The third European Microcompuiter Chess tournament. Anthony Harrington.
The Third European Microcomputer Chess Tournament
The Third European Microcomputer chess tournament, held at the Barbican Centre during the PCW exhibition, September 9-12, 1982 proved once again that computer chess is a thriving force.
A seven-round Swiss tournament is always an exciting affair with something of a lottery about it. The basic rule of such a tournament, for those who are unfamiliar with Swiss events, is that winners play winners and losers play losers. Each time the draw is made for a new round, the tournament director tries to pair opponents with the same scores against each other.
Because of this, the going gets easier for the weaker players (programs in this instance) and more and more difficult for those who are doing well. It is a sure way of finding the strongest player/program in the tournament. And it produces some marvellous final rounds.
This year's concluding round was in a class by itself, with the eventual overall winner, Mark Johnson and David Wilson's Advance 2.4, saving its unbeaten record at the last moment by a quirk of fate peculiar to computer chess.
But more of that final round later. Like all chess tournaments, even the human kind, this event began long before any processors found their way into the playing hall in the Barbican Centre.
This year there was an added complication to the usual organizational difficulties that surround such an event. Six months before the tournament was due to begin PCW changed hands. Bunch Books, the former owners of PCW, who staged the first European tournament back in 1978, sold the magazine to VNU Publications, a Dutch publishing group.
On top of this change, David Levy, the international chess master, and the man responsible for orgainising the tournament in previous years, decided, quite rightly, that his commercial interests in computer chess programs made it impossible or him to continue to organize the event.
The upshot of all this was that invitations and entry forms went out on short notice, a month before the tournament was due to start. Nevertheless, the tournament attracted a very strong field, totalling 14 entries in all. This was two more than last year, though program failure in the case of two entrants meant that the numbers actually contesting the tournament turned out to be identical.
The final line-up included four commercial entries (programs or dedicated chess computers actually on sale to the public or which would be available before the end of the year) and four experimental entries from commercial companies.
It might have been even stronger, but Fidelity, of Miami, FL, whose Sensory Chess Challenger 9 is one of the strongest chess computers around, decided against entering a week before the start. (According to John Renaldo, Fidelity's marketing manager they were considering entering either a commercial model or an experimental entry or both.)
The West German company Hegener and Glaser, who make the Mephisto, also almost entered, but decided at the last moment that they didn't like Swiss tournaments and would concentrate instead on the US World Championships in March 1984.
One has to sympathize with commercial companies. They have a lot at stake as far as potential sales are concerned. And if their machine, for some reason, fails to win one of the top three placings in a tournament it is tempting for their managing directors to feel that the public will take the hint, and buy the machine which did better.
My own feeling here is that they would do better to compete. If they do feel their commercial model is a little delicate, there's no harm in entering an experimental version. It would give their programmers a chance to test their ideas, and the results wouldn't affect their market share, provided the name given to the experimental entry bore no relation to the market model.
Still, if one remembers that the commercial class in the World Championships, held at L ubeck-Travem unde, West Germany, in September last year, consisted of only four entries, one of which withdrew after a few rounds, the commercial line up wasn't bad. And the amateur entries were excellent.
It is the amateur entries that make these events. Unburdened by the financial anxieties of the commercial entrants, the amateurs give the tournament its enthusiastic, convivial atmosphere. Whereas commercial competitors have a tendancy to get a little grim at the board and to be sharp and short with opponents, the amateurs tone down the seriousness of it all--and are just as willing to laugh at the odd strange move generated by their own program as they are at their opponent's.
In all there were five amateur entries. From Hamburg, Dieter Steinwender, a computer science student, led a team of four German students with an entry called Micromurks, running on an Apple II.
Also on an Apple II was White Knight (Mk 10), written by Mark Bryant. This, like Richard Lang's Cyrus II, running on an Osborne, was an improved version of a program entered into last year's PCW tournament. Mark Johnson and Dave Wilson, with Advance 2.4 and Jeff Cooper with Chess '86, also seasoned PCW tournament campaigners, completed the amateur turnout.
