2250XL-ExPro match?

This forum is for general discussions and questions, including Collectors Corner and anything to do with Computer chess.

Moderators: Harvey Williamson, Steve B, Watchman

Forum rules
This textbox is used to restore diagrams posted with the fen tag before the upgrade.
Four.nine
Full Member
Posts: 911
Joined: Tue Mar 21, 2017 7:30 pm

Post by Four.nine »

Error alert:

I am on Game 11, and I can't get the machines to cooperate on the reverse color English.
Turns out I have transposed White's moves 3 and 4 on Game 1 notation. May not mean much to anybody, but I am a stickler for accuracy.

One more note: RS doesn't understand transpositions. In English, RS will NOT play 2. .. Nf6, 3. .. Nc6 then 4. .. d4 (always plays 3. .. d4). It will however, if the line goes 2. .. Nc6, 3. .. Nf6: then RS WILL play 4. .. d4.
Four.nine
Full Member
Posts: 911
Joined: Tue Mar 21, 2017 7:30 pm

Post by Four.nine »

Game 4: Meanwhile, back in reality: ExPro smokes 2250

[Event "EPMS-2250"]
[Black "2250"]
[Date "Jan 30, 2018"]
[Result "1-0"]
[Round "4"]
[White "EPMS"]

1.e4 e5 2.Nc3 Nf6 3.Bc4 Nc6 4.d3 Bb4 5.Nf3 d6 6.O-O Bxc3 7.bxc3 Na5 8.Bb3 Bd7 9.c4 Qe7 10.c3 Nxb3 11.axb3 O-O 12.Be3 a6 13.Qe2 Bc6 14.Rfd1 Rae8 15.b4 Nd7 16.Qb2 Nf6 17.b5 axb5 18.cxb5 Bd7 19.Ra7 Ra8 20.Rxb7 Bc8 21.Ra7 Rxa7 22.Bxa7 Bb7 23.Ra1 Rd8 24.c4 Ra8 25.Be3 Ng4 26.Rxa8+ Bxa8 27.Bg5 f6 28.Bd2 f5 29.Qa2 Bb7 30.Qa7 Bc8 31.b6 fxe4 32.dxe4 Kf7 33.b7 Bxb7 34.Qxb7 Nf6 35.Kf1 h6 36.Bc3 c5 37.Qc6 g6 38.h3 Nh5 39.Qd5+ Qe6 40.Bxe5 dxe5 41.Nxe5+ Kf6 42.Ng4+ Ke7 43.Qxc5+ Ke8 44.Qd5 Ke7 45.Nxh6 Nf6 46.Qxe6+ Kxe6 47.f3 Nd7 48.Ke2 Kd6 49.Nf7+ Ke7 50.Ng5 Kd6 51.f4 Kc5 52.Kd3 Kb4 53.Ne6 Ka5 54.g4 Kb4 55.e5 Kb3 56.c5 Nb8 57.h4 Kb4 58.h5 gxh5 59.gxh5 Nc6 1-0

C28n: Vienna: 3. Bc4 Nc6 4. d3 Bb4 in book through White move 8. RS move 28. .. f5?? was the final nail. ExPro incompetence forced me to play another 30 moves to make sure. ExPro: 3-1

Game 5: A draw, until RS mistake, gives ExPro a win. The finish was...........exhausting.

[Event "2250-EPMS match"]
[Black "EPMS"]
[Date "Jan 31, 2018"]
[Result "0-1"]
[Round "5"]
[White "2250"]

1.d4 Nf6 2.c4 e6 3.Nf3 b6 4.a3 Bb7 5.Nc3 d5 6.e3 Bd6 7.Bd3 c5 8.cxd5 exd5 9.Nb5 Be7 10.O-O c4 11.Bf5 O-O 12.Re1 Re8 13.Nc3 Nc6 14.e4 dxe4 15.Bxe4 Nxe4 16.Rxe4 Na5 17.d5 Nb3 18.Rb1 b5 19.Be3 Bf6 20.Rxe8+ Qxe8 21.Nd4 Bxd4 22.Bxd4 a5 23.Qg4 Nxd4 24.Qxd4 b4 25.Ne4 Qb5 26.Rd1 Rd8 27.a4 Qe8 28.Qxc4 Qxa4 29.Qc1 Rc8 30.Qb1 Qd7 31.Ng5 f5 32.d6 a4 33.Qd3 b3 34.Nf3 Be4 35.Ne5 Bxd3 36.Nxd7 Bb5 37.Ne5 a3 38.bxa3 b2 39.d7 Bxd7 40.Nd3 Rb8 41.Rb1 Rb3 42.Ne1 g6 43.f3 f4 44.Nc2 Bf5 45.Nd4 Bxb1 46.Nxb3 Bc2 47.Nd2 g5 48.g3 Kf7 49.Kf2 Ke6 50.Ke2 fxg3 51.hxg3 h5 52.Kf2 b1=N 53.Nxb1 Bxb1 54.a4 Kd5 55.f4 g4 56.a5 Kc5 57.a6 Kb6 58.Ke2 Kxa6 59.Ke1 Bd3 60.Kf2 Kb5 61.Ke1 Kc4 62.Kd1 Kd4 63.f5 h4 64.f6 h3 65.f7 h2 66.Kd2 h1=Q 67.f8=Q Qg2+ 68.Qf2+ Qxf2+ 69.Kd1 Qe2+ 70.Kc1 Qc2# 0-1

E12m: Queen’s Indian: Petrosian in book through White move 5. A draw until RS 35. Ne5?? GAME OVER. ExPro “gits it”, but then, doesn’t git 36. .. a3! or 42. ..f4 or 52. .. Kd5. But, DOES calculate 52. .. b1=N. WTF? I made ExPro play it out, because (with the aforementioned ExPro endgame “expertise”) I still thought White had drawing chances until the very end.

