My 'occasional player level, class D' chess computers

Fidelity Sensory Chess Challenger 6

Year: 1982
Programmer: Ron Nelson
CPU: 8048 @11Mhz
ROM: 4Kb
Elo level: 1212 (1300 FIDE)
CMhz: 0.85
Rperf: 73%
Square size: 1"

A relatively insignificant electronic chessboard, as it succeeded the Sensory 8 displayed in this page (see below), at a lower cost at the expense of removing the diods on squares, replaced by a reduced display (only two LED characters, while a CC7 for example displays four of them), featuring a weaker processor and less playing levels to choose from... Nevertheless, a few small enhancements including a slightly larger openings book (but still embryonic), and enablement to display a hint move, and also to take back two half-moves. The program is placed in a removable cartridge, but Fidelity never offered any stronger play module; only a "book openings" one for classic openings practice, and a "greatest games" module for master games study. By the way, both modules were as well available for the Mini Sensory (see previous category), but used different references: the cartridges were not compatible (!). To make it short, the modular feature was quite useless for such  mass market devices, by design limited by their low-end CPU. Indeed the playing level weakened with regards to the Sensory 8, but is ahead of the Mini Sensory+CAC, mainly thanks to a faster clocking. So, little reasons to add this one to my collection... but opportunity: €5 on french "le bon coin", mint condition, complete with the original mains adapter and set of pieces, the small user guide like new, and the original packaging. How would you resist? And to be fair, without any protuberant LED on the board, it is a clean looking chess computer.

VEB RFT Chess-Master Diamond

Year: 1987
Programmer: Rüdiger Worbs & Wolfgang Pähtz
CPU: U880 @4Mhz
ROM: 16Kb
Elo level: 1231 (1315 FIDE)
Elo level with PM11: 1283 (1354 FIDE)
CMhz: 1.35
Rperf: 72% (with PM11: 75%)
KT: 1134 (with PM11: 1171)
Square size: 1.38"

The Chess-Master Diamond is the ultimate evolution of RFT VEB Mikroelektronik Erfurt products, from the former East Germany. If the processor is still a copy of the Zilog Z80 (U880), it runs notably faster than the SC2 one (displayed in the previous category), featuring 4Mhz instead of 2.5Mhz. It is a still more massive (weighting around 4kg!) chess computer than the SC2, despite not including a transformer (the power source is an external mains adapter); as it is made of wood, and of imposing size. Everything is pretty large: the chessboard squares (1 3/8 inches sides), the control panel, the display (3 1/8 inches, including the surround!). To realize how much it is nice and impressive, you need to see it in real life, pictures are not enough. Mine deserves no value for collection, as only the electronic chessboard (and its cardboard box) are original: the mains adapter is a modern one, the set of pieces is not the original one, and the optional modules are missing. The offerred pieces were part of my decision to buy; the former owner had chosen them with great taste, as they are VERO chessmen from the former GDR (which is a consistent choice), they are much nicer than the original ones; and I do love their design - I had already purchased a similar set to better fit out my King Performance. He bought this set €50 (I trust him, I bought mine €49.90!) and adapted the required magnets to let the sensitive board detection work. Total cost €58 including the set of magnets; I payed €150 for the whole lot; that is to say €92 for the chess computer itself, including the adapter; in mint working and cosmetic condition. The price seemed deserved to me, and it was! As you surely understood, the board is an auto-response one, detecting the moves thanks to Hall effect sensors. With regards to the thickness of the wooden board, the inserted magnets (South pole below) have to be strong. The program shows its own moves using one LED in each square, the from-square permanently lightened, the to-square blinking. By default, the display shows who is on the move ("Y" for "You", else  "C" for "Computer"), but shows alike which side (up or down arrow) - and possibly additional information, such as '+' to announce a check situation, or 'MATE' if so. But you can set the display to show, according to your preference, any one from a bunch of other data. You can choose to display the played move, the move number, or the computer score, or the depth of computation, or any of the many available informations. Of course you can browse thru all those, and the last displayed one, before resuming playing, will stay enabled from then on. Simple and convenient! In addition to displaying information, the offered features are many, sometimes advanced. In an article for the 'Modul' magazine published in January 1989, Thomas Mally summed it up this way: 'a high technology dinosaur, with features from today, in a dressing from yesterday, with a program from the day before'. Actually, the level of play is weak, despite enhanced with regards to the preceeding devices from this brand! The computer includes a connector to enable inserting an external module, two were offerred: PM10 ("opening module", 8K) containing a library of openings, and PM11 ("endgame module", 16K) providing enhanced algorithms. I will get back to the importance of these with detailed comments, leveraging the Khmelnitsky test outcomes. The lack of PM10 openings module could indeed be a concern, as the basic program computes as soon as its first move; this could result in always the same moves to be played, assuming same thinking time. Well, as a good demonstration of the extensive features, those can solve this point: the program offers a "BEST" setting to always choose the move with the highest score, if enabled; or provide more variety based on a random choice among the best moves, if disabled. I experienced playing openings with BEST disabled, four moves using this mode are enough to get a diversity of variants at least equivalent to a small library featured by a Chess Challenger 6 or 7, and after five to six moves played, the variety of games is more than enough. I use this mode combined with the display of the move number, and I enable again the BEST mode before inputing my sixth move. So, openings variety, then optimal moves (everything is relative)!

