Homework 5 - Peace War Tournament

The Prisoner's Dilemma problem is a classic problem in game theory that attempts to explain agreement and dissension among two parties. Here is a description of the game taken from Wikipedia:

Two men are arrested, but the police do not possess enough information for a conviction. Following the separation of the two men, the police offer both a similar deal- if one testifies against his partner (defects / betrays), and the other remains silent (cooperates / assists), the betrayer goes free and the cooperator receives the full one-year sentence. If both remain silent, both are sentenced to only one month in jail for a minor charge. If each 'rats out' the other, each receives a three-month sentence. Each prisoner must choose to either betray or remain silent; the decision of each is kept quiet. What should they do?

If the two convicts are cooperating, then the clear choice is to both remain silent. That way, each serves only the minimum one month sentence instead of one person receiving a full year or both receiving three months. However, if both convicts distrust the other, then it turns out that the optimal decision to be made by both is to rat the other convict out. To see this, let's examine the payoff matrix that describes the time each prisoner gets based on the decisions they make.

  A is silent A rats out B
B is silent A and B get 1 month A goes free; B gets 1 year
B rats out A A gets 1 year; B goes free A and B get 3 months

From this we can see that, for example, if prisoner A decides to be silent, then he will either 1 month or 1 year of team. On the other hand, if prisoner A rats out B, then he'll get either serve no time or 3 months. Clearly if A does not trust B, then ratting him out is the best idea.

The traditional prisoner's dilemma consists of only a single opportunity for the prisoners to be silent or rat out their partner. Academics have also considered the iterated prisoner's dilemma in which the prisoners make several such decisions in a row, potentially changing their strategies based on previous rounds. Of course, this is slightly unrealistic, so instead we phrase the iterated prisoner's dilemma game in terms of another game called A Peace War Game

Peace War Games

The peace and war game is similar to the prisoner's dilemma except that instead of prisoners we have neighboring countries. And instead of either staying silent or ratting, each year the countries declare peace or war against the opposing country, scoring a number of points based on who declares peace or war. The game continues for a number of years and the winner is the country with the most points.

The pay-off matrix for the peace and war game is as follows:

  A declares peace A declares war
B declares peace A and B get 3 points A gets 5 points; B gets no points
B declares war A gets no points; B gets 5 points A and B get 1 point
Our goal is to write a program, PeaceWarGame that plays a series of peace and war games against an AI opponent. Each game last 10 turns/years which should be represented in your program by a class constants (similar to homework 2). After the game, the program prompts the user if they would like to play again. When the user indicates they no longer want to play, the program prints out the total number of games played, the number of games the player won, the resulting win percentage, and then exits.

Sample Output

Here is a sample execution of the peace war game program in action. Like previous assignments, you should reproduce this output exactly. Because of the length of the sample output (two games at 10 turns a piece), I've included it as a text file linked below.

hw5-sample-output.txt

The following sections outline the specific format for the output of each part of the program.

Format of output

Years. Each round of the peace war game consists of a single year with a particular format of output. The general format is given below: 

=====
Year <# of year>
Our score: <player's score>
Their score: <AI's score<

What is your strategy this year? <Input echoed by Scanner>
You chose <Player's strategy>
They chose <AI's strategy>
<Peace war message>

The peace war message is based on the strategies the player and AI chooses as follows:

  Player declares peace Player declares war
AI declares peace >>> Peace for everyone! >>> We crushed them!
AI declares war >>> They crushed us! >>> Everyone to arms!

Keep in mind that the ">>>" is part of the output so that you can visually tell the outcome of the year easily.

Final report. At the end of a game, a final report is given similar to the format of the output for each year: 

=====
Final
Our score: <player's score>
Their score: <AI's score<

<winner message>

The winner message is based on who won the game and has the following format:

Like above, the > and < are part of the output.

Summary. Finally, the program reports statistics about the games played before exiting. The format of this final part is: 

Total games played: <total games played>
Total games won: <total games won>
Win percentage: <win percentage>
The win percentage should be given as a two place decimal of the form X.XX. You should use printf for this task.

Repeated and generous prompting

Unlike the previous homework, with while loops, we can now enforce that the user enters the correct input and repeatedly prompt them until they get it right. There are two places where the program prompts the user for input. Here are examples of repeated prompting in action for each of them: 

What is your strategy this year? cat
Invalid strategy.  Enter "peace" or "war": 5
Invalid strategy.  Enter "peace" or "war": peace
...
Play again? donkey
Invalid response.  Please enter "y" or "n". y

The valid inputs for the first prompt are peace or war , and the valid inputs for the second prompt are y or n.

