11 -- Ultimate Pig -- Ryan Burton

[rryanburton]

Ultimate Pig

Description

Use Monte Carlo simulations to find the ultimate player of Pig Solitaire.

Tasks

• [x] Blank slate
  • [x] Create a GitHub repo called ultimate-pig
  • [x] Put README.md and requirements.txt files in it
  • [x] Make sure ipython[all] and any necessary plotting libraries are in requirements.txt
  • [x] Make sure your ultimate-pig folder has its own virtual environment (.envrc), but do not commit .envrc or .direnv
  • [x] Install the requirements using pip install -r requirements.txt
• [x] Normal mode
  • [x] Create an IPython Notebook file for simulating the game
  • [x] Create a class for keeping track of the game state
  • [x] Create a Player base class for a Pig player
  • [x] Create at least two subclasses of Player that use two different strategies for playing the game
  • [x] Run a large number of trials of the game with each subclass of Player
  • [x] Plot the results of your players
  • [x] Attempt to improve upon your best strategy, and use charts and data to determine if your new player is better
• [ ] Hard Mode
  • [ ] Create a simulation of a 2-player game of Pig
  • [ ] Create a new Player class for the 2-player version, and try pitting it against your optimal Pig Solitaire player
  • [ ] Create charts and data analyzing the 2 players' performance over a large number of trials

    Objectives

    Learning Objectives

After completing this assignment, you should understand:

• Modeling games and simulations

  Performance Objectives

After completing this assignment, you should be able to:

• Play Pig well
• Use subclasses
• Create Monte Carlo simulations

  Details

  Deliverables

• A Git repo called some-pig containing at least:
  • an IPython notebook
  • a requirements.txt file

    Requirements

• At least 5 different charts

  The Rules of Pig Solitaire

In 7 turns, you are attempting to get the highest score possible.

Each turn, you have two choices:

• Roll. Roll a six-sided die. If it comes up one, your turn is over and you add nothing to your score. If it comes up two through six, you add that number to your turn total and choose again.
• Hold. If you hold, you add the turn total to your score.

An example:

The first turn, I roll a 1 immediately. I receive no points.

The second turn, I roll a 5. I choose to roll again. I roll a 3, for a turn total of 8. I choose to roll again. I roll a 3, for a turn total of 11. I hold, receiving 11 points. My score is 11.

The third turn, I roll a 5 again. I choose to roll again. I roll a 6, for a turn total of 11. I choose to roll again. I roll a 4, for a turn total of 15. I decide to roll once more. I roll a 6, for a turn total of 21. I hold, receiving 21 points. My score is 32.

The fourth turn, I roll a 6. I choose to roll again. I roll a 3, for a turn total of 9. I choose to roll again. I roll a 3, for a turn total of 12. I choose to roll again. I roll a 4, for a turn total of 16. I hold, receiving 16 points. My score is 48.

The fifth turn, I roll a 6. I choose to roll again. I roll a 6, for a turn total of 12. I choose to roll again. I roll a 3, for a turn total of 15. I choose to roll again. I roll a 4, for a turn total of 19. I choose to roll again. I roll a 6, for a turn total of 25. I hold, receiving 25 points. My score is 73.

The sixth turn, I roll a 6. I choose to roll again. I roll a 1, and receive no points.

The seventh turn, I roll a 5. I choose to roll again. I roll a 1, and receive no points.

My final score is 73.

Normal Mode

Create a simulation of different types of players of Pig Solitaire. You should have a base Player class that only rolls once per turn. Then create Player subclasses -- at least 2 -- that have different strategies. Create charts showing how they do in a large number of trials. Attempt to create an optimal player and back up your assertion with charts and data.

Your classes, findings, and charts should be in an IPython notebook.

Hard Mode

In addition to the requirements from Normal Mode:

Once you have your optimal Pig Solitaire player, create a second simulation. This could be in a separate notebook or the same one. This simulation will be of the game of Pig. Pig works like Pig Solitaire, but you play against another person, and are trying to be the first to score 100 points. There is no round limit.

Create a new type of player for Pig. Your new player should play against the strategy as your optimal Pig Solitaire player. (You will probably have a new class for your optimal Pig Solitaire player, as playing Pig will need the class to take in more data each turn.) Try to create an optimal competitive Pig player, when the opponent is your optimal Pig Solitaire player. Back up your assertions with charts and data.

[rryanburton]

https://github.com/rryanburton/ultimate-pig feeling pretty good about this one after all. @powder-river @jamesmallen