Claude Shannon (1916 – 2001) was an American mathematician and electrical engineer, known today for his contributions to communications and information theory and computer science. He made contributions to many fields , such as digital design , cryptography, genetics, communications systems, and of course “information theory”, a branch of math that he basically founded.
He devised the way to measure and quantify information, regardless of whether it was sound, text, images, or anything else. He was the first to use the word ”bit” in print as the unit of information and to determine the amount of information in anything. He is known today as the Father of Information Theory.
But he was not all work and no play either. Indeed, it would seem the line between the two was very porous for Shannon. He was known to take flying lessons, ride the unicycle, and practice juggling. "I've always pursued my interests without much regard to financial value or value to the world" he once said. In that spirit he invented several machines, such as a gasoline powered pogo stick, a machine to solve Rubik’s cube, a computer that operated using Roman numerals, and famously, The Ultimate Machine.
And he made machines to play games. This being BGG, this geeklist is about the games and what Claude Shannon did with them.
Claude Shannon was known as a strong chess player, and he was one of the first to not only propose that a computer could play Chess (Konrad Zuse and Norbert Wiener were other pioneers), but to also provide details on how that could be done. In his paper Programming a Computer for Playing Chess published in 1950, he describes how a computer could play chess using the minimax algorithm and a predesigned evaluation function. In so doing, he laid the foundation for virtually all abstract computer game programs used today.
Shannon devised and constructed a mechanical mouse, which by means of magnets and relays embedded in a rectangular maze, would learn to navigate its way through the maze. The maze could then be changed and the mouse would "forget" the old path and "learn" the new path through the maze.
Shannon did invent one game, known today as the Shannon Switching Game, though he called it Bird Cage at the time. Bird Cage is a connection game, in which two players seek to connect their edges. In the more general Shannon Switching Game, two players ("short" and "cut") seek to either connect or disconnect two vertices in a graph.
It was later shown to be equivalent to the game of BRIDG-IT, which was invented independently by another mathematician, David Gale. BRIDG-IT was published commercially by Hasbro in 1960.
Shannon devised a machine to play the game which used measurements of electrical voltages to determine where to place the next connecting segment. His machine now rests at the MIT museum. From the way it has been labeled as an electronics education aid, I don't think they know what it is.
Hex was invented by mathematicians (Piet Hein and John Nash), and has been assiduously analyzed by mathematicians ever since. It would be unusual if it had escaped notice by Mr. Shannon. He took it on like he did with the game of BRIDG-IT, by teaming up with a colleague (E.F. Moore) and building a machine which measured electrical voltages to determine the best move. He reported its performance in his article "Computers and Automata":
"With first move, the machine won about seventy per cent of its games against human opponents. It frequently surprised its designers by choosing odd-looking moves which, on analysis, proved sound".
A solution to the game of Nim has been known for a long time, even in Shannon's time. This would not stop him from building a machine to play the game, but he did allow the humans a chance to win against the machine. The machine would start the game in a position that would allow the human to win, but only if the human played the game perfectly. If the human makes a single error, the machine would seize the initiative and win.
His machine was called Nimwit and is now at the MIT museum.
In his article "Game Playing Machines", he reports building a Tic Tac Toe machine, based on an earlier design by W. Keister. The machine included an anti-cheat feature: a "tilt" light which would come on if its opponent tried to move twice.
Claude and Mathematician Edward O. Thorp, together built an analog device to improve their betting odds on Roulette wheels. The device was small enough to be concealed on the gambler's person and counts as one of the first ever wearable computers. The user would set switches in his or her shoes to denote the speed of the Roulette wheel, and the device would then communicate back with tones to an earpiece as to which section of the roulette wheel to place a bet.
Seems the device wasn't very reliable, and hence wasn't used very much. Its' existence was kept a secret until revealed by Thorp in his book " Beat The Dealer" in 1966.
Shannon teamed up with with mathematician Edward Thorp and physicist John Kelly, and they developed a method of card counting to beat the odds in Blackjack, which they put to good use in Las Vegas casinos. They gambled using a formula dubbed the Kelly Criterion, which determines the amount of a series of bets given the payoff and win/loss probabilities such that the long term payoff is maximized. Years later, their methods were adopted by the infamous MIT Blackjack Team, which was the subject of the film 21.
And oh, by the way, the same Kelly Criterion was successfully applied in financial investments by Elwyn Berlekamp, who was Kelly's assistant at the time. BGGers may know Elwyn Berlekamp from his books about games, "Winning Ways for your Mathematical Plays", "Mathematical Go", and "The Dots-and-Boxes Game".
The Penny Matching Game, similiar to the ancient game of Morra, consists of two players guessing their opponent's chosen penny side of either heads or tails. One player scores if the sides are the same, the other scores if they are different.
Shannon constructed a machine to play the Penny Matching Game against a human opponent. His machine was a variation on another machine, made by one D. W. Halbarger which would attempt to learn the pattern of head/tail choices made by its' opponent and play accordingly.
As it turns out, humans are not very good at choosing a purely random sequence since the machines would win about 55% of the time by making a good enough guess about what its opponent was going to do next.
Shannon and Halbarger tried to analyze which of the two strategies embedded in their machines would prevail against the other, but unwilling to work through the complex statistics mathematically, they decided to build a third machine to interface the two game machines and have them play against each other. This may be the first ever recorded game competition between game playing computers, thus predating the Computer Olympiads by decades.
Shannon must be one of the very few people ever to have published a mathematical treatment of juggling in a peer reviewed journal. His analysis resulted in an equation that related the time a hand is empty, the time a ball is in the air, and the number of hands and balls needed to sustain a juggling act.
To illustrate his ideas, he built juggling machines in the images of clowns and one in the image of W. C. Fields. Only the latter actually did juggling though, in the form of bouncing three balls on a drum and maintaining the strict timing needed to keep the balls bouncing and catching them on the way back.