Game 6: A draw, until RS mistake, gives ExPro a win. The finish was...........abysmal.

[Event "2250-EPMS match"]
[Black "2250"]
[Date "Jan 31, 2018"]
[Result "1-0"]
[Round "6"]
[White "EPMS"]

1.d4 d5 2.Nf3 e6 3.Bf4 Nf6 4.Nbd2 Nbd7 5.e3 c5 6.c3 Be7 7.Bb5 O-O 8.Rb1 Nh5 9.Bg3 a6 10.Bd3 c4 11.Bc2 b6 12.Ne5 g6 13.Nc6 Qe8 14.Nxe7+ Qxe7 15.O-O Nxg3 16.fxg3 f6 17.e4 Bb7 18.exd5 Bxd5 19.Qg4 b5 20.Rbe1 Rfd8 21.Ne4 Rab8 22.Qh4 Rf8 23.b3 Rbd8 24.Rf2 a5 25.bxc4 Bxc4 26.Bb3 Rc8 27.d5 e5 28.Bxc4 Rxc4 29.Rb2 Rb8 30.Rf2 Rb6 31.g4 Qg7 32.d6 Rxe4 33.Rxe4 Rxd6 34.Re1 Nc5 35.Rf3 Qe7 36.a3 Qd8 37.Rfe3 Rd2 38.Rf1 Qa8 39.Qg3 Qc6 40.Qf3 Qxf3 41.Rexf3 Nd7 42.Rb1 Rd5 43.Rf2 Nb6 44.Rxf6 Nc4 45.Ra1 Rd3 46.g5 a4 47.h4 Kg7 48.Kh2 e4 49.Kg1 Ne3 50.Rc1 Rd2 51.Rf2 Rxf2 52.Kxf2 Nc4 53.Ra1 Kf7 54.Kg3 e3 55.Ra2 Nd2 56.Kf4 e2 57.Ra1 Nc4 58.Re1 Nxa3 59.Rxe2 Nc4 60.Ke4 Ke6 61.Kd4+ Kd6 62.Rf2 a3 63.Rf6+ Kc7 64.Kc5 Kb7 65.Rf1 Nd2 66.Rf7+ Kb8 67.Kb6 Nc4+ 68.Ka6 Nd6 69.Rf1 Kc7 70.Ra1 Kc6 71.Rxa3 Kc5 72.Rb3 Kc4 73.Rb1 Kxc3 74.Rc1+ Kb4 75.Rc7 Ka3 76.Rd7 Nf5 77.Rxh7 b4 78.g4 Ne3 79.h5 gxh5 80.g6 Nd5 81.Rd7 Nf6 82.g5 Nxd7 83.g7 b3 84.g8=Q b2 85.Qh7 Ne5 86.Kb5 Kb3 87.Qb1 Nf7 88.Qd3+ Ka2 89.Ka4 Ka1 90.Qa3+ Kb1 91.Kb3 Kc1 92.Qxb2+ Kd1 93.g6 Nh6 94.Kc3 Ke1 95.Qd2+ Kf1 96.Qxh6 Kg2 97.g7 Kf3 98.g8=Q h4 99.Qf6+ Ke2 100.Qge6+ Kd1 101.Qf1# 1-0

D02n: Queen’s Pawn: London in book for 6 full moves. A draw until RS 39. .. Qc6?? ExPro played “Atlanta-ish” maneuvering, maneuvering, maneuvering, until an opponent’s mistake. Unfortunately, the finish was anything but Atlanta: totally clueless, brainless and inept. Idiotic 65. Rf1??????????? actually handed RS a draw. RS declined.
Four.nine
Full Member
Posts: 911
Joined: Tue Mar 21, 2017 7:30 pm

Post by Four.nine »

Game 7: A computer generated game: EPMS out of book at move 2, RS at move 3. ExPro crushes RS early on, but, the finish was...........abysmal.

[Event "2250-EPMS match"]
[Black "EPMS"]
[Date "Jan 31, 2018"]
[Result "0-1"]
[Round "7"]
[White "2250"]

1.e4 e5 2.Bc4 Bc5 3.Nf3 d6 4.O-O Nf6 5.Nc3 O-O 6.d3 Be6 7.Na4 Bxc4 8.Nxc5 dxc5 9.dxc4 Qxd1 10.Rxd1 Nc6 11.Nd2 Rad8 12.f3 Nd4 13.c3 Nc2 14.Rb1 Rd3 15.Kf2 Rfd8 16.g3 a5 17.h4 b6 18.g4 h6 19.h5 g6 20.hxg6 fxg6 21.Ke2 h5 22.gxh5 Nxh5 23.Kf2 Nf4 24.a3 Re3 25.Kg3 Re2 26.Rg1 Rexd2 27.Bxd2 Ne2+ 28.Kf2 Nxg1 29.Rd1 Rd3 30.Kxg1 Ne3 31.Re1 Nxc4 32.Bc1 Rxf3 33.Rd1 g5 34.Rd8+ Kh7 35.Rd7+ Kg6 36.Rxc7 Rd3 37.Kf2 Rd1 38.b3 Rxc1 39.bxc4 Rxc3 40.Rc6+ Kh5 41.Rxb6 Kg4 42.Rc6 Kf4 43.Rxc5 Rxa3 44.Rc8 g4 45.Rf8+ Kxe4 46.Rg8 Kf5 47.Rf8+ Kg5 48.Re8 g3+ 49.Kg2 Kf4 50.Rf8+ Ke4 51.Rg8 Kd4 52.Rxg3 Rxg3+ 53.Kxg3 0-1
C23: Bishop’s Opening: Classical in book for 4 moves, but as stated EPMS calculating at move 2 and RS at move 3! After RS 22. .. gxh5?? the game was over. Well, should have been. At move 26, ExPro has a 10 point eval lead (26. .. Ne3!), but screws it up with RxN? I made ExPro “prove” it could actually win...and that took nearly 30 moves. But, ExPro is killing RS 6-1; Match performance of +311. And I thought RS2250XL was equal to ExPro?

