El Ajadrecista. The mechanical chess player by Leonardo Torres Y Quevedo
Some inventors have become famous for a unique, original creation. Others have been remarkably prolific. The Spanish engineer Leonardo Torres Y Quevedo belongs to the latter category. In his 84 years of life, he invented the telekino, history’s first radio control, several airships, countless calculation machines and quite a few cable cars. Today his Aerocar, designed in 1916, still galvanises hundreds of tourists, making them fly over the Niagara Falls. Every device he invented had to be the best in their class because Leonardo Torres Y Quevedo was persuaded that this was the only way for machines to overcome their limits.
At that time, automatic devices had proven they could take over repetitive tasks in workshops, factories, and countless everyday activities. But could a machine make decisions? The Spanish engineer thought it was time to answer this question. In 1912, Torres embarked on a highly ambitious project: to create a device capable of playing chess. This game had always been synonymous with intelligence due to the calculating skills required and the need for the players to adapt their plans according to the opponent’s moves. Playing chess, for a machine, meant interacting on an equal footing with a human being.
The Hungarian civil servant Wolfgang Von Kempelen had already acknowledged the potential of such an instrument two centuries earlier. He had a passion for mechanics and built The Turk: an automaton with human features, dressed in a flashy turban and baggy oriental trousers. The device astonished Europe, beating Napoleon Bonaparte, Benjamin Franklin and many professional chess players. Purchased by another inventor, Johann Nepomuk Maelzel, The Turk landed in America, where it fuelled a heated debate about whether men could one day be replaced by a thinking machine. But some had doubts. Edgar Allan Poe dedicated an essay to the automaton, setting out his reasons for believing it to be a hoax. “If it were a real machine”, Poe argued, “then it would be the most extraordinary machine ever made by man”. It took several years for The Mechanical Turk to be finally recognised for what it was: a fraud. An expert chess player hid inside the device.
Torres’ machine, however, left no room for doubt. It had no human features but was an agglomeration of cables, switches, relays and mechanisms between which a mini chessboard appeared. There was no place to hide anything. Moreover, his chess skills were minimal. In fact, he couldn’t even play a whole game.
El Ajadrecista (the chess player), as Torres named his creature, was presented at the 1914 World Exhibition in Paris and played only an endgame of king and rook (moved by the machine) against a king (moved by a human being). Basically, the human being had no chance of winning. Nevertheless, the device never made the best move or got to checkmate within the 50 moves beyond which, according to chess rules, the game was a draw. But if the opponent was forbearing, El Ajadrecista always won.
Such a device appears insignificant by modern artificial intelligence standards. Still, for the first time in history, a mechanical device could adapt its responses to the interaction with a human being. When the opponent made a forbidden move, such as moving the king two squares instead of one, El Ajadrecista would express its resentments by lighting up a huge light bulb. When it checked the opponent’s king, he would emphasise this by lighting another bulb. Several light bulbs at the same time meant checkmate.
The second version of the Ajadrecista, built by Torres in 1920, could even pronounce “check” and “checkmate” through a phonograph. Instead of the mechanical arm previously used, the pieces were now moved by magnets. This way, the machine seemed to be animated by magical energy.
In 1951, Norbert Wiener, the American engineer considered to be the father of “cybernetics”, publicly acknowledged before the scientific community Torres’s contribution to studies on artificial intelligence and chess was for a long time one of the main instruments of investigation in this field. In the same year, Alan Turing developed a chess program, which remained only on paper, capable of playing an entire game. In 1996, the Deep Blue computer was the first to beat a world champion, Garri Kasparov. Today, we can install software on our smartphones that can win a chess game against any human being.
To know more:
- “Torres and his remarkable automatic devices”. Scientific American Supplement No 2079 (6 November 1915).
- Andrew Williams. History of Digital Games: Developments in Art, Design and Interaction. CRC Press, 2017.
- Randell, Brian. “From the Analytical Engine to the Digital Electronic Computer: The Contributions of Ludgate, Torres, and Bush.” IEEE Annals of the History of Computing 4, no. 4 (October 1982): 327–41.
- JJ Velasco. “Historia de la tecnología: El ajedrecista, el abuelo de Deep Blue”. Hipertextual, 22 July 2011.
- Ramón Jiménez. “The automaton of the Torres y Quevedo tower endgame”. Chess News, 20 July 2004.