Release Date: November, 1951
Platform: Ferranti Mark 1
Genre: Chess Program
Developer(s): Dietrich Prinz
Publisher(s): Ferranti
Before we get to Draughts, we're taking a little sidestep into the adjacent world of chess programs. Dietrich Prinz's Mate-in-Two, or Robot Chess, appears in a couple of online databases, so I felt it was worth brief consideration. Gaming-History creatively calls this program Chess-Playing Program. Which is... a true statement. Sort of.
What I mean is that Mate-in-Two isn't a full chess-playing computer program. You couldn't run a full, computerised game of chess on even the newest tech back in 1951 - Alan Turing's Turbochamp experiment proved that. The hulking, room-sized beasts, even of the calibre of Ferranti's Mark 1, just didn't have enough memory to manage all the thousands upon thousands of possible combinations of moves available in Chess - not to mention additional rules like castling and en passant. To make a working Chess program of any form, you'd have to focus on one phase of the game, rather than the whole. Dietrich Prinz realised this, and decided that the most manageable phase to focus on was endgame. Less pieces, less moves to contend with. Specifically, Prinz designed the program to deal with specific chess puzzle scenario where the task is to checkmate within two turns - hence "mate-in-two."
A Prinz-ly Parallel
Prinz himself has a rather interesting story, in some ways paralleling that of Josef Kates, creator of Bertie the Brain. Born on the 29th of March, 1923 in Berlin, Germany, Prinz's family would, like Josef Kates', make their way across the English Channel prior to the onset of the Second World War - Prinz's mother was Jewish. Prior to leaving Germany, he had received quite a prestigious education at the University of Berlin - he got to learn from Albert Einstein! Prinz graduated with a Ph.D. in Physics, and worked for a time with German TV and Radio manufacturer Telefunken, before deciding in 1935 that it wasn't worth hanging around Germany anymore.
 |
| Dietrich Gunther Prinz (1923 - 1989) |
Britain was a profitable safe haven for Prinz, settling quite nicely into a job for the British General Electric Company (GEC - not to be confused with the American General Electric!) However, this is a case of "second verse, same as the first." Just like Josef Kates, Prinz was interned when war hit Britain, over suspicion of any German (or Austrian) refugees being potential spies. Also like Kates, Prinz was booted out to Canada!
Unlike Kates, however, Prinz would return to England after the war's end. He started up north in Leeds, before being picked up by Ferranti in 1947 for a computer project at their Manchester site. That project would become the Ferranti Mark 1 - a landmark computer that I touched briefly on in the Nimrod article. The Mark 1 was the first commercially available digital computer and, for us gamers, is the oldest computer with extant games and software. It's only taken four years to transition from purpose-built hardware made to play a single game, to the more familiar model of a single computer able to have multiple games programmed and stored on it.
Unfortunately, Prinz's Mate-in-Two is not part of the suite of extant software - hence why this is a Gaiden article. Prinz got the idea to look into chess programming from a couple of sources - one, a research paper on the subject, and seminars he'd been attending on programming run by none other than Alan Turing himself. However, like I said at the beginning of the article, the Ferranti Mark 1 wasn't capable of running a full chess program, so Prinz focused his efforts on endgame situations, hence the commonly given "Mate-in-Two" title for the program. He began work on this program in 1949.
However, Prinz's program would not be completed before he got involved in a little side project that will be familiar to us - Nimrod. It shouldn't come as a surprise that a man of Prinz's calibre, working at Ferranti during Nimrod's development, would be left out. How much of the design process he was involved in is unknown, but his German heritage proved incredibly serendipitous when Nimrod made its way over to Germany for the Berlin Industrial Fair. Prinz was able to teach German students and dignitaries on the operation of Nimrod. According to Raymond Stuart-Williams, Prinz also had to invent an entirely new German computing vocabulary to do so!
More than Chess Problems
With that little Nimrod side quest out of the way, Prinz set about completing his chess program. Remember the timeline here - Nimrod was in Berlin in October, 1951, and Prinz had Mate-in-Two working by November the same year. From what I've read online, both the process of inputting a game state for the program to solve, and the computer solving the problem were exceptionally slow and arduous. The piece placements were provided to the computer on punched tape, which is then stored on the computer's magnetic tape memory, which the computer then converts to digital memory.... and only then it can start solving the problem.
The first problem given to Mate-in-Two took around 15 minutes to solve. It tested 450 moves - 100 of which ended up being illegal moves - and ran a self-check diagnostic every time a move was completed (this tested to see if the king was in check or not), which meant that each of those 450 moves took roughly 2 seconds to execute. The eventual solution was literally moving a rook two squares to the left. It took me about 10 seconds to figure out the winning move after I saw the layout - so the computer's not exactly winning any awards for quick thinking, is it? The technology will catch up, eventually...
 |
| This is the first problem Mate-in-Two solved. Can you figure it out in under 15 minutes? |
Summarising Mate-in-Two's legacy is a tricky task. Much of that is because I'm not sure if it has one. One source reports that Mate-in-Two was never used again after its first outing, due to more pressing matters computers were needed to address. There seemed to be some awareness of the program by contemporaries, as Prinz published an article about the program in a journal titled Research in 1952. Not a terribly descriptive title for a journal, and so I wasn't able to find said article in the archives. That article was reprinted in a 1988 book called Computer Chess Compendium, however, but I'm not shelling out $30 for an ebook copy of that, sorry-not-sorry.
Prinz's article was responded to in some manner in a 1953 book titled Faster Than Thought, by B. V. Bowden (the title is inspired by Nimrod). The author is apparently quite harsh in his assessment of Mate-in-Two, citing several improvements that could've been made to the code, and that there wasn't much of a future in computer chess if this was the only programming means available. But the first of anything is hardly ever going to be a success, right? The foundation's been laid for computer chess, at the very least.
And, just so we all know for future reference - I won't be covering every chess program ever made. I think that would be a fruitless exercise, and would be a distraction from my chief goal. I'll consider making mention of significant ones to video game development in the future - that's why we've been discussing Mate-in-Two to begin with. It's one of the first known examples of a "game" program developed for a programmable, digital storage computer, even predating the functional debut of Christopher Strachey's Draughts on the same system.