The Tournament of Tough Plastic

[Reinfeld]

The tournament matched four dedicated machines of roughly equal strength devised by four storied programmers.

Format: Double round-robin
Settings: 1 min/move, everything else default
Monitor: Shredder 12
Disputes: Settled by referee

The players -

Fidelity Designer 2265 (1990)
- The oldest, a brute-force terminator. Spracklen smash!
- Biggest opening book (28,000)
- The only 16-bit machine in the group
- Tricky buttons, limited features
- Worn-out Franco Rocco design
- Familiar Fidelity plastic pieces

Mephisto Milano v. 1.02 (1991)
- Ed Schroder’s long-awaited mid-level machine
- Efficient and cute with laptop cover
- Lots of extra features
- Big opening book (25,000)
- Buttons sort of annoying
- Pieces a bit stumpy

GK 2100 (1993)
- Morsch's improved mass-market monster.
- Big and bulky
- Extra features
- Buttons still annoying
- Smallest opening book (6,000)
- Boring pieces

Novag Obsidian (2005)
- Kittinger zombie (also Star Ruby?)
- Wooden pieces!
- Carrying case
- Smallish opening book (8,900)
- Extra features
- Feels good
- Big friendly buttons
- Amusingly unsound player

These babies are very close together, hovering in that mystical 2000 zone, rarely more than 20 points apart, no matter the measure. Active ratings give Designer 2265 a tiny edge. Tournament ratings are virtually equal – single digits separate the machines. My personal stat, an average of all available Elo numbers, yields these ratings:

Designer 2265 – 1991
Obsidian – 1983
GK 2100 – 1982
Milano – 1962

Each machine played six games. The victor: Milano.

Milano – 4
GK 2100 – 3.5
Designer 2265 – 3
Obsidian – 1.5

All of Milano’s points came from wins: three as Black, one as White, two wins against Obsidian. That was the difference. Milano split games with GK 2100 and Designer 2265 (the latter Milano loss was the final game of the tournament, a decisive Sicilian that salvaged some Spracklen pride.)

GK 2100 couldn’t take two from Obsidian; in the early rounds, the big box blundered away a clear win after punishing a series of foolhardy moves in a Queen’s Gambit accepted. Obsidian salvaged a draw.

Designer 2265 split with the leaders (two wins as Black), but couldn’t squeeze a win out of Obsidian. The first draw was predictable; the second was an opera, a performance that left spectators in awe. (See below)

Obsidian’s performance was tragic. The dashing Novag product played with verve and flair throughout the tournament, but could not muster a single win.

Overall, the black pieces dominated this tournament; the machines notched eight wins with the second move, only one with white (Milano v. Obsidian). There were three draws – all involved Obsidian.

Historians and crabby theoreticians will point to the endgame errors, and they will be right: there were many, typically turning obvious wins into dreary promotion maneuvers. But by and large, the side with the clear advantage prevailed after the middlegame.

If any truth emerges from the contest, it might be this: a two- or three-point advantage is not enough for these machines to win. They still screw up when faced with tenacious and witty defense.

Judges strayed from custom, and awarded the brilliancy prize to the following game, a draw that left both fighters spent. Obsidian is White. It’s a Sicilian, the Richter-Rauzer:

[Event ""]
[Site ""]
[Date "2012.12.06"]
[Round ""]
[White "Obsidian 60"]
[Black "Designer 2265 60"]
[ECO "B63"]
[Result "1/2-1/2"]

1.e4 c5 2.Nf3 d6 3.d4 cxd4 4.Nxd4 Nf6 5.Nc3 Nc6

[fen]r1bqkb1r/pp2pppp/2np1n2/8/3NP3/2N5/PPP2PPP/R1BQKB1R w KQkq - 3 6[/fen]

This is a fork in the road. Obsidian opts for aggression with the Richter-Rauzer.

6.Bg5 e6 7.Qd2 Be7 8.O-O-O O-O 9.Nb3 Qb6 10.f3 Rd8 11.Bb5?

[fen]r1br2k1/pp2bppp/1qnppn2/1B4B1/4P3/1NN2P2/PPPQ2PP/2KR3R w - - 2 11[/fen]

A positional blunder: Obsidian (remember the smaller opening book) has open space and attacking chances on the king side, a natural progression in this variation - but it jumps a bishop to the queen side, playing a move that doesn’t show up in the books or the databases. Fidelity takes its first long think.