Game 8: Another “Atlanta” type effort by ExPro: just wait and wait for a mistake... then pounce.

[Event "EPMS-2250"]
[Black "2250"]
[Date "Feb 1, 2018"]
[Result "1-0"]
[Round "8"]
[White "EPMS"]

1.e4 e5 2.Nc3 Nf6 3.Bc4 Nc6 4.d3 Bb4 5.Nf3 d6 6.O-O Bxc3 7.bxc3 Na5 8.Bb3 Nxb3 9.cxb3 O-O 10.h3 b6 11.Bg5 Bb7 12.Nh2 h6 13.Bh4 g5 14.Bg3 a5 15.Ng4 Nxg4 16.Qxg4 Qe7 17.Rfe1 Qe6 18.Qxe6 fxe6 19.d4 exd4 20.cxd4 Rad8 21.Rac1 c6 22.a3 h5 23.h4 g4 24.e5 d5 25.Rc2 Rc8 26.Rec1 Kf7 27.Bf4 Kg6 28.Be3 Kf5 29.Kh2 Rf7 30.Kg3 c5 31.dxc5 Rfc7 32.b4 Ba6 33.b5 d4 34.Bxd4 Bxb5 35.cxb6 Rxc2 36.Rxc2 Rb8 37.f3 gxf3 38.Kxf3 Rd8 39.Rd2 Rd7 40.Ke3 Ba6 41.a4 Rd8 42.Rf2+ Kg4 43.Rf7 Kxh4 44.Rg7 Rd5 45.b7 Bxb7 46.Bc3 Rd1 47.Rxb7 Rd5 48.Rb5 Rd1 49.Rxa5 Kg3 50.Ra6 Kxg2 51.Rxe6 h4 52.Rg6+ Kh3 53.e6 Rc1 54.Bb4 Rc6 55.e7 Rxg6 56.e8=Q Rg3+ 57.Kf2 Rg2+ 58.Kf1 Kh2 1-0

C28n: Vienna: 3. Bc4 Nc6 4. d3 Bb4. Yes, a repeat of Game 4..but with a difference: RS varies at move 8, with the book move, 8. ..Nxb3. The game was completely different than Game 4 and RS had a 100% draw for a long time, UNTIL 33. .. d4??? GAME OVER. But , ExPro misses 34. bxa6 (BxP??). But RS is still in bad shape: down a pawn and fighting for survival. After RS 43. .. KxP???? GAME OVER, this time for real. ExPro up 7-1.

Game 9: This is why repeat games should not be immediately halted. Game 7 opening, but this time RS2250XL wins!

[Event "2250-EPMS match"]
[Black "EPMS"]
[Date "Feb 1, 2018"]
[Result "1-0"]
[Round "9"]
[White "2250"]

1.e4 e5 2.Bc4 Bc5 3.Nf3 d6 4.O-O Nf6 5.Nc3 O-O 6.d3 Be6 7.Bxe6 fxe6 8.Na4 Bb6 9.c3 Nbd7 10.Qb3 Qe7 11.Bg5 h6 12.Nxb6 axb6 13.Bh4 Nc5 14.Qc2 g5 15.Bg3 Nh5 16.d4 Nxg3 17.fxg3 Nd7 18.a3 Nf6 19.dxe5 dxe5 20.Kh1 Ng4 21.Qe2 Qc5 22.Rae1 Ra4 23.h3 Nf6 24.Qc2 b5 25.h4 Nh5 26.Kh2 gxh4 27.gxh4 Nf6 28.Kh1 Nh5 29.Kh2 Nf6 30.Qd2 Kg7 31.Qd3 Ng4+ 32.Kg3 Nf6 33.Qc2 Qc6 34.Kh3 Qc5 35.Re2 Rd8 36.g4 c6 37.g5 hxg5 38.Nxg5 Qe7 39.Rg2 Kh8 40.Qe2 Rd6 41.Rxf6 Rd3+ 42.Kh2 Qxf6 43.Qh5+ Kg7 44.Nxe6# 1-0

C23: Bishop’s Opening: Classical in book for 4 moves, just like Game 7: EPMS calculating at move 2 and RS at move 3! In this game RS varies (Apparently thanks to Random=ON) at 7. Bxe6 (Na4 previously) and the game changed completely. Doubled “e” pawns and a Rook out on a5 Killed ExPro. ExPro was already in serious trouble when 37. .. hxg5???????? was GAME OVER.
Interesting Note: The only computer that I have seen do this: Helpmate! At move 43, ExPro “saw” its own Mate in 4 and instantly moved 43. .. Kg7 to give White MATE IN 1!!

