idsia / newtechnowar

New Techno War, an IDSIA project in collaboration with Armasuisse.

New Techno War

New Techno War is tabletop wargame developed by Helvetia Games in collaboration with Armasuisse. New Techno War makes it easy to simulate future systems integrating new technologies while stimulating discussions. The goal here is not to win, but to understand the strengths and weaknesses induced by these future systems in given tactical scenarios.

In this collaboration between IDSIA and Armasuisse, we aim to build a companion agent based on XAI that can help the players make strategical decisions and solving various combat scenario with different goals and constraints.

Structure

This repository is structured in three parts:

• code contains all the Python code necessary to play the game, develop agents or run experiments;
• frontend contains a WebApp build with ReactJS that can be used by humans to play the game in a web browser; and
• docs contains various kind of documentation.

Demo version

In order to run a demo version, which is the game with the WebApp as frontend, use the following standard Docker Compose commands:

docker-compose build
docker-compose up -d

The WebApp will be reachable at the port :8080.

More details

For more details on how to use this repository, please check the Wiki for all the details.

newtechnowar's Issues

Store games in database

For demo purposes, store the played games and recorded actions in a database. In this way, we can analyze some peculiar behavior of agents and build a record set of game that can be used in machine-learning based approach to the problem.

Functional Specifications

• Add a database (maybe using SQLAlchemy)
• The database need to keep track of:
  • parameters of game
  • if a game is completed or not
  • each action as a single transaction
• Add an agent to replay the game: this agent will play all the actions stored in the database
• (Optional): move forward and backward in time

Current Implementation

• Agents already stores an history of performed actions

Extends the Responses to follow Move path

It exists a variant of the response mechanism that allows to respond to a Move action. This is something not written in the rule book. The variant rule is the following:

When an enemy figure moves into the LOS of an ally figure, this ally figure can "respond" to the movement.
In any case, the enemy figure is stopped in the location where the response happens.

The response is intended to happen at the same time of an action. This variant allows the defending player to create ambushes and slow the advance of the enemy team.

It is still in doubt if to implement this mechanic or not.

Functional Specifications

• Allows response to Move actions.
• Allows to attack the path (hexagon) of a unit.
• When these response happens, move back the moving unit where the response happens.
• All other rules of responses apply.

Current Implementation

• Move already keep track of the path.

Implement use of smoke

In the tabletop game it is possible to use "smoke screens", a defensive support weapon that creates a temporary block to LOS on an hexagon. This is still not implemented.

Functional Specification

• The smoke launcher can target only an empty hexagon.
• From the shot moment, the hexagon will have the smoke status.
• Smoke can have different sizes: single hex or groups of three. For now, keep single hex.
• The smoke has a timeout of 3 turns.
• Each turn update will reduce the smoke by 1. Reaching 0 the smoke has no effect.
• Show in the frontend a counter on the smoke.

Current Implementation

• A smoke layer is available on the GameState object.
• An AttackGround action already exists.