The advanced features of the Chess-Master Diamond enable to perform the Khmelnitsky test, they even facilitate it a lot! One can permanently set the display to show the computer score; the setup of positions is really efficient and fast; even the pre-selection of moves before analysis fits for a practical use. Actually, as with any chess computer, the played move sometimes do not belong to the list or responses that are expected according to the test; in such a situation I play each proposed move (allowing a reduced thinking time) and I write down the resulting score, from the opponent's point of view. The worst score is then the program's best choice... Nevertheless, if the score is displayed with only one decimal digit, sometimes it is not accurate enough to identify the best choice of the program... Well, in such a situation, you can reduce the scope to let the computer analyze only the suggested moves, leveraging the "MOVE" function (still granting a reduced thinking time, of course). This enforces the computer to choose from the "valid" moves, and easily solves the issue! In addition, weak chess computers able to run the test are very rare; getting such a profile is unprecedented as far as I know! I performed the test without any module, consistently with my idea to test a weak program (using the basic 16K program so, as it is the configuration I own). The test confirms one can cheerfully do without the opening module (PM10), as this development phase is the strongest domain of skills of the program. It achieves better results in this domain than the Kittinger's Novag Emerald Classic, than the Donninger's Tiger Grenadier (who said "that's easy!"?) and plays on par with Kaplan's programs used in the Saitek Blitz or the Turbo King II D+! It is as well incisive with counterattacks, on par with a Fidelity Excellence or a Saitek Blitz. To complete the positive aspects, the strategic knowledge can be compared to the one of the Mephisto Mirage, and it has a rather good recognition of threats (again, everything is relative!). As with many other programs, the strength drops down at endgame, and initiative for attack is lacking. I trust you would like to see the contribution from the PM11 module? This "endgame" module provides 16K additional program, it has been developped with support from IGM Rainer Knaak. A bit surprisingly, the user's manual recommends to insert it as soon as the openings module ends, and up to suggest to withdraw the first move computed by the basic program, in order to compute again with the inserted module. Is the basic program so bad? Or, is the PM11 module so stronger? Anyway, for a named endgame module, one would rather expect some other rule, for example according to the remaining forces. Thanks to the excellent CB-Emu emulator, I could achieve the test with the module, despite not owning it (*):


As an actual outcome, the module displays a more balanced profile, more rounded, achieving a higher global score with 37 more Elo points (thus 1171). To help sizing the difference: on a statistical basis, assuming twenty games played, a 1200 Elo player wins 12 of them against the basic program, while winning 11 against the PM11 module... Obvious gains relate to spotting sacrifices, and tactics. Both middlegame and defense profit from the module, with close to 90 Elo points more. Endgame is only enhanced in standard positions (nevertheless +60 Elo points). Otherwise, the additional focus applied to these domains lowers the crest of strength for opening and counterattack, which does not matter much, considering the wide margin available compared to the average level of play. The loss in the strategy domain (-74 Elo points) is on the other hand sort of a pity.
(*) Three months after I wrote these lines, I had an opportunity to purchase a PM11 module; it was offered on eBay, buy it now and priced €49.

Fidelity Chess Challenger 7

Year: 1979
Programmer: Ron Nelson
CPU: Z80 @3.6Mhz
ROM: 4Kb
Elo level: 1259 (1336 FIDE)
CMhz: 1.21
Rperf: 74%
Square size: 0.87"

My very first one! Bought brand new once available... Then bought again about thirty years later for €45. This device made electronic chess affordable for most ones. The design is restrained and pleasant, despite being fully plastic made. The animated while thinking big red LEDs display is nice, it helps granting the Chess Challenger some personality (it seems to enter panic mode when put under pressure!). Even pieces are nicely designed, to such a point I sometimes use it as a passive chessboard. Playing level is weak, especially since I know its play inside out...

Novag Chess Champion Super System III

Year: 1979
Programmer: Mike Johnson
CPU: 6502 @2Mhz
ROM: 8Kb
Elo level: 1270 (1344 FIDE)
CMhz: 2
Rperf: 73%

A high-end chess computer at that time, featuring a strong hardware (2Mhz 6502), an advanced 8K program, expansion capabilities with a full chessboard LCD display and a printer; both can be plugged sideways after removing plastic covers. Sophistication mostly concerns chess features, not playing strength: under-promotion, 'en passant' capture, draw rules including stalemate, three-fold repetition and 50-moves rule, instant display of selected piece once 'from' coordinates are input... A very complete chess computer at that time, somewhat complicated to use. Mike Johnson belonging to David Levy's team will also author the Novag Chess Champion Pocket Chess / Chess Traveller (using 4Mhz Fairchild F8) and the Chess Partner 2000. Playing level is set, such as the Boris, with a timer, to the nearest second. I like it... The one I own, bought at a low price (€20) on German eBay, featured a quite unreliable keyboard: sometimes, randomly, a key input was erroneously scanned (for instance, a 'B' displayed instead of a 'A', a '4' instead of a '6'...), however carefully the key was hit (a brief or long press, a clear split between two hits, and so on...). Maybe one input out of twenty to thirty required to be cleared and restarted, this did not prevent playing but was uncomfortable. I realized some day this issue completely disappeared when I set the time switch to on. I used to play 'time off' to display the preferred move during program thinking time; whereas 'time on' displays the remaining time down to computer move. I faithfully get keyboard errors again whenever I set 'time off', and 'time on' results in a never failing keyboard scan! I do not know if this issue is specific to the SSIII I own, or is commonly shared by other ones. It also happened to my SSIII to be found stuck after leaving it a while (such as a quarter of an hour) on but unused; while during continuous play I never encountered such an issue.