11. …d5 12.exd5 Nxd5 13.Bxe7 Ndxe7 14.Qg5 Bd7

[fen]r2r2k1/pp1bnppp/1qn1p3/1B4Q1/8/1NN2P2/PPP3PP/2KR3R w - - 2 15[/fen]

Obsidian stabs the king side, then inexplicably retreats, inviting liquidation at the wrong time. Why help Black de-cramp himself?

15.Qc5

[fen]r2r2k1/pp1bnppp/1qn1p3/1BQ5/8/1NN2P2/PPP3PP/2KR3R w - - 3 15[/fen]

Novag moves in these odd sequences at times, seeing the start of an attack and failing to (thematically) follow through.

On the other hand, the move is powerful. The threat is winning a piece via 15...Qxc5 16. Nxc5, with a simple two-prong attack on d7. Black backs up. White forces the exchange of Queens, doubles the rooks, and seizes the open file.

15. …Qc7 16.Qd6 Qxd6 17.Rxd6 Be8 18.Rhd1 Rxd6 19.Rxd6 a6

[fen]r3b1k1/1p2nppp/p1nRp3/1B6/8/1NN2P2/PPP3PP/2K5 w - - 0 20[/fen]

The attacker has to get something out of this flurry, even if it's just better placement. Shredder likes 20. Bxc6, but Obsidian retreats, losing a little time.

20.Be2 Nf5 21.Rd2 b5 22.Bd3 Ne3?

[fen]r3b1k1/5ppp/p1n1p3/1p6/8/1NNBnP2/PPPR2PP/2K5 w - - 2 23[/fen]

Impetuous. Shredder likes 23. Re2 here, gaining time, but Obsidian chooses a more passive option, allowing Black’s knights to charge.

23.a3 Ne5 24.Be4 Rb8?

[fen]1r2b1k1/5ppp/p3p3/1p2n3/4B3/PNN1nP2/1PPR2PP/2K5 w - - 3 25[/fen]

This was Fidelity’s first real mistake. The right answer was 24. …Bc6. Black dreams of a knight on c4, but White should shove his knight to c5, forcing Black to fiddle around with defending the loose a-pawn.

25.Nc5 a5 26.Na6 Rb6

[fen]4b1k1/5ppp/Nr2p3/pp2n3/4B3/P1N1nP2/1PPR2PP/2K5 w - - 2 27[/fen]

Here, White has the nasty 27. Nc7!, but Obsidian, somewhat understandably, opts for a tempting and seemingly more forceful alternative.

27.Rd8 (Yow!) f5 28.Nc7 fxe4 29.Rxe8+ Kf7 30.Nxe4

[fen]4R3/2N2kpp/1r2p3/pp2n3/4N3/P3nP2/1PP3PP/2K5 w - - 0 30[/fen]

Holy crap, look at that stack of knights.

30. …h6 31.Rc8 Nxg2 (Black regains his pawn. Novag's failure to get these boys involved earlier could be telling.) 32.Nxb5!

[fen]2R5/5kp1/1r2p2p/pN2n3/4N3/P4P2/1PP3nP/2K5 w - - 0 32[/fen]

Nice. Takes a pawn, threatens 32...Rxb5 33. Nd6+!, winning the exchange. Fidelity responds instantly. 32. …Nxf3

[fen]2R5/5kp1/1r2p2p/pN6/4N3/P4n2/1PP3nP/2K5 w - - 0 33[/fen]

Fidelity lashes back, regaining the pawn, but allowing the White knights to centralize.

33.Nbd6+ (Look at all those horses.) Kg6 34.Rf8 Nge1 35.Re8 Nxh2?

[fen]4R3/6p1/1r1Np1kp/p7/4N3/P7/1PP4n/2K1n3 w - - 0 36[/fen]

A lemon. White has a check here (36. Rxe6) but Shredder likes the trickier 36. Nc4! Obsidian, with a full minute to calculate, doesn’t see it.

36.Rxe6+ Kh7 (Better to hide) 37.Re7 Rc6?

[fen]8/4R1pk/2rN3p/p7/4N3/P7/1PP4n/2K1n3 w - - 3 38[/fen]

Another lemon. Fidelity is blowing it. This time, Obsidian catches on.

38.Nf6+ Kg6 39.Nfe8!

[fen]4N3/4R1p1/2rN2kp/p7/8/P7/1PP4n/2K1n3 w - - 6 39[/fen]

Heroic – the type of move I never see.

