Final Rating Test attempt.

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

Final Rating Test attempt.

Post by spacious_mind »

I have recently been trying to come up with a set of test positions with the hope of being able to rate and compare from the lowest rated dedicated chess computers to the highest.

My first attempt was to create a set of positions similar to the BT-Tests, which at first seemed to work quite well until I tried out CXG 3000. This computer caused me to grind to a stop because it had real difficulty in the test, scoring very low. Well below it's actual game rating. When looking at it's results a little closer it made me realize that there are many dedicated computers who may not be able to always find a very best move, however an alternative second best move might be sufficient to win. It is exactly this dilemma that made me pause and realize that it would be almost impossible to find a good set of positions which would suit the play characteristics of every dedicated computer and rate every dedicated computer fairly.

Not wanting to give up so easily, I then continued looking through my chess book collection to see if there is anything there that could provide me with some ideas, when I happened to open "How good is your chess?" by Leonard Barden. This book rates your skill by you participating in 35 Master Games and rates you on your chosen move with a scoring system. Your total points can then be compared to a chart which tells you if you are a beginner, amateur, master or a grandmaster.

So anyway I tried some of the games in this book with some dedicated computers and this seemed to work very well from a perspective that these dedicated computers could play along very well from weak computer to strong computer. In these Tests the CXG 3000 performed again at a more expected skill level. Since this test is more of a test under Match Play conditions, I also liked the fact that with this kind of test I found it easier to get a feel of a dedicated thought process and planning or lack of planning depending on which direction the move choices were going.

However, I also quickly realized that since this book and several contemporary books of this "rate yourself" idea were written in the 40's, 50's and 60s, well before our computer generation, that a lot of the evaluations and scoring were incorrect as well as not having enough acceptable move choices provided with a score. In some of these books it was either "guess the correct move or get nothing" which is just not an acceptable way to rate a human or a computer in my humble opinion. Actually this really is the same basic flaw that BT-Tests and other positional tests also have as well, coming to think of it.

So not wanting to give up completely, since I hate giving up on something, I then thought why not continue with the Barden Test positions but use modern technology instead to create an evaluation criteria. So I started thinking about a set of goals and rules that I wanted to achieve.

1) I will not pick and chose a game = staying unbiased - I continued with Barden (hence not picked by me) since these positions he picked back in 1957.
2) The test has to be verifiable by anyone who wants to verify it = use an engine that is easily available for anyone.
3) The engine has to be very strong = I chose Critter 1.6a because it is freely available and strong
4) The engine and the computer program has to have the ability to search multi-variations simultaneously and be exportable to a spreadsheet and be available for everyone = I chose Arena.
5) The search depth has to be deep enough to ensure that the test can be used by any chess program at setting of 30 seconds per move or 3 minutes per move. = I chose 21 ply deep and in some positions 22 ply deep.
6) Create a scoring system which calculated the score and not a system where points are set based on opinion.

So with these goals in mind I then restarted the test following the above rules.

Image

The above table shows how I am calculating the points. I started with 10 choices evaluated in game 1. From game 2 onwards I have increased it to 15 choices. So in game 1 I might go back sometime in the future and re-run it with 15 to align game 1 with the rest of the test games.

Image

The above table shows the results of game 1. The programs shown are not all the programs since the spreadsheet is wider than what I can show here.

You can see in the top left corner a score of 2424 below the word FEN. This is the evaluated strength of this game played by Botvinnik against Grob in 1956 per the evaluation from Critter 1.6a and my point rating system.

In other words Mikhail Botvinnik himself made 64.7 points out of a maximum of 80 points with his move choices = 80.88% or a Nick rating of ELO 2424 for this particular game.

Since I used Critter and on my computer being an AMD 2 Core laptop 21-22 ply took about 15 to 30 minutes per move to calculate through those ply's, I can be safe in saying that my max rating is going to be 3000 ELO for these tests. And to test this I then ran Critter 1.6a at 30 seconds per move and it performed with a score of 74.8 points out of 80 = 92.25% or Nick's ELO of 2768.

As a note Botvinnik's highest ELO Rating was ELO 2885 in 1945 and Henry Grob's highest rating was in 1939 at ELO 2491. So my rating of this game at ELO 2424 I think is acceptable because overall this game was below par for both players I am sure.

So since I now have the ability to compare the Human, the engine and the dedicated computers as well as DOS programs etc here is how these computers would have faired against Botvinnik if he had played at this sub par strength against them. You can also kind of say that a computer performance above ELO 2424 in this Test would also have most likely beaten Henry Grob in this particular game.

Test Game 1 Rankings

1. Critter 1.6a 64 Bit - AMD Phenom 2 Core 2.8GHZ = 2768 ELO
2. Radioshack 2250XL Brute Force = 2453 ELO
3. Mikhail Botvinnik = 2426 ELO
4. Saitek Travel Champion 2100 = 2333 ELO
5. Tasc CM 512K – 15 MHZ – KING 2.54 =
6. Radioshack 2250XL Selective = 2306 ELO
7. Mephisto TM Vancouver 68030 36 MHz = 2288 ELO
8 . Saitek Corona = 2269 ELO
9 . Mephisto Nigel Short = 2261 ELO
10. Mchess Pro 5 - P75 = 2209 ELO
11. Novag Jade 2 = 2138 ELO
12. CXG 3000 = 1800 ELO
13. Palm Chess Tiger 14.9 33MHz = 1703 ELO
14. Saitek MK 12 Trainer LV 5 90S/Move = 1594 ELO
15. Fidelity Sensory 9 = 1553 ELO
16. Novag Constellation JR = 1470 ELO
17 Saitek MK 12 Trainer LV 4 15S/Move = 1361 ELO

Now please remember these ratings are based on comparing how the computer may have played the above game and therefore is a comparison of its performance against a human in this particular game.

Please don't try to compare the ELO against existing computer ratings because I think these are apples and oranges in a comparison.

I will try and post Test Game 2 results very soon.

http://spacious-mind.com/forum_reports/ ... _final.ods

The above link allows you to download the spreadsheet for the 1st 4 tests and the currently played results. I have prepared a couple columns at the of the spreadsheet columns for you, in case you would like to test one of your computers. Please do not alter the scoring system or any of the formulas as this will mess it up. Just enter the move played by your computer into the row where its move fits and its corresponding score. The spreadsheet will calculate the rest for you.

Also if you do not see the move listed as a scoring choice then please enter the move into the red row with 0.

There is still room for improvement..ie use the average of 3 or 4 top engines for the evaluation or narrow down the scoring system. But all this can be done later once all the 16 tests are established first.

I hope you enjoy,

ps. you could say that RadioShack Brute Force may have given Botvinnik a harder game than Henry Grob did in this particular game :)

best regards,

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

Post by spacious_mind »

I forgot to post the actual PGN game for Test Game 1. You can copy and paste this into your PGN player:

Test Game 1

1. Nf3 Nf6 2. c4 d5 3. cxd5 Nxd5 4. e4 Nf6 5. Nc3 e6 6. d4 c5 {Test Start}

[fen]rnbqkb1r/pp3ppp/4pn2/2p5/3PP3/2N2N2/PP3PPP/R1BQKB1R w KQkq c6 0 7[/fen]

[Event "Leonard Barden Test Game 1"]
[Site "Zurich 1956"]
[Date "????.??.??"]
[Round "?"]
[White "Botvinnik, M."]
[Black "Grob, H."]
[Result "1-0"]
[ECO "D06"]
[Annotator ",Microsoft"]
[PlyCount "59"]
[EventDate "1956.??.??"]

