The History of Computer Chess - Part 4 - Alex Bernstein continued...
In time, Bernstein's chess computer programme on the IBM 704 proved too large for the machine's memory capability; he had just four thousand words of memory to work with, and although in this time the computer's memory theoretically doubled, Bernstein still had two hundred words of overflowing memory. Bernstein used his considerable chess knowledge and six months of study of Modern Chess Openings to re-write the programme to concentrate on a tree of legal possible moves that were likely to prove fruitful, and eliminate the unnecessary clutter of evaluating all legal possible moves in any given position. This strategy didn't actually increase the computer's playing power; in fact, it inevitably led to disappointing omissions of the unusual or brilliant moves and plans. Bernstein was forced to economise the legal possible moves in order to fit the memory capability of the 704 and maximise its overall working power, but this unwillingly led to the suppression of some of its specific chess playing power.
Eventually, Bernstein's programme was recognised by IBM for its potential, and he and his team were moved to IBM headquarters on the corner of East 56th Street and Madison Avenue in New York City. There, they continued to work on the programme during their standard contracted third shift. Working full time with his assistants, Bernstein's team shared their machine with another research team who were designed FORTRAN, a new computer language that aimed to revolutionize computer science languages to make them more accessible and natural to humans. This easy-to-code language would then be directed translated by machines themselves into their relevant code. Because both teams were sharing the machine, they could only work in shifts over the night, alternating with one group on and the other off. They do say that New York City is the city that never sleeps!
After two years, all the code was finalised and it was time to try out the machine. Bernstein opted, quite sensibly, to do so in the absence of an audience, in the privacy of the lab at IBM. Initial attempts were not promising; "There was a bug. The very first move the machine ever made was to resign!", Bernstein would later tell a colleague. But like many computer programmes, even modern ones in 2019, bugs were continued to be found by the Bernstein team and other outside researchers for some years afterwards. The Bernstein programme was ready as it would ever be for release. "It played, I think, a sort of respectable beginner's game," Bernstein says, "and every once in a while it made a move which was remarkably good.". It was time to release it to the world!
The programme garnered some unexpected global interest. Besides all the expected scientific conferences Bernstein was invited to address, the national media and international journals took note. His article with Michael de V. Roberts was published in Scientific American in June of 1958; entitled 'Computer V Chess Player', it explored, 'Can a machine be made to think creatively? One reply is: First, can a machine can be made to play a good game of chess? How a computer was programmed so that it could defeat an inexperienced human opponent'. Interested readers might take note that this article appears in the Scientific American digital archive, and can be accessed online!
Bernstein was also featured in a short film by the Educational Testing Services; you can see him playing a game with the programme on the 704 and a chess board here: https://www.computerhistory.org/chess/mov-433190c26dd21/
Then came the national media. After an article in the New York Times, Life magazine requested an article, interview and accompanying photographs. After photographing Bernstein himself with a computer, the editor asked if he could find a chess player to pose with a board. Bobby Fischer, then a young 14 year old teenage prodigy, US Champion and resident of New York, was requested. His fee would be $2500, a not-so-inconsiderable sum that would have roughly the purchasing power of $25000 today!
Life magazine declined this fee, and asked Bernstein if he knew anyone else suitable. What they didn't know was that Bernstein did know another famed chess player, one who served as one of the first human test subjects of his chess programme, and who Bernstein kept in regular telephone contact with. It was none other than Dr Edward Lasker, a German-born, American-resident International Master, 7 time US Open Chess Champion, friend and distant relative of World Champion Emanuel Lasker. Dr Lasker, described by Bernstein as "a charming man and a gentleman", commanded the princely sum of $1 for the photoshoot.
A chess set was sought by the magazine, and they rented a board and set from an antique dealer on nearby Madison Avenue. Made of delicate filigree ivory, the board and set hailed from sixteenth century India, with the pieces alone costing $2500. The board commanded a further $1200, and despite being insured for the shoot, everyone present was said to be very scared in case on the pieces fell over!
The shoot went off without a hitch, but the development of the photos did not. The film did not turn out, and so the board, set, Alex Bernstein and Dr Lasker were re-rented for a second shoot, at a considerable cost, one that Bernstein would later calculate cost more than if they had paid Fischer's appearance fee in the first place! After this delay and disaster with the photos, the piece was eventually quietly dropped from Life magazine, though the photographer would later sell the photos to be published in the early 1960s in a series of books on mathematics by Time-Life, a book sale club who mailed monthly instalments of factual books to households in America.
But this saga did not mark the end of IBM's involvement in chess, as we shall see in later installments of my blog charting the course of chess and computers in history! In fact, some 40 years later they would receive publicity far beyond an article in Life magazine...