I watched this game in real time with Shredder’s analysis off, just trying to think what I would do as White. My only thought was defense. Black has two checks that look dominant: 39…Nd3+ or 39…Rxc2+

It’s so hard (for me, anyway) to look past those checks, to think one move deeper, and visualize one more after that, and realize that the White king can sidestep, Black won’t have a useful move and White can cash in with initiative. The nasty checks don’t matter. So then what? Soltis talks about “two-and-a-half move chess.” This is a perfect example.

White ignores the checks and finds a multi-purpose move: protect the d6 knight, threaten the g7 pawn, threaten the knight at e1. Black must choose – he keeps the e1 knight, but that leaves an opening for White’s cavalry to march the King up the board in spectacular fashion.

39. …Nhf3 40.Rxg7+ Kh5 (forced) 41.Nf6+ Kh4 42.Nf5+

[fen]8/6R1/2r2N1p/p4N2/7k/P4n2/1PP5/2K1n3 w - - 4 42[/fen]

The horses! Christ, you almost hear the trumpets.

42. …Kh3 43.Rg3+ Kh2 44.Ng4+ Kh1 45.Nfe3 Nd3+ 46.Kd1 h5

Just stop and look at this position. Obsidian has the game in hand – Shredder gives it a 4.06 advantage. The finish should be easy...

[fen]8/8/2r5/p6p/6N1/P2nNnR1/1PP5/3K3k w - - 0 47[/fen]

47.Rh3+ (Here, Fidelity pondered blocking with 47. …Nh2? – certain death) Kg1 48.Rg3+ Kh1 49.Rh3+

[fen]8/8/2r5/p6p/6N1/P2nNn1R/1PP5/3K3k w - - 5 49[/fen]

Novag, strangely, toys with perpetual check. If Fidelity wants a draw, speak now, or give up the knight.

49. …Kg1 50.Rxf3 (Novag rejects the draw. The Black horse stumbles, wounded, choking in a death spray of blood and sputum. Black still has a check at b2.) Nxb2+ 51.Kc1 hxg4 52.Rf1+ Kh2 53.Kxb2 g3

And look again. The black horses are gone. Obsidian commands the field. Black has next to nothing left, but the g-pawn must advance.

This point - 54 moves in - carried the game to sublime heights.

Obsidian, above all, needs to play a waiting game and stay on the first rank. Instead, it opts for a dubious venture. Fidelity blunders in response. Obsidian blunders again. Fidelity blunders once more, and Obsidian fails for a third time.

[fen]8/8/2r5/p7/8/P3N1p1/1KP4k/5R2 w - - 0 54[/fen]

54.Rf7? (Loses dominance.) Kg1? (Leaves an escape hatch for Obsidian to rush back with 55. Rf1+) 55.Kb3? (Obsidian, you're an idiot.) Re6? (Again allows the check at f1) 56.Nf5??

[fen]8/5R2/4r3/p4N2/8/PK4p1/2P5/6k1 w - - 5 56[/fen]

Possibly the worst move of the game, sitting at the edge of Obsidian's horizon. This blunder restores near-equilibrium.

56. …g2 57.Nd4? (Compounding the loss of advantage. Perhaps Obsidian is too fixated on the loose Black pawn at a5. Now Black gains time.) Re3+ 58.Kc4

[fen]8/5R2/8/p7/2KN4/P3r3/2P3p1/6k1 w - - 3 58[/fen]

The promotion threat on the a-file erases White’s material advantage. Shredder sees a dead draw.

58. …Kh2 59.Rh7+ Kg3 60.Nf5+

[fen]8/7R/8/p4N2/2K5/P3r1k1/2P3p1/8 w - - 7 60[/fen]

This is Obsidian's hoped-for payoff - forking the rook and king. But Black still gets a Queen out of it. We end up with unbalanced relative values: R+N+P+P vs Q + P. Will Fidelity see the threads?

60. …Kf2 61.Nxe3 g1=Q 62.Re7 Qa1 (Not best, says Shredder.)
63.Kb3 Qb1+ 64.Ka4 (Not best. Will Novag blow this one completely?) Qb6 65.c4?

[fen]8/4R3/1q6/p7/K1P5/P3N3/5k2/8 w - - 0 65[/fen]

Novag has frittered a 4-point lead into a 1-point deficit.