Game 10: A see-saw battle in an extremely complex position. Ultimately, 2250 made the last mistake. ExPro 1-0.

Event "2250-EPMS match"]
[Black "2250"]
[Date "Feb 1, 2018"]
[Result "1-0"]
[Round "10"]
[White "EPMS"]

1.d4 c6 2.e4 d5 3.Nc3 dxe4 4.Nxe4 Nd7 5.Bc4 Ngf6 6.Ng5 e6 7.Qe2 Nb6 8.Bb3 h6 9.N5f3 c5 10.Be3 cxd4 11.O-O-O Bd6 12.Bxd4 O-O 13.Nh3 Nbd5 14.Rd2 b6 15.Bxf6 Qxf6 16.c4 Qg6 17.cxd5 Qxg2 18.Rhd1 Qxh3 19.dxe6 Bf4 20.e7 Bxd2+ 21.Rxd2 Re8 22.Ba4 Ba6 23.Qxa6 Qxf3 24.Bxe8 Rxe8 25.Rd8 Qf4+ 26.Kd1 Qf3+ 27.Qe2 Qc6 28.Qd3 f5 29.Qxf5 Qa4+ 30.b3 Qc6 31.Qf8+ Kh7 32.Qxe8 Qf3+ 33.Ke1 Qe4+ 34.Kd2 Qf4+ 35.Kd3 Qf3+ 36.Kc4 Qe2+ 37.Kb4 Qe1+ 38.Kb5 Qf1+ 39.Kc6 Qg2+ 40.Kc7 Qxh2+ 41.Kb7 Qe5 42.Qg8+ Kg6 43.e8=Q+ Qxe8 44.Qxe8+ 1-0

B17s: Caro-Kann: Steinitz in book through White move 19. Both sides made mistakes, both sides had winning opportunities. 2250 grabby 23. .. QxN?? (Qc8+!) dooms Black. At move 33, ExPro must calculate a successful King “flight”. ExPro did it.

10 games and ExPro leads 8-2. I’m not surprised. I expected it.
All games were fair, no takebacks, no cheating, no forced moves, no forced openings.
How anyone expects a weaker program, running 62.5% SLOWER (or if you like, ExPro 60% faster) than an opponent computer is beyond me.
Next prediction: colors reversed: ExPro 7-3.
User avatar
spacious_mind
Senior Member
Posts: 3999
Joined: Wed Aug 01, 2007 10:20 pm
Location: Alabama
Contact:

Post by spacious_mind »

Four.nine wrote:Thank you to Nick for Brute Force blessing. Also: Maybe the uber performance of your 2250 in 2007 is the reason for the unrealistically high Elo rating continually foisted by Aktivschach.
I seriously doubt it, the ELO reflects 125 games. Mine are just a few.

Regards
Nick
User avatar
spacious_mind
Senior Member
Posts: 3999
Joined: Wed Aug 01, 2007 10:20 pm
Location: Alabama
Contact:

Post by spacious_mind »

This is a big enough sample that would reflect the rating accurately:

90 RadioShack Champion 2250XL 2058 125 (+ 43,= 33,- 49), 47.6 %

Mephisto Magellan 3 (+ 0,= 0,- 3), 0.0 %
Phoenix Resurrection Ruffian 5 (+ 0,= 1,- 4), 10.0 %
Mephisto Academy 2 (+ 0,= 1,- 1), 25.0 %
Novag Citrine 13 (+ 6,= 4,- 3), 61.5 %
Novag Diablo / Scorpio 68000 5 (+ 1,= 3,- 1), 50.0 %
Saitek GK 2000 2 (+ 1,= 1,- 0), 75.0 %
Mephisto Milano 5 (+ 4,= 1,- 0), 90.0 %
Mephisto Monte Carlo IV LE 2 (+ 1,= 0,- 1), 50.0 %
Novag Super Expert / Forte C 6 MHz 1 (+ 1,= 0,- 0), 100.0 %
Mephisto Supermondial / Monte Carlo 1 (+ 1,= 0,- 0), 100.0 %
Mephisto Supermondial II / Monte Carlo IV 2 (+ 2,= 0,- 0), 100.0 %
Fidelity Kishon Chesster 2 (+ 0,= 1,- 1), 25.0 %
Mephisto Almeria 68000 3 (+ 1,= 1,- 1), 50.0 %
Fidelity Excel 68000 2 (+ 1,= 1,- 0), 75.0 %
Mephisto MM IV 18 MHz / Revelation MM IV 1 (+ 0,= 0,- 1), 0.0 %
Fidelity Mach III / Designer 2265 / V2 3 (+ 1,= 1,- 1), 50.0 %
Fidelity Mach IV / Designer 2325 / V6 2 (+ 2,= 0,- 0), 100.0 %
Mephisto Milano Pro / Master Chess 2 (+ 0,= 0,- 2), 0.0 %
Mephisto Roma 68020 2 (+ 1,= 0,- 1), 50.0 %
Novag Sapphire / Diamond 1 (+ 0,= 1,- 0), 50.0 %
Novag Zircon II / Jade II 1 (+ 0,= 1,- 0), 50.0 %
Mephisto Polgar 2 (+ 2,= 0,- 0), 100.0 %
Saitek Brute Force 3 (+ 1,= 1,- 1), 50.0 %
Saitek Cosmos / Centurion 24 MHz 6 (+ 1,= 1,- 4), 25.0 %
Mephisto Lyon 68020 24 MHz 2 (+ 0,= 0,- 2), 0.0 %
Phoenix Revelation Polgar 1 (+ 0,= 0,- 1), 0.0 %
Saitek Risc 2500 2 (+ 0,= 1,- 1), 25.0 %
Novag Star Sapphire / Star Diamond 1 (+ 0,= 0,- 1), 0.0 %
Novag Turquoise / Emerald classic plus 6 (+ 1,= 4,- 1), 50.0 %
Saitek Cougar / Mephisto Explorer Pro 5 (+ 2,= 1,- 2), 50.0 %
Novag Super Nova 1 (+ 1,= 0,- 0), 100.0 %
Excalibur Grandmaster / Igor 2 (+ 1,= 1,- 0), 75.0 %
Novag Star Ruby / Obsidian 1 (+ 1,= 0,- 0), 100.0 %
Novag Super Expert / Forte B 6 MHz 1 (+ 0,= 0,- 1), 0.0 %
Tasc ChessMachine 32 MHz Gideon 3.0 2 (+ 0,= 0,- 2), 0.0 %
Fidelity Elite 68030 V9 1 (+ 0,= 0,- 1), 0.0 %
Mephisto Polgar 10 MHz 1 (+ 1,= 0,- 0), 100.0 %
Mephisto Portorose 68020 13 (+ 4,= 4,- 5), 46.2 %
Mephisto Schachakademie 1 (+ 0,= 0,- 1), 0.0 %
Fidelity Designer 2100 1 (+ 1,= 0,- 0), 100.0 %
Mephisto Almeria 68020 24 MHz 1 (+ 0,= 0,- 1), 0.0 %
Saitek D++ 16 MHz 1 (+ 1,= 0,- 0), 100.0 %
Mephisto Nigel Short 1 (+ 0,= 1,- 0), 50.0 %
Phoenix Revelation Amsterdam 2 (+ 0,= 0,- 2), 0.0 %
Phoenix Revelation Rebell 5.0 1 (+ 0,= 1,- 0), 50.0 %
Conchess T8 Amsterdam 1 (+ 1,= 0,- 0), 100.0 %
Excalibur Alexandra 2 (+ 2,= 0,- 0), 100.0 %
Saitek D++ 10 MHz 2 (+ 0,= 1,- 1), 25.0 %
Tasc ChessMachine 14 MHz King 2.2 1 (+ 0,= 0,- 1), 0.0 %
Saitek D+ 10 MHz 1 (+ 0,= 0,- 1), 0.0 %

