So again what is going to be your major work in this class is going to build a hex program. The first version of your program is going to be a simple version. In that version you're not going to do much in the way of what I'm calling AI artificial intelligence. There's going to be a detailed explanation on the course site of the homework, but the first version of your program should basically be able to make a legal and reasonable move. By a legal and reasonable move, think about the moves you tried to do when you were playing tic-tac-toe. What were reasonable moves in tic-tac-toe? Tic-tac-toe as a reasonable move is you tried to connect whatever your long set of xes were to make them longer. If you had a single x, you tried to make another x near it so that you could ultimately have three in a row. Or if you were in five in a row, if you had a three in a row, you wanted to try and make it a four in a row. So in tic tac toe what gave you a reasonable move was extend your longest paths and block your opponent's longest paths. Those are sort of what any kid would sort of come up with after playing the game for awhile or reading about the game. So that's what I want you to do. The analogous thing in Hex is try and create longer paths from your borders or block longer paths for your opponent's borders. Now, later, we're going to build a very smart Hex player. And the way we're going to build a very smart Hex player so, we're going to use a Monte Carlo strategy. Again, Monte Carlo is the very famous gambling resort. James Bond is always playing in the Monte Carlo Casino, of course, he's game of choice is baccarat. Our game of choice is going to be a dice game. And we're going to take probabilities by using a pseudo random number generator and then in this, in this way we're going to choose our moves at random. And we're going to use randomness in simulation to try and build a smart player. It's sort of a contradiction. You can use dice thrown at random, find moves at random, and yet they're going to lead to us building a very smart player. That by itself, is a very interesting result in the artificial intelligence community. So Monte Carlo programs are probabilistic simulations. And when you do them, here's a short quiz, what is your standard pseudo random number function in the C and C++ community? And in the C and ++ community, what does srand do for you and how and why should it be used? Take a second. So the standard function, we should have used it already, is int rand void. It's in the C standard library. It returns a pseudo random number generator, a positive number between zero and the predefined constant randmax. Randmax is important because we can use it if we do double arithmetic to get us a probability. If we divide the random number by randmax using double, we'll get a probability between zero and one. S ran time null is a way to seed the random number generator so that each call to the pseudo random number generator doesn't start at the same point in the list and that's important. We don't want to continually simulate the same set of events. Every time we restart the program we want it to be using a different part of the sequence. Okay, so that's going to be what we're doing with hex. We're going to aim ultimately at a super good hex program.