In passing, one should probably say that the "amateur status' of several of these programmers simply means that they are not selling or about to sell their entry programs, (though they might, if an offer came along). Johnson, Bryant, and Lang are all employed by Intelligent software, a company set up by David Levy and Kevin O'Connell to write and market chess programs (and other intelligent games).
In the experimental commercial class (programs which won't go before the public in their present form), there were two entries by SciSys, the Hong Kong based company whose machine, Mark V, won the Travem unde world championships. These were the Mark 5.01 Experimental and the President Turbo Experimental. The Austrian chess programmer Wim Rens, sponsored by the UK firm Microtrend entered a program called Gambeit '86, and Intelligent Software put in an entry called Philidor.
The commercial section proper, consisted of La Regence, a "sensoryboard' chess computer designed by Intelligent Software; another sensoryboard machine, called Bogol, by the Hamburg company Omikron (a newcomer to the commercial chess machine market); Spectrum Chess, which runs on the ZX Spectrum, designed by Arctic Computers; and an Irish entry, Conchess, another sensory-board machine.
On the morning of September 9 the competitors assembled in the playing area in the Barbican Centre, set up their machines and programs and promptly experienced the first (and, thankfully, the last) major delay of the event. Someone had wired up a plug incorrectly and blew the fuse for that entire floor sector of the Barbican exhibition hall.
Bear in mind the fact that the tournament was part of the PCW micro exhibition and hundreds of micro suppliers were putting the finishing touches to their stands while we were hunting for an electrician and you will get a fair picture of the start conditions.
Finally, after a half-hour delay an electrician was found to fix the fault and uncover the culprit, who shall be nameless, and the first round began. The draw and results for the first round are shown in Figure 1.
Play for the whole of this tournament was at the rate of two minutes a move, with the first time control at 30 moves in the first hour and then 23 moves in 46 minutes. This is a somewhat faster rate than is usual in tournament play, but it was forced on us by the tight closing hours of the exhibition.
Doors opened for the players and exhibitors at 9:00 a.m., (10:00 a.m. for the public), and closed at 7:00 p.m. In order to get through two games a day, we had to limit each round to approximately 3 1/2 hours.
Players were allowed a maximum of 30 minutes down time to rectify any program crashes, so the possible duration of each round was 4 1/2 hours. Since the theoretical starting time for the morning round was 10:00 a.m., and for the afternoon round, 2:30 p.m., a little mental arithmetic will tell you it was a close squeeze fitting both rounds in.
Because of the time pressure, it was decided to adjudicate all unfinished games at move 53. This set the tournament director, Stuart Reuben, some complicated adjudication problems in the course of the four days.
This first round was characterised by two events that looked more like absences than things that would shape the outcome of the tournament. Advance 2.4, running on a home constructed bitslice machine found itself playing the white side of what must be one of the most boring Benoni defenses of all time.
Philidor, the opposing program constructed a blocked middle game in which the best either computer could think of doing was to shuffle rooks back and forth. The programmers read the papers while waiting for the inevitable draw by repetition. Fortunately Advance 2.4 went on to win the next five games in a row, in more aggressive fashion.
While this was going on, the Omikron team of Martin Suhl and Gerd Krey were busy missing the first round. Unfamiliar with London traffic, they'd made the mistake of taking a taxi from Oxford Street instead of a tube. Their program, Bogol went on to win four out of the remaining six games, and who knows, win in this opening round might have seen them draw level for top commercial entry with La Regence on 5 points.
The Second Round draw and results are shown in Figure 2.
As the keen-eyed will have noted, the tournament had 13 entrants at this stage, though this was soon reduced to even figures for most of the tournament. Chess '86 collapsed during its game against Gambeit '82 and Jeff Cooper only managed to rectify it in time for the final round.
The day was exceptionally warm and the Barbican Centre was attracting record crowds. Computers don't like heat and several of them began to show their dislike. One of the SciSys entries, President Turbo overheated in a complex middle game against La Regence. It was a miracle that the President Turbo started the tournament at all, as Andrew Page, the SciSys UK general manager only received the module late the previous evening by airmail from Hong Kong.