The more important question is should Explorer pro games be split away from Cougar games based on what you are showing.

Regards
Nick
Four.nine
Full Member
Posts: 911
Joined: Tue Mar 21, 2017 7:30 pm

Post by Four.nine »

spacious_mind wrote:This is a big enough sample that would reflect the rating accurately:

90 RadioShack Champion 2250XL 2058 125 (+ 43,= 33,- 49), 47.6 %


The more important question is should Explorer pro games be split away from Cougar games based on what you are showing.

Regards
2058 10 MHz vs. 2056 for 16 MHz AND similar, but stronger program????
We all need to get some of what the tester was smoking.
And the 2058 is also based on Brute + random...default ????
I can assure at power up settings, the 2250 is actually very close to a stock GK2100.

Yes, ExPro needs its own rating, the tester may even dare use the Max Strength methodology.
User avatar
spacious_mind
Senior Member
Posts: 3999
Joined: Wed Aug 01, 2007 10:20 pm
Location: Alabama
Contact:

Post by spacious_mind »

Four.nine wrote:
spacious_mind wrote:This is a big enough sample that would reflect the rating accurately:

90 RadioShack Champion 2250XL 2058 125 (+ 43,= 33,- 49), 47.6 %


The more important question is should Explorer pro games be split away from Cougar games based on what you are showing.

Regards
2058 10 MHz vs. 2056 for 16 MHz AND similar, but stronger program????
We all need to get some of what the tester was smoking.
And the 2058 is also based on Brute + random...default ????
I can assure at power up settings, the 2250 is actually very close to a stock GK2100.

Yes, ExPro needs its own rating, the tester may even dare use the Max Strength methodology.
The ratings are based on the games played. Cougar/EX Pro played 358 games combined. Hey if they have a larger weighting playing against much better opponents then you are going to get knocked down some.

The ratings are just a guideline to give you a feel for who they played and not absolute. You don't have to take them as gospel, because they are not. If Carlsen played you 100 times and beats you 100-0 does that make you a 200 ELO player?

Regarding "smoking" that's pretty harsh towards all the dozens of people people that might have contributed to the 125 RS games and 360 or whatever Cougar and ExPro games... LOL

As I said my influence is about 10 games combined between all these programs so I can safely say my influence is almost zero! :wink:

Or are you asking for someone to selectively remove some games to suit? Would that be more accurate? This solution would kind of sound like selectively cheating to me :)

Best regards
Nick
Four.nine
Full Member
Posts: 911
Joined: Tue Mar 21, 2017 7:30 pm

Post by Four.nine »

spacious_mind wrote:
Four.nine wrote:
spacious_mind wrote:This is a big enough sample that would reflect the rating accurately:

90 RadioShack Champion 2250XL 2058 125 (+ 43,= 33,- 49), 47.6 %


The more important question is should Explorer pro games be split away from Cougar games based on what you are showing.

Regards
2058 10 MHz vs. 2056 for 16 MHz AND similar, but stronger program????
We all need to get some of what the tester was smoking.
And the 2058 is also based on Brute + random...default ????
I can assure at power up settings, the 2250 is actually very close to a stock GK2100.