1. Nf3 Nf6 2. c4 d5 3. cxd5 Nxd5 4. e4 Nf6 5. Nc3 e6 6. d4 c5 {Test Start} 7.
d5 {7. d5 2.5; 7. Bf4 3.0; 7. Be2 2.3; 7. Bb5+ 1.8; 7. Be3 1.1; 7. a3 1. 1; 7.
Bg5 1.1; 7. e5 1.0} a6 8. Bg5 {8. Bg5 1.6; 8. Bf4 3.0; 8. d6 2.0; 8. Bd3 1.9;
8. a4 1.8; 8. Qa4+ 1.1; 8. g3 1.0} Qb6 9. Bxf6 {9. Bxf6 1.0; 9. Bc4 3.0; 9.
Bd3 2.6; 9. Qa4+ 2.2; 9. Qc2 1.9; 9. Qd2 1.7; 9. Qe2 1.7; 9. Be2 1.6; 9. Rc1 1.
2; 9. Ne5 1.2} gxf6 10. Qd2 {10. Qd2 2.6; 10. Qc2 4.0; 10. a3 3.1; 10. Rb1 2.2;
10. Qc1 1.6; 10. Qe2 1.1; 10. Nd2 1.1; 10. Qb1 1.0} h5 11. Be2 {11. Be2 3.0;
11. Bc4 3.0; 11. Rd1 3.0; 11. a4 2.4; 11. Qc2 1.6; 11. Bd3 1.0; 11. g3 1.0} Nd7
12. O-O {
12. 0-0 3.0; 12. Rd1 2.4; 12. dxe6 1.7; 12. a4 1.3; 12. Qc2 1.1; 12. h4 1.0} h4
13. a4 {13. a4 1.1; 13. Rad1 3.0; 13, Rfd1 2.8; 13, h3 2.7; 13. Rfe1 2.0; 13.
dxe6 1.4; 13. a3 1.2; 13. Qc2 1.0} Bh6 14. Qc2 {14. Qc2 2.0; 14. Qd3 1.2} Bf4
15. a5 {15. a5 1.4; 15. dxe6 2.0; 15. Rfd1 1.3; 15. h3 1.0} Qc7 16. Rfd1 {
16. Rfd1 1.2; 16. e5 3.0; 16. h3 1.4; 16. Na4 1.2; 16. dxe6 1.0} Ne5 17. Nxe5 {
17. Nxe5 1.6; 17. h3 3.0; 17. Na4 2.7; 17. Nd2 1.2; 17. g3 1.0} Bxe5 18. h3 {
18. h3 3.0; 18. Kh1 3.0; 18. Na4 2.4; 18. dxe6 1.0} Bd7 19. Na4 {19. Na4 3.0}
Bxa4 20. Rxa4 {20. Rxa4 2.1; 20. Qxa4 3.0} Rc8 21. Rc4 {21. Rc4 4.0; 21. Bg4 2.
3; 21. b3 2.1; 21. Ra3 2.0; 21. dxe6 1.9; 21. Kf1 1.5; 21. Qd2 1.5; 21. Bc4 1.
4; 21. Qc1 1.4; 21. Qb3 1.0} Qxa5 22. b4 {
22. b4 4.0; 22. dxe6 2.1; 22. Qc1 1.0; 22. Bg4 1.0} Qa3 23. Rxc5 {23. Rxc5 4.4;
23. dxe6 5.0; 23. bxc5 1.8; 23. Bg4 1.6; 23. Qd2 1.5; 23. d6 1.3; 23. Kh1 1.0;
23. Kf1 1.0; 23. Rd2 1.0} Rd8 24. dxe6 {24. dxe6 4.0} Bd6 25. Rxd6 {25. Rxd6 3.
2; 25. exf7+ 5.0; 25. e5 4.7; 25. Rc8 2.3; 25. Rh5 2.2; 25. Rc4 2.0; 25. Bh5 2.
0; 25. Rcd5 1.9; 25. Rc3 1.3; 25. Rf5 1. 0} Rxd6 26. Rc8+ {
26. Rc8+ 4.0; 26. exf7+ 2.0} Ke7 27. Qc7+ {
27. Qc7+ 2.0; 27. Rc7+ 3.0; 27. Rxh8 1.0} Kxe6 28. Bg4+ {
28. Bg4+ 4.0; 28. Rxh8 4.0; 28. Bc4+ 2.0; 28. Qc4+ 1.0} f5 29. Bxf5+ {
29. Bxf5+ 3.0; 29.exf5+ 2.0} Ke5 30. Qc5+ {
30. Qc5+ 3.0; 30. Qe7+ 3.0; 30. f4+ 2.0; 30. Rxc8 1.0} 1-0

Best 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 »

Test Game 2

The second Test Game is match played between A. Mangini against Alexander Kotov played in Mar Del Plata, Argentina in 1957.

Kotov's highest ELO rating was recorded in 1950 at 2753 ELO. In this game Kotov was evaluated at ELO 2629. Which means he played this game at Grandmaster strength but still likely had some reserves left in him. I could not find any ratings for A. Mangini.

[fen]r1bqkb1r/pp2pppp/2np1n2/8/3NP3/2PB4/PP3PPP/RNBQK2R w KQkq - 0 6[/fen]

1. e4 c5 2. Nf3 d6 3. d4 cxd4 4. Nxd4 Nf6 5. Bd3 Nc6 6. c3 {Start Test}

[Event "Leonard Barden Test Game 2"]
[Site "Mar del Plata, 1957"]
[Date "????.??.??"]
[Round "?"]
[White "Mangini, A."]
[Black "Kotov, A."]
[Result "0-1"]
[ECO "B55"]
[Annotator ",Microsoft"]
[PlyCount "56"]
[EventDate "1957.??.??"]

1. e4 c5 2. Nf3 d6 3. d4 cxd4 4. Nxd4 Nf6 5. Bd3 Nc6 6. c3 {Start Test} e5 {
2.5 = 6. /\ e5; 3.0 = 6. /\ g6; 2.6 = 6. /\ e6; 1.7 = 6. /\ Bd7; 1.7 = 6. /\
Ne5; 1.7 = 6. /\ Qc7; 1.5 = 6. /\ h6; 1.5 = 6. /\ Rb8; 1.3 = 6. /\ a6; 1.2 = 6.
/\ Qd7; 1.2 = 6. /\ Bg4; 1.2 = 6. /\ Qb6; 1.1 = 6. /\ Qa5; 1.0 = 6. /\d5} 7.
Nc2 d5 {4.0 = 7. /\ d5; 1.7 = 7. /\ Bg4; 1.3 = 7. /\ Be7; 1.0 = 7. /\ Be6} 8.
exd5 Nxd5 {3.0 = 8. /\ Nxd5; 2.2 = 8. /\ Qxe5; 1.0 = 8. /\ e4} 9. O-O Be6 {
4.0 = 9. /\ Be6; 2.8 = 9. /\ Be7; 2.8 = 9. /\ Bc5; 1.9 = 9. /\ h5; 1.8 = 9. /\
Nf6; 1.5 = 9. /\ a6; 1.0 = 9. /\h6} 10. Qf3 Bd6 {1.9 = 10. /\ Bd6; 4. 0 0. /\
Qd7; 3.9 = 10. /\ Qf6; 3.1 = 10. /\ Nf6; 2.7 = 10. /\ a6; 2.5 = 10. /\ Qc8; 2.
5 = 10. /\ Rc7; 2.5 = 10. /\ Rc8; 2.5 = 10. /\ Bc5; 2.1 = 10. /\ Be7; 1.9 = 10.
/\ Nb6; 1.6 = 10. /\ Qh4; 1.4 = 10. /\ h5; 1.3 = 10. /\ g6; 1.0 = 10. /\ Qb6}
11. Ne3 Nxe3 {
4.0 = 11. /\ Nxe3; 1.7 = 11. /\ Nf4; 1.0 = 11. /\ e4; 1.0 11. /\ Qf6} 12. Bxe3
O-O {4.0 = 12. /\ 0-0; 4.0 = 12. /\ f5; 1.1 = 12. /\ Qc7; 1.0 = 12. / \ Qd7; 1.
0 = 12. /\ Qe7} 13. Nd2 f5 {4.0 = 13. /\ f5; 2.4 = 13. /\ Be7; 1. 4 = 13. /\
Qc8; 1.3 = 13. /\ Qh4; 1.0 = 13. /\ Ba3} 14. Bc4 Qe7 {4.0 = 14. /\ Qe7; 3.6 =
14. /\ Qd7; 3.1 = 14. /\ Qf6; 2.6 = 14. /\ Qe8; 2.0 = 14. /\ Bf7; 2.0 = 14. /\
Qc8; 1.0 = 14. /\ Bxc4} 15. Bxe6+ Qxe6 {1.0 = 15. /\ Qxe6} 16. b4 e4 {
4.0 = 16. /\ e4; 3.1 = 16. /\ Rac8; 3.0 = 16. /\ a5; 2.7 = 16. / \ Bc7; 2.0 =
16. /\ Be7; 1.9 = 16. /\ Rad8; 1.7 = 16. /\ Rfe8; 1.7 = 16. / \ Rfd8; 1.6 = 16.
/\ h6; 1.6 = 16. /\ Rfc8; 1.5 = 16. /\ Rae8; 1.4 = 16. / \ Rf6; 1.3 = 16. /\
a6; 1.1 = 16. /\ Kh8; 1.0 = 16. /\Rf7} 17. Qe2 Qe5 {1.0 = 17. /\ Qe5; 4.0 = 17.
/\ Be5; 3.3 = 17. /\ Kh8; 3.0 = 17. /\ Ne5; 2. 6 = 17. /\ Rfe8; 2.6 = 17. /\
Rae8; 2.5 = 17. /\ Rac8; 2.2 = 17. /\ Ne7; 1. 8 = 17. /\ Rfd8; 1.5 = 17. /\
Rf7; 1.5 = 17. /\ Bc7; 1.3 = 17. /\ Rfc8; 1. 3 = 17. /\b6} 18. Qc4+ Kh8 {
2.0 = 18. /\ Kh8; 1.0 = 18. /\ Rf7} 19. g3 Rac8 {4.0 = 19. /\ Rac8; 2.9 = 19. /
\ Qf6; 1.5 = 19. /\ Qe7; 1.5 = 19. /\ f4; 1. 0 = 19. /\ Qe8} 20. Rfd1 Qf6 {
4.0 = 20. /\ Qf6; 2.0 = 20. /\ Qe7; 1.5 = 20. /\ Bb8; 1.2 = 20. /\ b5; 1.0 =
20. /\ Be7} 21. Qb3 Be5 {4.0 = 21. /\ Be5; 3. 6 = 21. /\ b6; 1.6 = 21. /\ Bc7;
1.5 = 21. /\ Rcd8; 1.2 = 21. /\ Ne5; 1.0 = 21. /\ Rfe8; 1.0 = 21. /\ Bb8} 22.
Nc4 f4 {4.0 = 22. /\ f4; 1.7 = 22. /\ b6; 1.0 = 22. /\ b5} 23. Bc5 e3 {
1.7 = 23. /\ e3; 4.0 = 23. /\ Na5; 2.4 = 23. /\ Rfe8; 1.0 = 23. /\ Qf5} 24.
Bxf8 exf2+ {4.0 = 24. /\ exf2+; 3.8 = 24. /\ fxg3} 25. Kg2 Rxf8 {
3.0 = 25. /\ Rxf8; 1.3 = 25. /\ fxg3} 26. Rd3 fxg3 {
3.0 = 26. /\ fxg3; 2.5 = 26. /\ f3} 27. Nd2 gxh2 {2.0 = 27. /\ gxh2} 28. Nf1
Qg6+ {1.0 = 28. /\ Qg6+; 3.0 = 28. /\ h1Q+; 3.0 = 28. /\ h1B+; 2.0 = 28. / \
h1R+; 1.0 = 28. /\ Qg5+; 1.0 = 28. /\ Qh4} 0-1


