ins_apmj@jhunix.UUCP (Patrick M Juola) (03/04/86)
In article <2078@jhunix.UUCP> ins_bbdg@jhunix.UUCP (James T. Kirk) writes: >In article <2054@jhunix.UUCP> Ken Arromdee writes: >> In article <4474MIQ@PSUVMA> miq@psuvm.bitnet.UUCP writes: > >>> There is a major hitch here-- the fact that a piece of logical machinery, >>>in doing the most logical thing, is perfectly predictable. This would spell >>>doom in any battle or conflict, whether in space or on a chessboard. > >> No. There are situations where the most logical thing to do is to make a ran- >> dom decision, for precisely the reason that it's less predictable--making the >> unpredictable decision has a higher probablilty of producing the desired >> result than any predictable decision. If a machine makes a predictable >> decision in such a situation, it isn't "doing the most logical thing". >> Kenneth Arromdee >I disagree. If two giant computers were playing chess, both had equal >information, and both had equal capabilities, then both would be able to >figure out all possible moves, all possible consequecnes of all possible >moves, all possible consequenses to all possible consequences, etc. Then, >each would try to find the most advantageous moves possible. >Therefore, the one who moved first, should win, since he chooses the most >advantageous move with which to open. Similarly, if two gigantic battle >computers were working on ships, the battle would never exist, since each >computer would be working on the best time and the most advantageous plan >along with the most favorable situation, while also computing what the other >computer would be computing, trying to second-guess it, and work *that* >into the situation. If you're curious for further explination, see the Dr. >Who episode entitled "Destiny of the Daleks" with Tom Baker. >Lad, you're gonna need somethin' ta wash thaet down with. Y'ever try any >Saurian Brandy? Wrong, wrong, wrong. Chess is a poor example, since chess is what is known as a "strictly determined" game. Let's take a really simple game for a while, all right? Rules of our game : We both take a coin and (secretly set it to be either heads or tails. If we match, you win $1, if not, I win $1. What is your best strategy? Basically, just flip the darn thing, so it lands randomly. If you use a *REAL* strategy (heads only on tries divisible by three, etc), I can predict your strategy and use it against you. If, on the other hand, you're totally random, the best I will hit is 50/50, no matter what I do. See any decent games theory (or finite mathematics) text if you don't believe me. P.S. There are games that the second player can force a win, despite the first player playing perfectly. Also, for anyone who considers Dr. Who a scientific source, I have a perpetual motion machine to sell him... Pat Juola Hopkins Maths "Reverse the polarity of the neutron flow," indeed!