Yes, ExPro needs its own rating, the tester may even dare use the Max Strength methodology.
The ratings are based on the games played. Cougar/EX Pro played 358 games combined. Hey if they have a larger weighting playing against much better opponents then you are going to get knocked down some.

The ratings are just a guideline to give you a feel for who they played and not absolute. You don't have to take them as gospel, because they are not. If Carlsen played you 100 times and beats you 100-0 does that make you a 200 ELO player?

Regarding "smoking" that's pretty harsh towards all dozens of people people that might have contributed to the 125 RS games and 360 or whatever Cougar and ExPro games... LOL

As I said my influence is about 10 games combined between all these programs so I can safely say my influence is almost zero! :wink:

Or are you asking for someone to selectively remove some games to suit? Would that be more accurate? This solution would kind of sound like selectively cheating to me :)

Best regards
No. I don't want any cheating. The machine speed and the numbers simply don't add up.
I throw up my hands now and finish the match.

One added note: I used the Elo calculator and found 2250 is -22 (quick add=125 games).

Maybe a curious student will add up the opponents to see if they truly average 2080.
Four.nine
Full Member
Posts: 911
Joined: Tue Mar 21, 2017 7:30 pm

Post by Four.nine »

Correction 2250 is -17
User avatar
spacious_mind
Senior Member
Posts: 3999
Joined: Wed Aug 01, 2007 10:20 pm
Location: Alabama
Contact:

Post by spacious_mind »

Four.nine wrote: No. I don't want any cheating. The machine speed and the numbers simply don't add up.
I throw up my hands now and finish the match.

One added note: I used the Elo calculator and found 2250 is -22 (quick add=125 games).

Maybe a curious student will add up the opponents to see if they truly average 2080.
Here let me show you how this happens, which is why I also take the ratings as a guideline only, with a bit of a pinch of salt as it is all based on how the computer played against the opponents it played, with loads of additional unknowns such as opening books played. Here is the example from combined Cougar/Ex Pro.

91 Saitek Cougar / Mephisto Explorer Pro 2056 359 (+133,= 77,-149), 47.8 %

Mephisto Vancouver 68000 13 (+ 4,= 2,- 7), 38.5 %
Phoenix Resurrection Ruffian 2 (+ 0,= 0,- 2), 0.0 %
Mephisto Risc 2 / CM 14 MHz Gideon 3.1 20 (+ 4,= 2,- 14), 25.0 %
Mephisto Academy 1 (+ 0,= 1,- 0), 50.0 %
Novag Citrine 8 (+ 2,= 2,- 4), 37.5 %
Novag Diablo / Scorpio 68000 4 (+ 2,= 1,- 1), 62.5 %
Saitek GK 2000 2 (+ 1,= 1,- 0), 75.0 %
Saitek GK 2100 32 (+ 14,= 7,- 11), 54.7 %
Novag Super Expert / Forte C 6 MHz 8 (+ 5,= 2,- 1), 75.0 %
Mephisto Supermondial / Monte Carlo 1 (+ 1,= 0,- 0), 100.0 %
Mephisto Supermondial II / Monte Carlo IV 2 (+ 2,= 0,- 0), 100.0 %
SciSys Turbo King II 5 (+ 5,= 0,- 0), 100.0 %
Fidelity Kishon Chesster 2 (+ 1,= 1,- 0), 75.0 %
Mephisto Rebell 5.0 10 (+ 6,= 3,- 1), 75.0 %
Mephisto Almeria 68000 3 (+ 0,= 1,- 2), 16.7 %
Mephisto Berlin 68000 14 (+ 1,= 2,- 11), 14.3 %
Mephisto MM IV 18 MHz / Revelation MM IV 1 (+ 0,= 1,- 0), 50.0 %
Mephisto MM V + HG550 7 (+ 3,= 2,- 2), 57.1 %
Fidelity Mach III / Designer 2265 / V2 9 (+ 3,= 2,- 4), 44.4 %
Fidelity Mach IV / Designer 2325 / V6 14 (+ 3,= 4,- 7), 35.7 %
Mephisto Milano Pro / Master Chess 18 (+ 1,= 8,- 9), 27.8 %
Mephisto Roma 68020 22 (+ 9,= 4,- 9), 50.0 %
Novag Sapphire / Diamond 11 (+ 2,= 3,- 6), 31.8 %

I just took the info from the top end of the list. Where it loses its points is being overall hammered by Risc 2. Berlin, Master, Diamond, Vancouver. Poor showing against Roma, Mach 4 (just shows you that best way to beat Mach IV might be to play it without book, who knows). Only 14-11 against GK2100.

As I said lots of negatives here that knock it down.

Regards
Nick
User avatar
spacious_mind
Senior Member
Posts: 3999
Joined: Wed Aug 01, 2007 10:20 pm
Location: Alabama
Contact:

Post by spacious_mind »

PS. If you take the above as example you have 69 wins 95 losses = 42. 1% which is considerably less than the 47.8% average.

Also what did the programs play? 30S, 60/30 or 30 minute game, Did the opponent play same setting or was it 60/30 against 30S or 30S against 30 minute game. What was the random factor? Which book? Selective on?

regards
Nick
Four.nine
Full Member
Posts: 911
Joined: Tue Mar 21, 2017 7:30 pm

Post by Four.nine »

Last post today...tomorrow 11-13.

Final note:

In all my 2250/Aktivschach bashing, I overlooked maybe the biggest factor:
Time.