65. …Qc6+ 66.Kb3 a4+ 67.Kc2 Qc5 68.Nd1+ Kg2 (White is stretched very thin, but Fidelity falters just a bit. The horizon effect looms - we are in six- to seven-ply territory. The position is complicated.) 69.Re2+ Kf1 70.Rf2+ Kg1 71.Rf4 (Concedes the neglected a3 pawn.)

[fen]8/8/8/2q5/p1P2R2/P7/2K5/3N2k1 w - - 9 71[/fen]

71. …Qxa3 72.Rg4+ Kf1 73.Rf4+ Kg2 74.Nc3 Qb3+ (A thread of darkness oozes across the horizon...) 75.Kd3 a3 76.Re4 Qb6 77.Re5? Qg6+ 78.Kd4

[fen]8/8/6q1/4R3/2PK4/p1N5/6k1/8 w - - 5 78[/fen]

Decisive. It's over now. Obsidian has thrown this game away. Even the drawing chances are bleak. Shredder gives Black a 4-point advantage, recommending 78. …Qf6, pinning the rook and gaining time.

78. ...Qg4+??

[fen]8/8/8/4R3/2PK2q1/p1N5/6k1/8 w - - 6 79[/fen]

But wait - Fidelity delivers the worst move of the game!

79.Kc5 Qd7

[fen]8/3q4/8/2K1R3/2P5/p1N5/6k1/8 w - - 8 80[/fen]

Incredibly hard to see – way beyond the horizon line – but this move gives White just enough room to gain time, march the pawn toward promotion and distract the wave of checks.

80.Re1??

Novag responds with an equally terrible move, handing back the advantage in spades. Shredder gives Black a 7-point advantage after 80. …Qd2!

80. …Qb7??

[fen]8/1q6/8/2K5/2P5/p1N5/6k1/4R3 w - - 10 81[/fen]

Fidelity answers with a horrific move. The sequence from White's 77th to here is astounding.

81.Re5??

[fen]8/1q6/8/2K1R3/2P5/p1N5/6k1/8 w - - 11 81[/fen]

Fidelity has a clear win here with 81. …Qb2 – and misses it. The referee, appalled by the punch-drunk fighters, began to ponder the exhaustion factor...

81. …Qa7+? 82.Kb4 Qb6+ 83.Rb5 Qd6+ 84.Kb3 Qd2 85.c5

[fen]8/8/8/1RP5/8/pKN5/3q2k1/8 w - - 0 85[/fen]

Unable to stand the strain, the referee stopped the fight and declared a draw. Pandemonium ensued, with both machines claiming an advantage.

1/2-1/2

- R.

[Steve B]

Settings: 1 min/move, everything else default

Good Time Control R

I especially enjoyed the physical description of each computer

Stumpy Regards
Steve

[Larry]

Settings: 1 min/move, everything else default

- R.

I never saw the GK2100 housing as bulky,
rather saw it as a good size, with good buttons and a good display.
I only wish it had a light for each square.
The franco rocco Designer style I always saw as tacky.
The Obsidian is a rather boring design.
The laptop style is pretty much the ideal design for a plastic machine.
Pity they are pressure sensory, not auto sensory.
The tourney was nice, and a good description of it too.
thanks.... Larry

[SirDave]

Very interesting Reinfeld- thanks for the effort. I enjoy reports that involve comparisons of various boards. We sometimes find out new pieces of information such as, in this case, the unexpected poor performance of the Obsidian, an unfortunate additional blackmark to its very poor openings diversity to e4 and d4.

Regarding Reinfeld's comment re: Fidelity 2265, 'Worn-out Franco Rocco design.' and Larry's: 'The franco rocco Designer style I always saw as tacky.':

IMO, the design was grossly overused and it does look tacky in the cheaper, weaker models, but there is one exception: the 2265. In its case, Fidelity got the color combination just right which seems to enhance the design itself. When the 2265 is turned on the combination of the red lights offset by the overall brown color of the board and the gold trim around the LCD display makes it downright attractive and distinctive from any other board.

It is because of that and other more practical parameters that make the 2265 one my favorite laptops. Others apparently feel the same, because the 2265 is one of the few plastic units currently selling at prices only seen with wood.

In support of this premise is the interesting fact that both the Designer 2100 Display and the Designer 2265 Mach have almost the same firmware and manual- the only major difference I can find is that the 2265 has a Selectable User Time Controls. Other than that the ELO difference is less than 50-75. And yet the 2100 Designer Display doesn't sell for much more than $75-150 on eBay while the 2265 is now going for $500+. One major reason methinks is that the 2100 is downright drab compared to the 2265 in addition to the fact that there are relatively fewer 2265s out there compared to the 2100 Display.