Test Game 2 Results


Image

Test Game 2 Rankings

Code: Select all

1 Critter 1.6a 64 Bit - AMD Phenom 2 Core 2.8GHZ	75.4	94.25%	2828
2 Herr Alexander Kotov	70.0	87.50%	2629
3 Novag Jade 2	57.6	72.00%	2160
4 Saitek Travel Champion 2100	53.0	66.25%	1988
5 Mephisto TM Vancouver 68030 36 MHz	52.3	65.38%	1961
6 Radioshack 2250XL Selective	49.5	61.88%	1856
7 Saitek Corona	48.6	60.75%	1823
8 MChess Pro 5 – P75	46.7	58.38%	1751
9 Tasc CM 512K – 15 MHZ – KING 2.54	45.7	57.13%	1714
10 Radioshack 2250XL Brute Force	44.0	55.00%	1650
11 CXG 3000	41.8	52.25%	1568
12 Fidelity Sensory 9	38.9	48.63%	1459
13 Novag Constellation JR	35.0	43.75%	1313
14 Saitek MK 12 Trainer LV 5 90S/Move	28.0	35.00%	1050
15 Saitek MK 12 Trainer LV 4 15S/Move	28.0	35.00%	1050
Well other than of course the engine Critter 1.6a, no one else could get close to the quality of performance displayed by Alexander Kotov in this game. Definitely Alexander Kotov would have easily beaten all the tested programs (except for Critter 1.6a)

Tasc King 2.54 was extremely bad in this game. This game did not suit a King. Best performance and only program showing consistency is David Kittinger's Novag Jade II. All the better programs were well below par. They obviously did not like and could not understand the positions arising in this game.

Average Ranking and Rating after 2 Games

Code: Select all

1	Critter 1.6a 64 Bit - AMD Phenom 2 Core 2.8GHZ	149.2	93.25%	2798
2	Grandmaster Performance Level 	134.7	84.19%	2528
3	Saitek Travel Champion 2100	115.7	72.31%	2169
4	Novag Jade 2	114.6	71.63%	2149
5	Mephisto TM Vancouver 68030 36 MHz	113.3	70.81%	2124
6	Radioshack 2250XL Selective	111	69.38%	2081
7	Radioshack 2250XL Brute Force	109.4	68.38%	2051
8	Saitek Corona	109.1	68.19%	2046
9	Tasc CM 512K – 15 MHZ – KING 2.54	107.9	67.44%	2023
10	MChess Pro 5 - P75	105.6	66.00%	1980
11	CXG 3000	89.8	56.13%	1684
12	Fidelity Sensory 9	80.3	50.19%	1506
13	Novag Constellation JR	74.2	46.38%	1391
14	Saitek MK 12 Trainer LV 5 90S/Move	70.5	44.06%	1322
15	Saitek MK 12 Trainer LV 4 15S/Move	64.3	40.19%	1206
Best Regards,

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

Post by spacious_mind »

Test Game 3

The third Test Game is match played between Arnold Yorvath Green against the author of the book whose games are being used in this Test Suite, Leonard Barden. The game was played in Bognor, England in 1956. As far as I can recall Bognor should really be named Bognor Regis which is a seaside town on the south coast of England.

I believe both are International Masters. Arnold Yorvath Green's highest ELO rating of ELO 2356 was established in 1952. Leonard Bardens highest ELO rating was yet to come in 1958 with a ratng of ELO 2497. This game was evaluated at ELO 2456 which fits well into Leonard Bardens chess strength rating.

[fen]r1bq1rk1/pp2ppbp/3p1np1/n2P4/8/2N3P1/PP1NPPBP/R1BQ1RK1 w - - 0 10[/fen]

1. d4 Nf6 2. c4 g6 3. g3 Bg7 4. Bg2 O-O 5. Nf3 d6 6. O-O Nc6 7. d5 Na5 8. Nfd2
c6 9. Nc3 cxd5 10. cxd5 {Test Start}

[Event "Leonard Barden Test Game 3"]
[Site "Bognor, 1956"]
[Date "????.??.??"]
[Round "?"]
[White "Green, A. Y."]
[Black "Barden, L."]
[Result "0-1"]
[ECO "E62"]
[Annotator ",Microsoft"]
[PlyCount "60"]
[EventDate "1956.??.??"]

