Game Design: Simple but Smart Automata
In the co-op mode “It Learns” – I wanted an difficult automated opponent that adapts to what the players (challengers) do. So, I wanted a smart Automata, but at the same time I wanted an Automata that is simple to play, one requiring not much effort from the players.
These are in sense conflicting goals … if you make the Automata smarter that basically means bigger decision tree, which means more difficult and time consuming to play. After all this is not a video game and a person will need to perform the actions of the Automata.
The Ark is invaded by a self-aware hybrid organic-computer virus. Players will try to purge the virus out of the Ark’s systems. But, it learns! More actions you take against the virus, more dangerous it will become.
The solution was to have a simple decision tree with 3 possible actions that the Automata can take.
Instead of changing the decision tree based on player’s action, the Automata will upgrade the actions on the decision tree. The Automata will still perform the basic 3 actions, but will perform more powerfully the action that is worse for the challengers.
The virus has a player board and 6 “Enemy Action” cards.
“Enemy action” cards are introduced to randomize actions for the Automata.
Each turn, the player taking actions for the Automata will need to flip a card, see how much that action is upgraded on the player board and apply the appropriate action. Simple.
Enter the Challenger Action cards.
There are 8 Challenger Action cards. Each Challenger Action card has 2 possible actions to choose from. Next to each action there is a small sign depicting what Automata action is a best contra-measure.
Each turn, a player/challenger chooses an action and rotates it one way or the other taking the appropriate action. By rotating, the small symbol on top is active.
If there are no cards to rotate or none that depict an action that the player wants to take, all the Challenger cards are reset. But, before the reset, the small symbols on top are counted – the most frequent one decides what next will Automata upgrade on it’s player board.
So, by upgrading the best contra-measure, the Automata adapts to the challengers actions. The Automata still performs the basic actions with similar decision trees, but performs better the actions that are most dangerous to the challengers.