[Reinfeld]

IMO, the design was grossly overused and it does look tacky in the cheaper, weaker models, but there is one exception: the 2265. In its case, Fidelity got the color combination just right which seems to enhance the design itself. When the 2265 is turned on the combination of the red lights offset by the overall brown color of the board and the gold trim around the LCD display makes it downright attractive and distinctive from any other board.

I was being a bit flip about the Franco Rocco design. I like the first Designer 2000, with the silver and black setup (no display) and the thin metal sheet, which felt sleek and classy.

The later versions (2000 Display and 2100 Display, also Chesster) are a bit less attractive to me, with a less rigid board - almost metallic tape. They felt slightly cheap.

I do agree with the point about the color scheme on 2265 - a complete shift. I like the chocolate-brown pieces as well. I still haven't tracked down a Designer 2325 (a big want on my list), but it looks cool in pictures, with the green and cream tournament-style colors.

It is because of that and other more practical parameters that make the 2265 one my favorite laptops. Others apparently feel the same, because the 2265 is one of the few plastic units currently selling at prices only seen with wood.

The 2265 is one of my faves as well, partly because the board is just a shade bigger than the Mephisto models, and things are easier to see.

However, I think the chief reason the 2265 fetches prices comparable to wood is strength. No more, no less. It plays tough and it plays fast. Tackle the thing straight up, and it kills you. It's the second-strongest tabletop in the Fidelity line (unless I'm mistaken, the EAG wood versions simply house the Mach III program in a classier box.)

The later Mephistos and Novags (and the RISC 2500) ultimately surpassed the 2265, obviously. But only the diehard collectors know that. And as this tournament shows, the 2265 still plays a good strong game. When you rank the best plastic machines (exclude wood, exclude clones), the 2265 still hovers in the top 15.

Other than that the ELO difference is less than 50-75. And yet the 2100 Designer Display doesn't sell for much more than $75-150 on eBay while the 2265 is now going for $500+.

A small quibble, but I find the Elo difference between these two to be more than 100 points, based on overall averages. I've pitted 2265 against 2100 a few times. It is no contest. I don't know what they did to juice up the firmware, but the difference is substantial. Maybe I'll try a short match between these two, just to test it.

I never saw the GK2100 housing as bulky,
rather saw it as a good size, with good buttons and a good display.

Again, I was teasing a little. The GK 2100 is a good machine, but like the Rocco design, the housing was copied for multiple machines, and grew tiresome. Not quite as big as the 2250 XL (nothing's bigger than that one), but certainly sizable.

My gripe about the buttons is more serious. I like a big button. Those little Mephisto tabs are sometimes hard to feel - and I'm never quite sure I've done the right thing.

The Obsidian is a rather boring design.

On this, I respectfully disagree. My eyes have never been good, and I like the clean, easy quality of Obsidian's buttons. I also like the pieces. The board is a bit thicker than the others we've been discussing, but the rounded corners make it pleasant to use.

(Also consider the unusual Diamond/Star Diamond design - again, rounded corners, though the slanted edges leave no space to stack taken pieces. The pieces are a big disappointment, however.)

I'm biased toward Novag, obviously. I forgive the inflated ratings. The company strikes me as more attentive to the casual user - in an Apple kind of way, if I can dare that comparison. The cases for Obsidian and Star Diamond are also a major plus - a genuine cover, the faux leather bags, etc.

By comparison, the Mephisto Master - the top of the plastic line for that company - comes with a flimsy "cover" that barely deserves the name. Milano's removable top was better.

With all that said, I'm still waiting for Steve Blincoe's hoped-for next-gen Novag machine...

Star Citrine regards,

- R.

[Reinfeld]

Footnote: What conclusions (if any) can be drawn from Black's dominance in this tournament? Only one win for White in 12 games, and 8 wins for Black...

Opening stats (simplified categories):

d4 openings (QGD-3, QGA-2, Reti-1)

White's record -
Wins: 0
Losses: 4
Draws: 2

e4 openings (Sicilian-3, Petroff-1, Caro-Kann-1, Vienna-1)

White's record -
Wins: 1 (a Vienna - Milano v. Obsidian)
Losses: 4
Draws: 1

- R.