1. d4 Nf6 2. c4 g6 3. g3 Bg7 4. Bg2 O-O 5. Nf3 d6 6. O-O Nc6 7. d5 Na5 8. Nfd2
c6 9. Nc3 cxd5 10. cxd5 {Test Start} Ng4 {1.9 = 10. /\ Ng4; 3.5 = 10. /\ Bd7;
2.9 = 10. /\ Bg4; 2.9 = 10. /\ Bf5; 2.8 = 10. /\ Qc7; 2.1 = 10. /\ a6; 1. 9 =
10. /\ Nh5; 1.8 = 10. /\ Ne8; 1.6 = 10. /\ Qb6; 1.0 = 10. /\ Nd7} 11. Rb1 Bd7 {
3.3 = 11. /\ Bd7; 3.5 = 11. /\ Bf5; 2.4 = 11. /\ Rb8; 2.3 = 11. / \ Ne5; 2.2 =
11. /\ a6; 1.7 = 11. /\ h6; 1.7 = 11. /\ Qd7; 1.4 = 11. /\ Qb6; 1.3 = 11. /\
Bxc3; 1.3 = 11. /\ Qe8; 1.3 = 11. /\ h5; 1.2 = 11. /\ b6; 1.0 = 11. /\ Re8; 1.
0 = 11. /\} 12. h3 Ne5 {3.5 = 12. /\ Ne5; 2.5 = 12. /\ Nf6; 1.0 = 12. /\ Nh6}
13. b3 Qc8 {3.5 = 13. /\ Qc8; 1.5 = 13. /\ b5; 1.0 = 13. /\ Rc8} 14. b4 Nac4 {
3.5 = 14. /\ Nac4; 3.2 = 14. /\ Bxh3; 3.1 = 14. /\ Qxc3; 1.1 = 14. /\ b6; 1.0
= 14. / \ Rb8} 15. Kh2 Nxd2 {2.5 = 15. /\ Nxd2; 4.0 = 15. /\ Ne3; 2.3 = 15. /\
b5; 2.2 = 15. /\ Nb6; 1.4 = 15. /\ Qc7; 1.4 = 15. /\ a6; 1.4 = 15. /\ a5; 1.4
= 15. /\ f5; 1.2 = 15. /\ Ng4+; 1.1 = 15. /\ Bf5; 1.0 = 15. /\ Qe8} 16. Qxd2
Bf5 {4.0 = 16. /\ Bf5; 2.9 = 16. /\ b5; 2.4 = 16. /\ a5; 2.3 = 16. /\ Nc4; 2.2
= 16. /\ b6; 2.2 = 16. /\ Qc7; 2.1 = 16. /\ Re8; 2.0 = 16. /\ a6; 1.7 = 16. /\
Rd8; 1.6 = 16. /\ h6; 1.4 = 16. /\ Qc4; 1.2 = 16. /\ Rb8; 1.2 = 16. /\ Qd8; 1.
0 = 16. /\ Qe8; 1.0 = 16. /\ h5} 17. Ne4 Qc4 {4.0 = 17. /\ Qc4; 1.9 = 17. /\
b5; 1,9 = 17. /\ Qd7; 1.6 = 17. /\ b6; 1.6 = 17. /\ Re8; 1.3 = 17. /\ Qc7; 1.3
= 17. /\ Nc4; 1.3 = 17. /\ h5; 1.2 = 17. /\ a6; 1.2 = 17. /\ Qd8; 1.1 = 17. /\
Rd8; 1.1 = 17. /\ Rb8; 1.1 = 17. /\ Qe8; 1.0 = 17. /\ Qb8} 18. f3 Rfc8 {
4.0 = 18. /\ Rfc8; 3.8 = 18. /\ Rac8; 3.5 = 18. /\ Rfe8; 3.1 = 18. /\ a6; 2.7
= 18. /\ Rfd8; 1.9 = 18. /\ Rab8; 1.8 = 18. /\ Rfb8; 1.3 = 18. /\ h5; 1.3 = 18.
/\ Bxe4; 1.0 = 18. /\ b6; 1.0 = 18. /\ h6; 1.0 = 18. /\b5} 19. Rd1 Qc2 {
3.2 = 19. /\ Qc2; 4.0 = 19. /\ Bxe4; 2.9 = 19. /\ Rc7; 2.2 = 19. /\ h6; 2.2 =
19. /\ a6; 2.0 = 19. /\ b6; 1.8 = 19. / \ b5; 1.4 = 19. /\ Bh8; 1.3 = 19. /\
a5; 1.2 = 19. /\ h5; 1.0 = 19. /\Rab8} 20. Rb2 Qa4 {
2.4 = 20. /\ Qa4; 4.0 = 20. /\ Qxd2; 1.0 = 20. /\ Qc7} 21. Qe1 Rc7 {1.1 = 21. /
\ Rc7; 4.0 = 21. /\ Nc4; 3.0 = 21. /\ a6; 2.3 = 21. /\ b6; 2.0 = 21. /\ Rc4; 1.
5 = 21. /\ Bxe4; 1.3 = 21. /\ Nxf3+; 1.3 = 21. / \ h6; 1.2 = 21. /\ h5; 1.2 =
21. /\ Nd7; 1.0 = 21. /\ b5} 22. Nd2 Bc2 {3.8 = 22. /\ Bc2; 4.0 = 22. /\ a5; 2.
6 = 22. /\ Rc2; 1.2 = 22. /\ Nc4; 1. 2 = 22. /\ Rac8; 1.0 = 22. /\Bd3} 23. Nb3
Bxd1 {4.0 = 23. /\ Bxd1; 3.4 = 23. /\ Nxf3+; 2.5 = 23. /\ Rac8; 1.3 = 23. /\
Ng4+; 1.0 = 23. /\ Nd3; 1.0 = 23. /\ Nd7} 24. Qxd1 Rac8 {4.0 = 24. ...Rac8; 3.
1 = 24. /\ Qxb4; 2.0 = 24. /\ Rd8; 1.8 = 24. /\ Re8; 1.8 = 24. /\ Rf8; 1.8 =
24. /\ Rb8; 1.4 = 24. /\ Nc4; 1.4 = 24. /\ a6; 1.3 = 24. /\ h6; 1.2 = 24. /\
h5; 1.0 = 24. /\ Nd7} 25. Rd2 Qxb4 {3.0 = 25. /\ Qxb4; 4.0 = 25. /\ Nc4; 1.7 =
25. /\ Qb5; 1.6 = 25. /\ Bh6; 1.5 = 25. /\ Nd7; 1.4 = 25. /\ h5; 1.2 = 25. /\
a6; 1.1 = 25. / \ a5; 1.0 = 25. /\ h6} 26. h4 a5 {4.0 = 26. /\ a5; 1.8 = 26. /
\ h5; 1.7 = 26. /\ Qb6; 1.6 = 26. /\ b6; 1.4 = 26. /\ b5; 1.3 = 26. /\ f5; 1.2
= 26. / \ Qb5; 1.2 = 26. /\ Bh8; 1.1 = 26. /\ Nc4; 1.1 = 26. /\ Rc4; 1.0 = 26.
/\ a6; 1.0 = 26. /\ Qa4} 27. a3 Qb6 {
3.5 = 27. /\ Qb6; 2.6 = 27. /\ Qb5; 1.7 = 27. /\ Qa4; 1.0 = 27. /\ Qc4} 28. a4
Rc3 {3.5 = 28. /\ Rc3; 3.3 = 28. /\ Bh6; 1.0 = 28. /\ h5} 29. Bh3 R8c7 {
1.1 = 29. /\ R8c7; 3.5 = 29. /\ Qf2+; 1.5 = 29. /\ R8c4; 1.0 = 29. /\ f5} 30.
f4 Qxb3 {1.0 = 30. /\ Qxb3; 3.5 = 30. /\ Qf2+; 1.5 = 30. /\ Rxb3; 1.5 = 30. /\
Rxc1; 1.3 = 30. /\ Ng4+; 1.2 = 30. /\ Qe3; 1.0 = 30. /\ f5} 0-1



Test Game 3 Results


Image

Test Game 3 Rankings

Code: Select all

1	Critter 1.6a 64 Bit - AMD Phenom 2 Core 2.8GHZ	74.1	92.63%	2779
2	Tasc CM 512K – 15 MHZ – KING 2.54	71.1	88.88%	2666
3	Leonard Barden	65.3	81.63%	2449
4	MChess Pro 5 - P75	63.8	79.25%	2393
5	Mephisto TM Vancouver 68030 36 MHz	63.1	78.88%	2366
6	Radioshack 2250XL Brute Force	62.3	77.88%	2336
7	Saitek Travel Champion 2100	61.8	77.25%	2318
8	Fidelity Sensory 9	59.6	74.50%	2235
9	Radioshack 2250XL Selective	58.5	73.13%	2194
10	Novag Jade 2	57.5	71.88%	2156
11	Saitek Corona	56.3	70.38%	2111
12	Saitek MK 12 Trainer LV 4 15S/Zug	48.4	60.50%	1815
13	CXG 3000	46.6	58.25%	1758
14	Saitek MK 12 Trainer LV 5 90S/Zug	44	55.00%	1650
15	Novag Constellation JR	40.7	50.88%	1526
All the chess computers seemed to like this game. However the supreme ruler this time is King 2.54 with a Grandmaster level performance of ELO 2666. King 2.54 surely would have beaten both Green and Barden had they dared play this game against the King. This makes up for the amateurish performance in Test Game 2.


