale correspondingly with the player's hero level. I also simplified the shop function: all three heroes are available for purchase every round. However, this greatly reduced the heroes' scarcity, so I decided that dead heroes would not revive at the end of the round, forcing players to adjust their strategy to ensure hero survival.
After playtesting, we found that at higher levels, players had to spend a lot of time merging from level one. The system was improved so that the base level of heroes in the shop increases by 1 every 7 rounds, with a corresponding increase in gold cost, eliminating the need to start merging from the bottom every time.
The "risk" theme was originally designed to include many elements:
Terrain: Some tiles on the board would be unknown terrain spaces, for example, representing a tree. A fragile archer could be placed on the "tree" tile, which would lower the enemy's attack priority and increase the archer's accuracy. However, there would be a probability each round that the archer would "fall from the tree" and die instantly.
Potions: Potions might greatly increase attack power, but with a chance of causing "mental derangement," leading the hero to also damage their own units.
Due to the limited development time, it was difficult to implement all of this in seven days. Drawing inspiration from the feeling of going "all-in" with gold in TFT, we implemented a mechanic where players could clear all, 60%, or 30% of their gold in exchange for a corresponding level of power boost. However, this boost had a chance to apply to the enemies instead of the player. The more gold the player risked, the more dramatic the boost, and the higher the chance it would empower the enemy.
After playtesting, we found that if the probability remained constant, a single major failure resulted in too strong a feeling of defeat. The system was adjusted so that the probability now changes based on the previous result: if the boost went to the enemy in the last attempt, the current attempt has an increased chance of benefiting the player. This probability is also displayed on the "Risk" button for the player's reference.
There are still some shortcomings, such as having only three heroes, the infinite leveling system becoming tedious at high levels, and the risk results being too random, which can make the experience feel a bit monotonous. However, this one-week game jam was my first time leading a team as a designer, explaining the design rationale, and communicating implementation with the programmers and artists. The entire team consisted of game developers I had met on Discord but never in person, and I consider it a relatively successful experience.
