The situation you described could be handled by the following: Imagine your board being a 10x10 Array of some Integer-variables. The single cells have different states for ‘water’, ‘ship1’, ‘ship2’, etc. So the top row could look something like this: ‘WWW1WW2WWW’. Now you would have generated a table of random numbers. You query the table like this: “The X coordinates is ___, the Y coordinate is ___, the content of this cell is ___” and the table gives you a random number it stores for this. You perform all queries for your current board and then you X-OR the results to get a new number, your hash value.

I hope I could clear things up a bit.

On the other hand, I think in the example of the game ‘battleship’ you might be better off with just analyzing probabilities. Imagine the following:

At the beginning of the game you know nothing about the state of the opponents board. So every choice of your move is equally as likely to yield a hit. This fact basically holds throughout the game with a few expections:

1.) You already hit a ship, but did not sink it. Then you know about the ship locations.
2.) You already covered most of the area with attacks, so that there are some free spots where your opponents ship cannot fit in anymore.

So, my guess would be, if you think about probabilities of a hit that a certain move has, and the way it changes probabilities in the upcoming round, you should get quite good results — After all you need luck in battleship, even if your strategy is good :)

Oh, and webOS? Awesome! The app store there really needs some decent additions!