Average Ranking and Rating after 3 Games

Code: Select all

1	Critter 1.6a 64 Bit - AMD Phenom 2 Core 2.8GHZ	223.3	92.63%	2791
2	Grandmaster Performance Standard	199.7	83.21%	2496
3	Tasc CM 512K – 15 MHZ – KING 2.54	179	74.58%	2238
4	Saitek Travel Champion 2100	177.5	73.96%	2219
5	Mephisto TM Vancouver 68030 36 MHz	176.4	73.50%	2205
6	Novag Jade 2	172.1	71.71%	2151
7	Radioshack 2250XL Brute Force	171.7	71.54%	2146
8	Radioshack 2250XL Selective	169.5	70.63%	2119
9	MChess Pro 5 - P75	169.4	70.58%	2118
10	Saitek Corona	165.4	68.92%	2068
11	Fidelity Sensory 9	139.9	58.29%	1749
12	CXG 3000	136.4	56.83%	1705
13	Novag Constellation JR	114.9	47.88%	1436
14	Saitek MK 12 Trainer LV 5 90S/Zug	114.5	47.71%	1431
15	Saitek MK 12 Trainer LV 4 15S/Zug 	112.7	46.96%	1409
King 2.54 jumps back into the top spot for dedicated chess computers, 258 ELO points below a Grandmaster standard.

best regards,

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

Post by spacious_mind »

Test Game 4

The fourth Test Game was played between Wolfgang Unzicker and Luís Sánchez in Saltsjobaden, Sweden 1952. Wolfgang Unzicker had a high rating of ELO 2686 in 1960 when he was ranked # 17 in the World. Luís Sánchez had a high rating of ELO 2557 in 1958 when he was ranked 96th in the World. This game is rated at ELO 2621 which again fits perfectly into the strength categories of these two players.

[fen]r1r3k1/2qbbppp/p2p1n2/n3p3/Pp1PP3/5N1P/1PB2PP1/R1BQRNK1 w - - 0 16[/fen]

1. e4 e5 2. Nf3 Nc6 3. Bb5 a6 4. Ba4 Nf6 5. O-O Be7 6. Re1 b5 7. Bb3 d6 8. c3
Na5 9. Bc2 c5 10. d4 Qc7 11. a4 b4 12. cxb4 cxb4 13. h3 O-O 14. Nbd2 Bd7 15.
Nf1 Rfc8 {Test Start}

[Event "Leonard Barden Test Game 4"]
[Site "Saltsjobaden, 1952"]
[Date "????.??.??"]
[Round "?"]
[White "Unzicker, W."]
[Black "Sanchez, M."]
[Result "1-0"]
[ECO "C90"]
[Annotator ",Microsoft"]
[PlyCount "69"]
[EventDate "1952.??.??"]

1. e4 e5 2. Nf3 Nc6 3. Bb5 a6 4. Ba4 Nf6 5. O-O Be7 6. Re1 b5 7. Bb3 d6 8. c3
Na5 9. Bc2 c5 10. d4 Qc7 11. a4 b4 12. cxb4 cxb4 13. h3 O-O 14. Nbd2 Bd7 15.
Nf1 Rfc8 {Test Start} 16. Ne3 {
16. Ne3 = 4.0; 16. Bd3 = 3.4; 16. Re2 = 2.3; 16. Bb3 = 1.0} Nc6 17. Bb3 {
17. Bb3 = 2.6; 17. Nd5 = 4.0; 17. dxe5 = 3.6; 17. Bb1 = 3.4; 17. d5 = 3. 0; 17.
Re2 = 2.8; 17. Nf5 = 1.9; 17. Bd3 = 1.6; 17. Kh1 = 1.0; 17. b3 = 1. 0; 17. a5
= 1.0} Na5 18. Bd5 {18. Bd5 = 4.0; 18. d4e5 = 3.5; 18. Ba2 = 3.2; 18. Nd5 = 3.
0; 18. Bd2 = 1.6; 18. Ng5 = 1.0; 18. Bc2 = 1.0} Bc6 19. Bxc6 {
19. Bxc6 = 2.1; 19, Nf5 = 4.0; 19, dxe5 = 3.3; 19. Bd2 = 1.0} Nxc6 20. Nf5 {
20. Nf5 = 4.0; 20. Rb1 = 2.5; 20. Bd2 = 2.5; 20. b3 = 2.4; 20. dxe5 = 2. 3; 20.
a5 = 2.1; 20. Re2 = 2.0; 20. Qd3 = 1.8; 20. Qe2 = 1.8; 20. d5 = 1. 8; 20. Nf1
= 1.5; 20. h4 = 1.4; 20. g3 = 1.4; 20. Qb3 = 1.2; 20. Nc2 = 1.0} Bf8 21. Bg5 {
21. Bg5 = 4.0; 21. dxe5 = 3.3; 21. Bd2 = 2.7; 21. Rb1 = 2. 6; 21. a5 = 2.4; 21.
b3 = 2.3; 21. Re2 = 1.8; 21. Qd3 = 1.7; 21. Qd2 = 1. 7; 21. Kh1 = 1.6; 21. Ng3
= 1.2; 21. h4 = 1.2; 21. d5 = 1.2; 21. Qb3 = 1. 2; 21. Qe2 = 1.0} Ne8 22. Rc1 {
22. Rc1 = 2.9; 22. dxe5 = 4.0; 22. Ne3 = 3.6; 22. Rb1 = 3.4; 22. Qd3 = 3. 4;
22. Qd2 = 3.1; 22. Qb3 = 2.9; 22. Bh4 = 2.9; 22. h4 = 2.4; 22. Re2 = 2. 4; 22.
Be3 = 2.1; 22. b3 = 1.7; 22. a5 = 1.4; 22. Rf1 = 1.4; 22. Kh1 = 1.0} Qb7 23.
Qd3 {23. Qd3 = 3.3; 23. d5 = 4.0; 23. Qd2 = 3.6; 23. b3 = 2.9; 23. Rc2 = 2.7;
23. Qb3 = 2.4; 23. Kh1 = 2.4; 23. Qe2 = 2.4; 23. h4 = 2.3; 23. Kh2 = 2.3; 23.
Rb1 = 2.0; 23. dxe5 = 1.6; 23. Rf1 = 1.6; 23. Be3 = 1. 0; 23. Kf1 = 1.0} Rc7
24. Ne3 {24. Ne3 = 3.0; 24, dxe5 = 4.0; 24. d5 = 3.7; 24. Red1 = 2.0; 24. h4 =
1.8; 24. Rcd1 = 1. 7; 24. Kh2 = 1.7; 24. Kh1 = 1.6; 24. Ref1 = 1.5; 24. Bd2 =
1.4; 24. Nd2 = 1.4; 24. Rc4 = 1.3; 24. b3 = 1.1; 24. Bh4 = 1.0; 24. Rc2 = 1.0}
Rac8 25. Nd5 {25. Nd5 4.0} Rd7 26. Bd2 {26. Bd2 = 2. 1; 26. a5 = 4.0; 26. Red1
= 2.4; 26. Qc4 = 1.6; 26. dxe5 = 1.3; 26. Qc2 = 1.3; 26. Qe2 = 1.3; 26. Qd2 =
1. 3; 26. Qb3 = 1.2; 26. Be3 = 1.0; 26. Rc4 = 1.0} g6 27. Rc4 {27. Rc4 = 4.0;
27. a5 = 3.7; 27. b3 = 2.8; 27. h4 = 2.5; 27. Rc2 = 2. 3; 27. Be3 = 2.2; 27.
Red1 = 1.7; 27. Ref1 = 1.7; 27. Qf1 = 1.5; 27. dxe5 = 1. 5; 27. Qe3 = 1.3; 27.
Kh1 = 1.2; 27. Kh2 = 1.2; 27. Qb1 = 1.1; 27. Bg5 = 1.0} exd4 28. Bxb4 {
28. Bxb4 = 1.9; 28. Rec1 = 4.0; 28. Nxd4 = 3.3; 28. b3 = 2.8; 28. Nxb4 = 1.6;
28. Red1 = 1.6; 28. Qe2 = 1.5; 28. Rea1 = 1.5; 28. Qb1 = 1.5; 28. Bg5 = 1.4;
28. Qf1 = 1.4; 28. Reb1 = 1.4; 28. Ref1 = 1.4; 28. a5 = 1.0} Bg7 29. Bd2 {
29. Bd2 = 4.0; 29. Ba3 = 2.1; 29. Nxd4 = 1.3; 29. a5 = 1.0} Rb8 30. Rec1 {
30. Rec1 = 4.0; 30. b3 = 1.9; 30. Rb1 = 1.8; 30. Qc2 = 1.4; 30. g3 = 1.4; 30.
h4 = 1.4; 30. Rd1 = 1.3; 30. g4 = 1.1; 30. Ra1 = 1.0; 30. a5 = 1.0} Ne5 31.
Nxe5 {31. Nxe5 = 4.0} dxe5 32. Rb4 {32. Rb4 = 4.0} Qa7 33. Rxb8 {
33. Rxb8 = 4.0; 33. Rc6 = 2.6; 33. a5 = 1.9; 33. b3 = 1.0} Qxb8 34. Qxa6 {
34. Qxa6 = 4.0; 34. Rc6 = 2.5; 34. Qc4 = 1.2; 34. Nb4 = 1.2; 34. Rc4 = 1.0} Rb7
35. Rc8 {35. Rc8 = 4.0; 35. Qc6 = 1.4; 35. b4 = 1.1; 35. a5 = 1.1; 35. Rc6 = 1.
0; 35. f3 = 1.0} 1-0