Ideally, he would have liked a few weeks to test the program before the tournament, and we had to admire the competitive spirit that chose to enter rather than withdraw under such circumstances.
Round Three draw and results are shown in Figure 3.
Richard Lang won last year's tournament and his program, Cyrus II had looked good up to this point. But La Regence proved too strong. Dieter Steinwender and the German team were becoming very disappointed with their program, which was showing definite flaws.
Round Four produced one of the key games of the competition.
La Regence put up a dour struggle against Advance 2.4, and the game went to the 53 move adjudication but Advance 2.4 was two clear connected passed pawns up in a rook ending and Stuart Reuben had no hesitation in awarding it a win.
No one knew it at this stage, but Cyrus II had won its last game, beating President Turbo on adjudication. Reuben had his work cut out in this round, for White Knight and the Mk 5.01 also had to be adjudicated. This last was a very complex position, which either computer could have mishandled with ease. A draw seemed the fairest conclusion.
Round Five saw White Knight and Philidor falter and lose their way, though Mark Bryant with a final score of 3 1/2 was still to finish strongly enough to take the second amateur placing and a cheque for 50, while Philidor, on 4 1/2, would take third place overall in the tournament.
The game against Bogol was a stiff test for David Levy's La Regence in Round Six, and it was not without its commercial appeal as well as its chess interest, for both these sensory-board machines are due to be marketed before the end of this year.
After this loss, Martin Suhl was undecided whether he was pleased or displeased with the performance of his machine. No obvious flaws had been detected and its three wins were very reasonable chess games. In retrospect, the two losses Bogol suffered turned out to be to the machines which came first and second overall, and after picking up another point in the final round, the German team went home smiling.
Philidor had an easy game against Spectrum, but Richard Turner, of Arctic Computing thought that the ZX program had held its own well against computers that were very much more powerful. It was the only cassette tape module program in the competition and this is one area where one can expect a greater number of entries next year. Advance 2.4 by this stage looked unstoppable, and it duly crushed President Turbo with a fine mating attack.
The final round was full of surprises. The Mk 5.01 played a marvellous game against Advance 2.4, won material in the opening and looked set to force mate by move 26. But Advance 2.4 held on and although two rooks down, managed to find a position where it could force perpetual check with its queen--the human equivalent of a "swindle.'
On board two meanwhile, the Intelligent Software entries finally came up against each other, and much to everyone's suprise, Philidor, which hadn't done much of note till then, chose that moment to play a complex, crafty game against the fancied La Regence and stopped its stable mate from winning 6 out of 7 games.
So although the final round had started with the faint possibility of a tie for first place between La Regence and Advance 2.4 (if the former won while the latter lost), the gods decided otherwise and gave us a clean, clear winner.
Mike Johnson and Dave Wilson took away just about every prize on offer. They won the trophy for the overall winner of the Third Microcomputer Chess Tournament, the trophy for the best amateur entry and a cheque for 150. La Regence was named as the best commercial entry in the tournament and Mark Bryant, as we said, collected a cheque for 50 for the second highest placed amateur entrant.
Table: Figure 1. Draw and results, Round One.
Table: Figure 2. Draw and results, Round Two.
Table: Figure 3. Draw and results, Round Three.
Table: Figure 4. Draw and results, Round Four.
Table: Figure 5. Draw and results, Round Five.
Table: Figure 6. Draw and results, Round Six.
Table: Figure 7. Draw and results, Round Seven.
Photo: The prototype chessboard for the German machine Bogol (foreground) was built by an excellent craftsman who knew nothing about chess (note the white square in the bottom left corner). Here it is playing against La Regence from David Levy's Intelligent Software Company, the eventual commercial victor.
Photo: Spectrum Chess from Arctic for the Sinclair ZX81 held its own against much more powerful competitors such as Sci Sys' MK 5.01 experimental entry (foreground).
Photo: Along with the 3rd Microcomputer Chess Tournament at the PCW Show, the Computer Scrabble Tournament proved popular with participants and spectators alike.