IBM is at it again, but this time their target for domination isn’t aimed at nerdy trivia buffs or intimidating chess masters. No, its electronic jaws are now chomping at the bit to take a megabyte out of sweet innocent little children.
The big story in science and technology this past week was the defeat of Jeopardy! champions Ken Jennings and Brad Rutter by a supercomputer named Watson. Developed by IBM engineers over a four-year period, Watson handily (with no hands, natch!) trounced its human competitors by quickly using language algorithms to analyze and interpret several terabytes of data to find the right answer. Or, in this case, find the right question.
The win is a triumph for the advancement of artificial intelligence (A.I.), following in the footsteps of its chess playing ancestor (or is that transistor), Deep Blue. However, these man vs. machine massacres are nothing compared to what IBM plans on unleashing next: a computer that can mercilessly crush a child’s self-esteem at the simple racing board game Candy Land.
First published in 1949, the game has become an American icon and is a ubiquitous staple in family closets everywhere. Requiring almost no reading skills, and minimal counting, Candy Land has brought immense joy, as well as a sense of friendly competitive spirit, to generations of children who love its simple storyline of Candy Cane Forests, Molasses Swamps, and Gum Drop Mountains.
Sadly, the days of innocently navigating the colorful sugary roads to get to the Candy Castle and find the lost King Kandy may be gone forever.
Shrouded in secrecy for the last several months, IBM is about to announce the unveiling of Ellie, the first computerized Candy Land champion. Named after the original designer, Eleanor Abbott, who conceived and designed Candy Land while recovering from polio in 1945, Ellie is poised to dominate the magical world of Gramma Nutt, Mr. Mint, Queen Frostine, Lord Licorice, Gloppy the Chocolate Monster, and the Gingerbread People.
Using a complex mathematical algorithm, Ellie factors in the number of players, including herself, number of actual cards (single, double, picture/character) in the 64 color card deck, as well as the 134 spaces on the board, then calculates the odds of landing on a shortcut, penalty space or, worse, landing on Gooey Gumdrops, getting lost in the Lollipop Woods, or stuck in the Molasses Swamp. She then uses advanced statistical analysis to determine the probabilities, averaging the length of play, and develops her game strategy accordingly.
The program also factors in whether the players decide that a win is determined by advancing past the final purple square or having to land exactly on it, a source of endless squabbling and tantrum throwing. Also, as Ellie’s ‘technically’ the youngest player, she always gets to go first, which gives her a slight advantage.
“The programming is quite sophisticated,” stated the project’s principal investigator Milton Bradley (no relation to the game company). “It requires mathematical skills akin to counting cards in gambling, the reasoning skills of a kindergartner, and a childlike burning desire for cheating. All of that takes enormous processing power,” he continued.
Ellie will be presented to the public for the first time in April at the 2011 World Candy Land Championship Tournament in San Diego, California, where she will compete with the top winners.
Future plans for Ellie include software upgrades for the games Chutes and Ladders and Sorry!, the latter of which is especially important, so that the children can get into the habit of apologizing to their computer overlords.