Test Game 4 Results


Image

Test Game 4 Rankings

Code: Select all

1	Critter 1.6a 64 Bit - AMD Phenom 2 Core 2.8GHZ	76.2	95.25%	2858
2	Grandmaster Performance Standard	69.5	86.88%	2621
3	Tasc CM 512K – 15 MHZ – KING 2.54	61.4	76.75%	2303
4	Saitek Travel Champion 2100	58.7	73.78%	2201
5	Radioshack 2250XL Selective	58.7	73.78%	2201
6	Radioshack 2250XL Brute Force	56.6	70.63%	2119
7	Mephisto TM Vancouver 68030 36 Mhz	56.3	70.38%	2111
8	Novag Jade 2	53.1	66.38%	1991
9	Mchess Pro 5 – P75	52.9	66.13%	1984
10	Saitek MK 12 Trainer LV 5 90S/Move	52.4	65.50%	1965
11	Saitek Corona	51.6	64.50%	1935
12	CXG 3000	51.6	64.50%	1935
13	Saitek MK 12 Trainer LV 4 15S/Move	51.4	64.25%	1928
14	Fidelity Sensory 9	46.7	58.38%	1751
15	Novag Constellation JR	42.3	52.88%	1586
King 2.54 again performed the best, with the Morsch's doing well again as well. Still slightly disappointed with TM Vancouver and MChess Pro 5.

Perhaps they will do better in later Tests.

Average Ranking and Rating after 4 Games

Code: Select all

1	Critter 1.6a 64 Bit - AMD Phenom 2 Core 2.8GHZ	299.5	93.59%	2808
2	Grandmaster Performance Standard	269.2	84.13%	2524
3	Tasc CM 512K – 15 MHZ – KING 2.54	240.4	75.13%	2254
4	Saitek Travel Champion 2100	236.2	73.81%	2214
5	Mephisto TM Vancouver 68030 36 Mhz	232.7	72.72%	2182
6	Radioshack 2250XL Selective	228.2	71.31%	2139
7	Radioshack 2250XL Brute Force	228.2	71.31%	2139
8	Novag Jade 2	225.2	70.38%	2111
9	Mchess Pro 5 – P75	222.3	69.47%	2084
10	Saitek Corona	217	67.81%	2034
11	CXG 3000	188	58.75%	1763
12	Fidelity Sensory 9	185.6	58.31%	1749
13	Saitek MK 12 Trainer LV 5 90S/Move	166.9	52.16%	1565
14	Saitek MK 12 Trainer LV 4 15S/Move	164.1	51.28%	1538
15	Novag Constellation JR	157.2	49.13%	1474
That is all the Tests that I have completed so far.

Moving on to Test Game 5 regards,

Nick
Nick
User avatar
Theo
Member
Posts: 132
Joined: Tue Mar 05, 2013 11:34 am

Nigel Short result GAME 2

Post by Theo »

Hi Nick,

here's the results for Nigel Short, GAME 2:

6.. e5 2,5
7.. Be7 1,3
8.. Nxd5 2,2
9.. Be6 4,0
10..Bc5 2,5
11..Nxe3 4,0
12..0-0 4,0
13..f5 4,0
14..Qd7 3,6
15..Qxe6 1,0
16..f4 0,0
17..f4 0,0
18..Kh8 2,0
19..Rac8 4,0
20..b5 1,2
21..Rcd8 1,5
22..Bxc3 0,0
23..Bxc3 0,0
24..exf2+ 4,0
25..Txf8 3,0
26..fxg3 3,0
27..Qh6 0,0
28..h1Q+ 3,0 M07 0:17

50,8 points

So far I think it is a very nice test, which needs fine tuning of the points/rating system. What is the ultimate goal of this test, to predict the strenth of a computer against humans?

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

Post by spacious_mind »

Hi Theo,

Below is where Mephisto Nigel Short is placed. It performed 4th best in this Test Game 2.

Test Game 2 Rankings

Code: Select all

1 Critter 1.6a 64 Bit - AMD Phenom 2 Core 2.8GHZ	75.4	94.25%	2828
2 Mr Alexander Kotov	70.0	87.50%	2629
3 Novag Jade 2	57.6	72.00%	2160
4 Saitek Travel Champion 2100	53.0	66.25%	1988
5 Mephisto TM Vancouver 68030 36 MHz	52.3	65.38%	1961
6 Mephisto Nigel Short	51.6	64.50%	1935
7 Radioshack 2250XL Selective	49.5	61.88%	1856
8 Saitek Corona	48.6	60.75%	1823
9 MChess Pro 5 – P75	46.7	58.38%	1751
10 Tasc CM 512K – 15 MHZ – KING 2.54	45.7	57.13%	1714
11 Radioshack 2250XL Brute Force	44.0	55.00%	1650
12 CXG 3000	41.8	52.25%	1568
13 Fidelity Sensory 9	38.9	48.63%	1459
14 Novag Constellation JR	35.0	43.75%	1313
15 Saitek MK 12 Trainer LV 5 90S/Move	28.0	35.00%	1050
16 Saitek MK 12 Trainer LV 4 15S/Move	28.0	35.00%	1050
And Nigel Shorts average after 2 games placing is 4th. Which is not bad really for a 6502 with 5 Mhz.

Average Ranking and Rating after 2 Games

Code: Select all

1	Critter 1.6a 64 Bit - AMD Phenom 2 Core 2.8GHZ	149.2	93.25%	2798
2	Grandmaster Performance Level 	134.7	84.19%	2528
3	Saitek Travel Champion 2100	115.7	72.31%	2169
4	Novag Jade 2	114.6	71.63%	2149
5	Mephisto TM Vancouver 68030 36 MHz	113.3	70.81%	2124
6 	Mephisto Nigel Short	111.9	69.94%	2098
7	Radioshack 2250XL Selective	111	69.38%	2081
8	Radioshack 2250XL Brute Force	109.4	68.38%	2051
9	Saitek Corona	109.1	68.19%	2046
10	Tasc CM 512K – 15 MHZ – KING 2.54	107.9	67.44%	2023
11	MChess Pro 5 - P75	105.6	66.00%	1980
12	CXG 3000	89.8	56.13%	1684
13	Fidelity Sensory 9	80.3	50.19%	1506
14	Novag Constellation JR	74.2	46.38%	1391
15	Saitek MK 12 Trainer LV 5 90S/Move	70.5	44.06%	1322
16	Saitek MK 12 Trainer LV 4 15S/Move	64.3	40.19%	1206
You asked what I am trying to achieve? Nothing sensational really. If you look back into the chess computer history, you read about situations where it was documented that a GM lost to a dedicated computer. For Example was it not Anatoli Karpov that lost to Mephisto Almeria? Well you can ask yourself why? How could this happen? When you consider that Karpov had a rating of around 2800 and Almeria a rating less than 2100. And these situations are only the ones that where recorded. Who knows how often a Karpov, Kasparov or someone else lost to a much lower rated computer and quickly turned and looked around sheepishly to make sure that no one noticed, before switching the computer off quickly !! :P

Well I am now looking at these games and using the evaluation criteria I can really see that it really is possible on occasion under the right situation. Just look at King 2.54 in game 3. It understood that particular game much better than the IM's that played the game.

But then on the flip side you can see that King 2.54 is also capable of losing to a U1800 player from time to time. That is a swing of about 800+ ELO. I think that it is things like this that make the old programs so interesting. They are fantastic one game and horrible the next. That sounds almost human like don't you think? Even CC9 has this great swing, it shows it can scare a 2000 ELO player but also lose to a 1200 ELO player. And I think if you search back through history you will find some of these scares posted in old tournaments across USA and the world.