Aktivschach is 30 sec/move...mine 1 min.
It is very likely the 2250 program was cleverly tuned to get to stronger moves faster at the expense of even stronger moves later.
I'm thinking now maybe 2058 IS correct, BUT for 30 sec/move.

So, take my match results....with the 1 min/move grain of salt.

As for harsh: the place I worked that "smoking" phrase was tossed around like "How you doin'".
No offense to the testers.
User avatar
spacious_mind
Senior Member
Posts: 3999
Joined: Wed Aug 01, 2007 10:20 pm
Location: Alabama
Contact:

Post by spacious_mind »

Four.nine wrote:Last post today...tomorrow 11-13.

Final note:

In all my 2250/Aktivschach bashing, I overlooked maybe the biggest factor:
Time.

Aktivschach is 30 sec/move...mine 1 min.
It is very likely the 2250 program was cleverly tuned to get to stronger moves faster at the expense of even stronger moves later.
I'm thinking now maybe 2058 IS correct, BUT for 30 sec/move.

So, take my match results....with the 1 min/move grain of salt.

As for harsh: the place I worked that "smoking" phrase was tossed around like "How you doin'".
No offense to the testers.
No one is taking offence...its why I put smiley behind it. I see you as being passionate... which is great for the hobby :)

ps I placed the games I listed through the beginner rating and Cougar/EXP showed 2044 ELO. But I am not that convinced that the beginner rating as shown at Schachcomputer.Info works 100% correctly. Maybe it does...but not sure.

Best regards
Nick
Reinfeld
Member
Posts: 486
Joined: Thu Feb 17, 2011 3:54 am
Location: Tacoma, WA

Post by Reinfeld »

Nick said,
The more important question is should Explorer pro games be split away from Cougar games based on what you are showing.
Four.nine said (in the Cougar v ExPro thread):
Ostensibly, Cougar and ExPro are the exact same program....until:
UNTIL, some point in the endgame, when new code was introduced making the ExPro program “considerably” stronger than Cougar. I’ll leave the definition of “considerably” to a curious student.
The results of the Cougar v ExPro match certainly show a big advantage for ExPro. So despite the near-identical guts and feature sets (when the "secret options" for ExPro are invoked), the argument that they're separate machines at some level deserves deeper scrutiny.

- R.
"You have, let us say, a promising politician, a rising artist that you wish to destroy. Dagger or bomb are archaic and unreliable - but teach him, inoculate him with chess."
– H.G. Wells
Four.nine
Full Member
Posts: 911
Joined: Tue Mar 21, 2017 7:30 pm

Post by Four.nine »

Game 11: Reversed Colors Game 1. ExPro varies very quickly and just sits back and waits for a mistake. ExPro 1-0.

[Event "EPMS-2250 default match"]
[Black "2250"]
[Date "Feb 2, 2018"]
[Result "1-0"]
[Round "11"]
[White "EPMS"]

1.c4 c5 2.Nc3 Nf6 3.g3 Nc6 4.Nf3 d5 5.cxd5 Nxd5 6.Bg2 e6 7.O-O Be7 8.Nxd5 exd5 9.d4 cxd4 10.Nxd4 O-O 11.Be3 Bf6 12.Rc1 Ne7 13.Qd2 Bg4 14.Nb5 Qd7 15.Nc7 Rac8 16.Bf4 Be6 17.Rfd1 Rfd8 18.Nxe6 fxe6 19.Bg5 Rxc1 20.Rxc1 Bxg5 21.Qxg5 Nc6 22.b3 Rc8 23.a4 Rf8 24.Rc3 b6 25.Qd2 Qd6 26.Rc2 a6 27.Rc1 Rc8 28.Bh3 Rf8 29.Qc3 Nd8 30.Bg2 Nf7 31.b4 e5 32.Qd2 d4 33.Rc6 Qb8 34.Qd3 a5 35.Qb5 axb4 36.Rxb6 Qc8 37.Bd5 Kh8 38.Bxf7 Qf5 39.f4 Qxf7 40.Qxe5 b3 41.Qxd4 Qg8 42.Qb4 Re8 43.e4 h6 44.e5 Ra8 45.Kf2 h5 46.h4 Qd8 47.Qxb3 Kh7 48.Qc2+ Kg8 49.Rb4 Ra7 50.Ke3 Qe8 51.Qc5 Qa8 52.Qd6 Kh7 53.Qd3+ g6 54.f5 Qe8 55.e6 Re7 56.Rb5 Kh6 57.Qd4 Rg7 58.Re5 gxf5 59.Qf4+ Kh7 60.Qxf5+ Kg8 61.Qxh5 Qxh5 62.Rxh5 Rxg3+ 63.Kf4 Ra3 64.Ra5 Kg7 65.Ra7+ Kf6 66.e7 Kf7 67.h5 Rh3 68.Kg5 Rg3+ 69.Kh4 Rg1 70.h6 Rh1+ 71.Kg5 Rg1+ 72.Kf4 Rf1+ 73.Kg3 Rg1+ 74.Kf2 Rg6 75.h7 Rf6+ 76.Ke3 Rh6 77.Ra8 Kxe7 78.h8=Q Rxh8 79.Rxh8 Kd6 1-0

A36d: English Symmetrical in book through White move 12, the full line shown by SmallFish! At move 8, ExPro CALCULATED the book move 8. Nxd5 (2250 calculated Qa4 as White). Commendable was that both programs calculated the full book line from about move 7 or 8 on!! Basically a DRAW for the first 31 moves; ExPro has been playing the “waiting” game and at 32. .. d4? a small error by RS (SF says e4). After RS 35. .. axb4? Black is in trouble. ExPro 38. Bxf7 should be GAME OVER, but the moron makes life very difficult with 39. f4? WTF are you doing? Now, ExPro turns total moron and takes 40 more inane moves to mate. A good win for ExPro, but the finish, like so often, was abysmal.

