Prinz's chess program

Dietrich Prinz's program on the Ferranti Mark 1 solved mate-in-two problems; no attempt was made to have it play a game.

To have something working quickly, the rules were cut down: no castling, no double pawn moves, no en passant or promotion, and stalemate counted as mate. The program did not discard unlikely moves; it tried every move until it found the solution. The simple problem in the report took 15 minutes, about 450 moves at roughly two seconds each, and Prinz wrote that a solution could take several hours. The report called the machine "disappointingly slow" at chess and saw a game against a human as far off. It is Shannon's type A strategy in its purest form. Prinz's article appeared in the journal Research in 1952: the reprint gives only the year and pages 261-266, and a bibliography of 1970 dates it to June 1952, volume 5. The bibliography of Levy's own collection (1988, p. 435) also gives volume 5, 1952, pp. 261-266, without a month. In Google Books snippets of that volume, p. 261 shows the title, the byline D. G. Prinz and the firm Ferranti Ltd. An abridged version of the article was printed in October 1952 by the MIT undergraduate magazine Tech Engineering News, and a list in Computers and Automation (May 1954) records a six-page Ferranti reprint from Research dated 1952. The printed issue of Research, with its month, was not seen. What the record does not claim: that Prinz himself dates the first run. The first two pages of his article (pp. 213-214 of Levy's reprint) were read in the free preview, and the technical description that follows in an abridged version in Tech Engineering News; neither dates the run. The abridged version does not name Research and lacks the paragraph on Cooke, and the more detailed analysis of the problem's structure that it promises at the end of the article is not in it either. The text of pp. 214-219 of Levy's collection, gathered through the Internet Archive's search, has no year, month or season, and Google Books' index has no "1951" on those pages; the pages themselves were not seen, so what was gathered may be incomplete. The only 1951 date in Levy's reprint stands on p. 213: in a report to the Manchester Guardian of 13 November 1951 Alistair Cooke mentioned an American firm, and Prinz adds that inquiries in the United States had not confirmed it. It is the date of a newspaper report, not of a run, though it shares month and year with November 1951. The original in Research was seen only in Google Books snippets: the start of the article on p. 261 and the reference list on p. 266, which again names the Manchester Guardian of 13 November 1951. Pages 262-265 of the original and the issue's month were not seen. November 1951 comes from Faster Than Thought.