Well these tests show us great examples where the old computers find great moves and then follow up with horrible moves. Look at game 2 closely. Every computer picks f4 which is not a good move it is drawish at best. Every computer picks Bxc3. The first time it is picked it is again drawish at best if the queen is sacrificed (if I remember correctly) after Rd6 or something. The second time it is played it loses a winning game for Black. Sorry I am currently multitasking on the computer with a lot of background programs running so I may be not remembering exactly correct. But in any case these moves are not good and every one from King 2.54 to Constellation Jr make the same mistakes. In the last move if you read the Leonard Barden book, he did not even see 28.h1Q+, he only gave Qg6+ a score of 2 Points. The computers of course see that 28.h1Q+ is better and faster. On move 22 Barden actually gives 22. … Bxc3 two points? Well that's the benefit of being able to use high speed computers top programs to analyze. They had to do all this back then over a chess board. It is amazing really how good they were at this without any assistance and just their chess knowledge.

You would rarely notice any of this without a test like this that allows you to line up a Constellation JR and King 2.54 ;)

Also I really think this is as close as we will be able to show how computers really could and would perform in comparison to humans. I picked 16 games because as you probably know, you need around a minimum of 16 games to get an ELO rating established so hopefully after 16 games there will be a preliminary comparison between GM performance and computers tested. So yes I do think this might give us a much better way to compare humans and computers.

Computer against computer play does one thing consistently well. It pushes the weaker ones down to such a low rating level that in the end you lose the real reality of where they really stand against human players.

Anyway don't you think it is fun to be able to play a Test Game in about 1/2 Hour ?? :)

Best regards and thanks for trying out the games!!!!!
Nick
Reinfeld
Member
Posts: 486
Joined: Thu Feb 17, 2011 3:54 am
Location: Tacoma, WA

Post by Reinfeld »

For Nick's spreadsheet, here's the record of actual moves for RS 2150L, RS 2200X and Explorer Pro in the game 4 test (Unzicker v Sanchez):

RS 2150L

16. Ne3
17. d5
18. Ng5
19. Bxc6
20. Nf5
21. Bg5
22. d5
23. Qd2
24. d5
25. Nd5
26. Qc4
27. Qc4
28. Nxb4
29. Qd2
30. Rec1
31. Nxe5
32. Rb4
33.Rc6
34. b4
35. Rc8


RS 2200X

16. Ne3
17. d5
18. Ng5
19. dxe5
20. d5
21. Bg5
22. Rc1
23. dxe5
24. d5
25. Nd5
26. a5
27. a5
28. Nxb4
29. Ba3
30. Qa3
31. Nxe5
32. Rb4
33. Rxb8
34. Qxa6
35. Rc8

Explorer Pro

16. Ne3
17. Nd5
18. Bc2
19. Qd3
20. d5
21. Bg5
22. Rc1
23. dxe5
24. d5
25. Nd5
26. a5
27. a5
28. b3
29. Rd1
30. Qe2
31. Nxe5
32. Rb4
33. Rxb8
34. Qxa6
35. Rc8


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

Post by spacious_mind »

Test Game 5

In the 5th Test Game, Julio Bolbochán plays Black against Gregorio Idigoras Guisasola. This game was played in Saltsjobaden, Sweden in 1956. Julio Bolbochán's highest rating was posted in 1954 at ELO 2703 when he was ranked 14th in the World. Idigoras had a high rating a year after this Test Game was played in 1957 of ELO 2451 when he was ranked #273 in the World.

This game was evaluated at ELO 2640 which again perfectly fits into the strength of Grandmaster Julio Bolbochán. I think it is 5 out of 5 hits on the Ratings up to now.

[fen]1r1q1rk1/3bppbp/p1pp1np1/n7/2PN4/1P4P1/PB1NPPBP/1R1Q1RK1 w - - 0 13[/fen]

1. d4 Nf6 2. c4 g6 3. g3 Bg7 4. Bg2 O-O 5. Nf3 d6 6. O-O Nc6 7. b3 Rb8 8. Bb2
a6 9. d5 Na5 10. Nd4 Bd7 11. Nd2 c5 12. dxc6 bxc6 13. Rb1 {Start test}

[Event "Leonard Barden Test Game 5"]
[Site "Saltsjobaden, 1956"]
[Date "????.??.??"]
[Round "?"]
[White "Idigoras, G."]
[Black "Bolbochan, J."]
[Result "0-1"]
[ECO "E60"]
[Annotator ",Microsoft"]
[PlyCount "81"]
[EventDate "1956.??.??"]