Game 12: Reversed Colors Game 2. A flatline draw...the iPad called it at move 78.

[Event "EP-2250 match"]
[Black "EPMS"]
[Date "Jan 30, 2018"]
[Result "1/2-1/2"]
[Round "12"]
[White "2250"]

1.e4 e6 2.d4 d5 3.Nd2 Nf6 4.e5 Nfd7 5.Bd3 c5 6.c3 Nc6 7.Ne2 cxd4 8.cxd4 Qb6 9.Nf3 f6 10.exf6 Nxf6 11.O-O Bd6 12.Qb3 O-O 13.Re1 e5 14.Nxe5 Nxe5 15.dxe5 Bxe5 16.h3 Bd7 17.Nc3 Qxb3 18.axb3 Rae8 19.Re2 a6 20.Bg5 Bc6 21.Rd1 Rf7 22.Red2 h6 23.Bxf6 Rxf6 24.Bc2 d4 25.Ne2 Bb5 26.Be4 Bxe2 27.Rxe2 b5 28.Rde1 Rd8 29.Bb7 Bc7 30.Re7 Bf4 31.Bf3 d3 32.Rd1 Kf8 33.Ra7 Be5 34.Rd2 Re6 35.Kf1 Bf4 36.Rd1 Red6 37.Be4 d2 38.Ke2 Bg5 39.Rc7 Rd4 40.f3 Rb4 41.Bc2 Re8+ 42.Kf2 Re7 43.Rxe7 Kxe7 44.g3 a5 45.f4 Bf6 46.Kf3 Bxb2 47.Rxd2 h5 48.Rd5 h4 49.gxh4 a4 50.bxa4 bxa4 51.Ra5 a3 52.Ra6 Bc1 53.Be4 Rb3+ 54.Kg4 Bb2 55.Bg2 Re3 56.Kf5 Rg3 57.Bf1 Bf6 58.h5 Bd4 59.h4 Rc3 60.Bg2 Rc5+ 61.Kg4 Rc3 62.Bd5 Rd3 63.Ba2 Re3 64.Kf5 Rh3 65.h6 gxh6 66.Re6+ Kf8 67.Rxh6 Rh2 68.Rd6 Bg7 69.Bc4 a2 70.Ke6 Bc3 71.Rd8+ Kg7 72.Bxa2 Rxa2 73.Rd3 Ra6+ 74.Kf5 Ba1 75.Kg4 Kg6 76.Rd8 Bg7 77.Rd7 Rb6 78.Rd1 Rc6 1/2-1/2

C06o: French: Tarrasch in book for 11 full moves. RS left book with 12. Qb3 (Game 2 went book to White move 13). Didn’t matter, the game was a total draw: First Draw! At move 78, the iPad “locked up”. Even the iPad had had enough. ExPro Max Strength leads 9.5-2.5 (+232 Elo performance). I predicted 15-5; ExPro is running slightly ahead of that.

Game 13: Reversed Colors Game 3. Drawn after ExPro missed a small chance

[Event "2250-EPMS match"]
[Black "2250"]
[Date "Feb 3, 2018"]
[Result "1/2-1/2"]
[Round "13"]
[White "EPMS"]

1.d4 d5 2.g3 Nd7 3.Bg2 Ngf6 4.Nf3 e6 5.O-O c5 6.c4 dxc4 7.Na3 Be7 8.Nxc4 O-O 9.dxc5 Bxc5 10.Bg5 Qe7 11.Rc1 Rd8 12.Qb3 a6 13.Nce5 Bd6 14.Nxd7 Bxd7 15.Rfd1 Bc6 16.Qc3 Rac8 17.Bxf6 Qxf6 18.Qxf6 gxf6 19.Nd4 Bxg2 20.Rxc8 Rxc8 21.Kxg2 Rd8 22.Nb3 b6 23.a4 f5 24.Kf3 Kg7 25.a5 Be7 26.Ra1 Rb8 27.Nd4 bxa5 28.Nc6 Rb3+ 29.e3 Bf6 30.Rxa5 Rb6 31.Nd4 Kg6 32.Kg2 e5 33.Nf3 e4 34.Nh4+ Bxh4 35.gxh4 f4 36.exf4 Rxb2 37.Kg3 Rb6 38.Kg4 Rc6 39.h5+ Kg7 40.h3 Kh6 41.Ra4 Re6 42.f5 Rb6 43.h4 Kg7 44.Kf4 Rc6 45.Rxe4 h6 46.Re8 Rb6 47.Ra8 Kf6 48.f3 Rb4+ 49.Kg3 Rb6 50.Kf4 Rc6 51.Kg4 Rd6 1/2-1/2

D00a: Queen’s Pawn 2.g3 in book through White move 7. Unlike Game 3, ExPro varied with book move 4. 0-0 (Nf3 previously by RS). Well, apparently RS knows something, because going book only got ExPro a draw (RS won in Game 3). ExPro may have had the slightest advantage with 13. Be3 (Be3 BxB NxB Rb8 Rd1 Nf8 Qb6 RxR RxR Bd7 Qa7 Qd8 Ne5), but that is StockFish talking. RS actually wins the D00a home and home 1W 1D, but ExPro has reached the 10 point mark in 13 games (10-3).
Post Reply