You should duplicate these error-and-reprompt messages exactly in your program. I highly recommend reviewing chapter 5.4 of the text to review how to handle user error in a robust way.

Furthermore, your handling of user input should be generous and allow the user to enter text of any case. For example, the following exchange should be valid: 

What is your strategy this year? PeAcE
You chose peace.
They chose peace.
>>> Peace for everyone!

Note how even though the inputted text is PeAcE, we still print out peace in lowercase letters. You will find the toLowerCase and equalsIgnoreCase methods of the String class helpful to emulate this behavior.

The peace war AI

Your program will need play rounds of peace and war against the player. For the initial version of your program, your program should randomly choose between peace and war each year. The random class should be helpful in implementing this behavior.

For the purposes of the tournament described below, we would like you to implement this behavior in a specific way. Create a class named AI in a file called AI.java. Your class should take on the following form: 

public class AI {
    // Returns the strategy the computer should do this turn, either "war" or
    // "peace".  Takes a Random object and the last play made by the player in
    // the previous year, a number that is the current year being played, and a
    // String that represents the last strategy chosen by the player (not the
    // AI).  This will also be "peace" or "war".
    public static String getStrategy(Random rand, int currentYear, String lastStrat) {
        // Implement me!
    }

    // Returns your AI's name that will be used during the tournament.  This
    // can be your name or any nickname as long as it is not offensive or
    // otherwise inappropriate.
    public static String getName() {
        // Implement me!
    }
}

You must use this getStrategy method in your PeaceWarGame game when the AI needs to make a move. Remember that this is a public static method in another class, so you should use dot notation to call the getStrategy within the AI class that you make.

The parameters of getStrategy allow your AI to make more complex decisions. currentYear is current year that is being played (starting at 1). lastStrat is the strategy chosen by the player in the previous year. For the first year, you should pass "peace" in for lastStrat, i.e., assume that the playr was peaceful before the game began. For your initial implementation, the AI will choose strategies at random and thus ignore these parameters, but even if you end up using this initial implementation in your final submission, we will check to make sure that you are passing the correct values for these parameters in your program.

The Tournament

We will be holding a class-wide Peace War Tournament using the AI classes that you make. The tournament will begin with a round robin group stage of approximately 12 people per group. The top two AIs (by point totals) from each group will advance to the finals where they will play to determine the winner. The format of the finals will be a second round robin consisting of the advancing players from each bracket. We will cast the finals live over the Internet later in the semester. Note that this is different from traditional peace war/prisoner's dilemma tournaments in that those are usually purely round robin tournaments.

Here are the rules:

Once you are done with your initial program, you may implement any algorithm you wish in the getStrategy method to decide if the AI declare peace or war each year (subject to the "no infinite loops" and 'return only "peace" or "war"' constraints). In particular you may use any programming APIs or constructs that you wish, even those not yet discussed in class. For inspiration, read about strategies for prisoner's dilemma from wikipedia and feel free to search around for ideas.

Extra credit: you will receive 1 point of extra credit if you

Note that we will only award extra credit to students that are able to get their main program (mostly) working. So please focus on the the main problem before getting into the extra credit. But otherwise, feel free to have fun with it!

For your final submission, you only need to turn in your final AI. In particular, you do not need to submit a random AI if you decide to implement something fancier.

Design

Like the previous assignment, this is a more traditional programming task than the first three homeworks. There are no visual patterns for you to decompose, but the principles of decomposition still apply! There are at least two static methods we'd like you to include:

  1. A method to play a single game of peace and war. Recall that a single game of peace and war for our purposes is 10 rounds/years of play. In particular, it should not be responsible for prompting the user to play another game.
  2. A method to print out the final statistics for all the games played. This method should do not other work other than printing these statistics to the screen.

In addition to these methods, main, and the required methods in your AI class, please include three other useful static methods in your program. You are of course welcome to include more methods if it makes sense to do so (and given the number of checks and prompts you need to do, it seems very reasonable).

Also, make sure that you use a class constant to represent the number of years/rounds each game should last. Make sure that if you change this value (e.g., if you want to test your AI for more rounds in preparation for the tournament) that you change it back to 10 for your final submission.

Submission

Please submit your Java source files, PeaceWarGame.java and AI.java, electronically via the course website. You will need to place these files into a zip archive called hw05.zipand submit that archive. On Windows you can create a new "Compressed (zipped) Folder" by right-clicking your desktop and selecting "New". On OSX, you can select your files in Finder, right-click and select "Compress".