Rock-Paper-Scissor is the simple game that all people can play and enjoy it

1. Game-presentation

   1. Introduction
   2. Object of the game
   3. Gameplay

2. Rules

   1. Starting-positions
   2. moves

3. Environments

   1. Episode termination
   2. Observations
   3. Actions
   4. Reward

4. Installation

5. Usage

   1. Opponent player
   2. Agents tournment
   3. Citation
   4. Contribute

https://en.wikipedia.org/wiki/rock_paper_scissor

Rock-Paper-Scissor is a game that it play with hand gesture (left or right).

gym-rpsgame is an OpenAI RL environment that we can train your agent to play this game.

In the game, player must show one of these three possible action (rock, paper, scissor) where rock beats paper, paper beats scissor and scissor beats rock.

During the game, player must be stratege by showing rock or paper or scissor and trying to beat its opponent player. If player shows rock and its opponent shows paper then player looses a game but if player shows scissor and its opponent shows paper then player win the game.

In begining, any players cannot show its hands.

Players can move following this three hands:

• Rock: the 5 fingers of its hand are closed.

• Paper: the 5 fingers of its hand are openned.

• Scissor: Just two fingers are openned.

The observation is a 1D-numpy array of size (3, ). This array is a dtype=np.float value. This observation are a different hand given by opponent player during a game, step count and score of the game. You can see image below

The action are:

- rock --> 0

- paper --> 1

- scissor --> 2

Player have three rewards:

1. If player shows an action greater than an action of opponent player then reward is 1 contrary reward is -1.

2. If player shows an action equal to an action of opponent player then reward is 0.

The episode is terminated if a step count of game down to zero agent.

• python 3.7+
• OpenAI gym
• Numpy
• Pygame
• PIL
• Keras or Tensorflow or Pytorch

Dependencies can be installed with pip or conda

$ git clone https://github.com/batalong123/gym-rpsgame.git
$ cd gym-rpsgame
$ pip install -e .

import time
import gym  
import gym_rpsgame


env = gym.make('rps-v0', opponentplayer='statistical')


for i in range(20):

	done = False
	total_reward = 0
	obs = env.reset()
	count = 0

	while not done:

		action = env.action_space.sample()
		obs, reward, done, info = env.step(action)

		total_reward += reward

		env.render()

	print("========== Rock-Paper-Scissor ===========")
	print(f'episode#: {i+1}')
	print(f'Total reward: {total_reward}.')
	print(f'Last game: {info}.\n') 

env.close()

The rock-paper-scissor environment have seven own agents to play against your agent we have:

• rock agent: this agent give everytime rock sign.

• paper agent give paper sign

• scissor agent give scissor sign

The others agent are copy_opponent, reactionary, counter_reactionay, statistical.

You can make agents tournment with rps-v1 environment like this:

import time
import gym  
import gym_rpsgame

env = gym.make('rps-v1')

game = {'win': 0, 'loose': 0, 'draw': 0}
for i in range(20):

	done = False
	total_reward = 0
	obs1, obs2 = env.reset()
	while not done:

		action1 = env.action_space.sample()# random agent1
		action2 = env.action_space.sample()# random agent2
		(obs1, obs2), reward, done, info = env.step([action1, action2])

		total_reward += reward

		if reward > 0: 
			game['win'] += 1 
		elif reward < 0: 
			game['loose'] += 1 
		else: 
			game['draw'] += 1

		env.render()
		env.game.make_gif()
		time.sleep(1.0)
	print("========== Rock-Paper-Scissor ===========")
	print(f'episode#: {i+1}')
	print(f'Total reward: {total_reward}.')
	print(f'Last game: {info}.') 
	#print(f'Last observation: {obs}.')
	print(f'Game: {game}\n')

env.close()

This environment allows us to compare different agent after trainning.

Bibtex if you want to cite this repository in your publications:

@misc{gym_rpsgame,
  author = {Massock Batalong M.B.},
  title = {rpsgame Environments for OpenAI Gym},
  year = {2023},
  publisher = {GitHub},
  journal = {GitHub repository},
  howpublished = {\url{https://github.com/batalong123/gym-rpsgame}},
}

Feel free to contribute to this project. You can fork this repository and implement whatever you want. Alternatively, open a new issue in case you need help or want to have a feature added.