1. d4 Nf6 2. c4 g6 3. g3 Bg7 4. Bg2 O-O 5. Nf3 d6 6. O-O Nc6 7. b3 Rb8 8. Bb2
a6 9. d5 Na5 10. Nd4 Bd7 11. Nd2 c5 12. dxc6 bxc6 13. Rb1 {Start test} c5 {
2.0 = 13. /\ c5; 1,.7 = 13. /\ Re8; 1,.7 = 13. /\ Rc8; 1,.7 = 13. /\ Qc7;
1,.7 = 13. /\ Ng4; 1,.7 = 13. /\ Qc8; 1.5 = 13. /\ e5; 1.5 = 13. /\ Qb6; 1.
4 = 13. /\ h6; 1.3 = 13. /\ h5; 1.3 = 13. /\ Ne8; 1.2 = 13. /\ Bh8; 1.1 =
13. /\ Bh6; 1.1 = 13. /\ Nh5; 1.0 = 13. /\Qe8} 14. Nc2 Qc8 {1.6 = 14. /\
Qc8; 3.0 = 14. /\ Qc7; 2.5 = 14. /\ Nc6; 2.2 = 14. /\ Qc6; 1.9 = 14. /\ h5;
1.7 = 14. /\ Qe8; 1.6 = 14. /\ Bc8; 1.5 = 14. /\ Re8; 1.4 = 14. /\ Bf5; 1.
2 = 14. /\ Kh8; 1.2 = 14. /\ Be8; 1.2 = 14. /\ Rc8; 1.2 = 14. /\ Be6; 1.2
= 14. /\ Bh8; 1.0 = 14. /\ Rb6} 15. Ne3 Nc6 {3.0 = 15. /\ Nc6; 3.0 = 15. /\
Bh3; 3.0 = 15. /\ Qd8; 2.9 = 15. /\ e6; 2.8 = 15. /\ Re8; 2.2 = 15. /\ Ne8;
2.2 = 15. /\ Nh5; 1.9 = 15. /\ Ng4; 1.8 = 15. /\ Qe8; 1.5 = 15. /\ Bc6; 1.
5 = 15. /\ Qc7; 1.2 = 15. /\ h6; 1.2 = 15. /\ Bh8; 1.1 = 15. /\ Rd8; 1.0
= 15. /\ Kh8} 16. Nd5 Nxd5 {3.0 = 16. /\ Nxd5; 2.4 = 16. /\ Nh5; 2.1 = 16. /
\ Ne8; 1.9 = 16. /\ Ng4; 1.7 = 16. /\ Bf5; 1.0 = 16. /\ Qd8; 1.0 = 16. /\
Bh3} 17. cxd5 Ne5 {3.0 = 17. /\ Ne5; 2.6 = 17. /\ Nd4; 1.8 = 17. /\ Na7; 1.
7 = 17. /\ Nd8; 1.0 = 17. /\ Na5} 18. Qc2 Bh3 {2.2 = 18. /\ Bh3; 3.0 = 18.
/\ a5; 2.7 = 18. /\ Qc7; 2,7 = 18. /\ Bf6; 2.6 = 18. /\ Qb7; 2.5 = 18. /\
Bb5; 2.0 = 18. /\ Re8; 1.9 = 18. /\ h5; 1.8 = 18. /\ Rd8; 1.8 = 18. /\ Ng4;
1.8 = 18. /\ f5; 1.2 = 18. /\ Bb4; 1.2 = 18. /\ Kh8; 1.1 = 18. /\ Rb7; 1.
0 = 18. /\ Bh8} 19. Nc4 Bxg2 {3.0 = 19. /\ Bxg2; 3.0 = 19. /\ Nxc4; 1.4 =
19. /\ Qf8; 1.4 = 19. /\ f6; 1.0 = 19. /\ Bf5} 20. Kxg2 Nxc4 {3.0 = 20. /\
Nxc4; 1.8 = 20. /\ Nd7; 1.5 = 20. /\ Qd7; 1.4 = 20. /\ f6; 1.3 = 20. /\
Rd8; 1.0 = 20. /\ Ng4; 1.0 = 20. /\ Qc7} 21. bxc4 Bxb2 {3.0 = 21. /\ Bxb2;
3.0 = 21. /\ Rxb2; 2.7 = 21. /\ Qd7; 2.0 = 21. /\ Qg4; 1.5 = 21. /\ Qf5; 1.
4 = 21. /\ Bh6; 1.4 = 21. /\ Bf6; 1.4 = 21. /\ Qc7; 1.3 = 21. /\ e5; 1.1
= 21. /\ f6; 1.1 = 21. /\ Re8; 1.0 = 21. /\ h5} 22. Rxb2 Rxb2 {3.0 = 22. /\
Rxb2; 1.7 = 22. /\ h5; 1.6 = 22. /\ Re8; 1.6 = 22. /\ Qc7; 1.5 = 22. /\
Qd7; 1.5 = 22. /\ Qd8; 1.5 = 22. /\ f6; 1.4 = 22. /\ Rb4; 1.3 = 22. /\ Qg4;
1.3 = 22. /\ Rd8; 1.2 = 22. /\ h6; 1.2 = 22. /\ f5; 1.1 = 22. /\ Rb7; 1.1
= 22. /\ Qf5; 1.0 = 22. /\ a5} 23. Qxb2 Qg4 {3.0 = 23. /\ Qg4; 2.9 = 23. /\
Qd7; 2.7 = 23. /\ Qe8; 2.5 = 23. /\ h5; 2.5 = 23. /\ Re8; 2.2 = 23. /\ e6;
2.0 = 23. /\ Qa8; 1.9 = 23. /\ Qf5; 1.6 = 23. /\ a5; 1.4 = 23. /\ f6; 1.1
= 23. /\ h6; 1.1 = 23. /\ Qd8; 1.0 = 23. /\ Rd8; 1.0 = 23. /\ Qc7; 1.0 =
23. /\ Qb8} 24. Qc2 Rb8 {
3.0 = 24. /\ Rb8; 1.4 = 24. /\ h5; 1.2 = 24. /\ Ra8; 1.0 = 24. /\ Re8} 25.
e4 a5 {1.7 = 25. /\ a5; 3.0 = 25. /\ Qg5; 1.7 = 25. /\ Kg7; 1.7 = 25. /\
Rb4; 1.7 = 25. /\ e5; 1.7 = 25. /\ Rb7; 1.7 = 25. /\ h5; 1.7 = 25. /\ Qh5;
1.7 = 25. /\ Qc8; 1.7 = 25. /\ Qd7; 1.4 = 25. /\ Kf8; 1.2 = 25. /\ h6; 1.
2 = 25. /\ Rb6; 1.1 = 25. /\ f6; 1.0 = 25. /\ Kh8} 26. h3 Qh5 {
2.1 = 26. /\ Qh5; 3.0 = 26. /\ Qg5; 2.6 = 26. /\ Qd7; 1.0 = 26. /\ Qc8} 27.
f3 Qe5 {3.0 = 27. /\ Qe5} 28. Rb1 Rxb1 {2.0 = 28. /\ Rxb1} 29. Qxb1 a4 {
1.2 = 29. /\ a4; 3.0 = 29. /\ Qc3; 1.2 = 29. /\ Kf8; 1.1 = 29. /\ h5; 1.1
= 29. /\ Kg7; 1.1 = 29. /\ Qd4; 1.1 = 29. /\ Qf6; 1.1 = 29. /\ h6; 1.0 =
29. /\ Kh8} 30. Qc2 a3 {2.0 = 30. /\ a3; 1.8 = 30. /\ Kg7; 1.8 = 30. /\ Kf8;
1.8 = 30. /\ h5; 1.5 = 30. /\ Qd4; 1.5 = 30. /\ Qh8; 1.4 = 30. /\ h6; 1.3
= 30. /\ Qf6; 1.3 = 30. /\ Qg7; 1.3 = 30. /\ Qa1; 1.3 = 30. /\ Kh8; 1.2 =
30. /\ f6; 1.0 = 30. /\ g5; 1.0 = 30. /\ e6; 1.0 = 30. /\ Qg5} 31. Qe2 g5 {
3.0 = 31. /\ g5; 3.0 = 31. /\ Qb2; 2.3 = 31. /\ Qg7; 2.2 = 31. /\ Qa1; 2.
2 = 31. /\ h5; 2.2 = 31. /\ Qd4; 1.9 = 31. /\ h6; 1.8 = 31. /\ e6; 1.8 =
31. /\ f6; 1.8 = 31. /\ Kf8; 1.8 = 31. /\ Kh8; 1.8 = 31. /\ Qc3; 1.8 = 31.
/\ Qf6; 1.8 = 31. /\ Qh8; 1.8 = 31. /\ Qg7; 1.0 = 31. /\Qg5} 32. Kf2 h5 {
3.0 = 32. /\ h5; 2.2 = 32. /\ Kf8; 1.9 = 32. /\ Kh8; 1.9 = 32. /\ f6; 1.9
= 32. /\ Qa1; 1.9 = 32. /\ Qb2; 1.9 = 32. /\ Qd4+; 1.2 = 32. /\ Qg7; 1.2 =
32. /\ Qf6; 1.0 = 32. /\h6} 33. Qd2 h4 {2.0 = 33. /\ h4; 3.0 = 33. /\ Kh7;
3.0 = 33. /\ Kg7; 3.0 = 33. /\ Kf8; 2.8 = 33. /\ Qb2; 2.7 = 33. /\ Kh8; 1.
9 = 33. /\ Qf6; 1.0 = 33. /\ f6} 34. gxh4 gxh4 {
2.0 = 34. /\ gxh4; 2.0 = 34. /\ Qb2; 1.0 = 34. /\ Qh2+; 1.0 = 34. /\ Qf4}
35. Kf1 Kf8 {2.7 = 35. /\ Kf8; 3.0 = 35. /\ Kg7; 2.8 = 35. /\ Kh7; 2.6 =
35. /\ Qb2; 2.6 = 35. /\ Qf6; 2.6 = 35. /\ Qh8; 2.6 = 35. /\ Qa1+; 2.4 =
35. /\ f6; 1.0 = 35. /\ e6} 36. Ke2 Ke8 {1.3 = 36. /\ Ke8; 3.0 = 36. /\
Qh2+; 1.3 = 36. /\ Qg7; 1.2 = 36. /\ Qa1; 1.0 = 36. /\ Qf6; 1.0 = 36. /\
Qb2; 1.0 = 36. /\ Kg7; 1.0 = 36. /\ Qh8} 37. Ke3 f5 {3.0 = 37. /\ f5; 2.3
= 37. /\ f6; 1.1 = 37. /\ Qg7; 1.1 = 37. /\ Qg5+; 1.1 = 37. /\ Qb2; 1.0 =
37. /\ Qa1; 1.0 = 37. /\ Qh8; 1.0 = 37. /\ Qf6; 1.0 = 37. /\ Qg3} 38. Qc2
Kf7 {2.6 = 38. /\ Kf7; 3.0 = 38. /\ Kf8; 2.9 = 38. /\ f4+; 1.2 = 38. /\ e6;
1.1 = 38. /\ Qd4+; 1.1 = 38. /\ Qb2; 1.0 = 38. /\ fxe4} 39. Qd3 f4+ {
3.0 = 39. /\ f4+; 3.0 = 39. /\ Kf6; 2.0 = 39. /\ Kg7; 1.0 = 39. /\ Qa1; 1.
0 = 39. /\Qb2} 40. Kf2 Qb2+ {
3.0 = 40. /\ Qb2+; 1.0 = 40. /\ Qg5; 1.0 = 40. /\ Qg7} 41. Qe2 0-1


Test Game 5 Results


Image

Test Game 5 Rankings

Code: Select all

1	Critter 1.6a 64 Bit - AMD Phenom 2 Core 2.8GHZ	74.2	92.75%	2783
2	Julio Bolbochán   	70.4	88.00%	2640
3	Nova Jade 2	69.0	86.25%	2588
4	Saitek Travel Champion 2100	63.2	79.00%	2370
5	Mephisto TM Vancouver 68030 36 MHz	63.2	79.00%	2370
6	CXG 3000	61.8	77.25%	2318
7	Radioshack 2250XL Brute Force	61.4	76.75%	2303
8	Tasc CM 512K – 15 MHZ – KING 2.54	60.7	75.88%	2276
9	Saitek Corona	60.4	75.50%	2265
10	Radioshack 2250XL Selective	60.3	75.38%	2261
11	MChess Pro 5 - P75	56.6	70.75%	2123
12	Saitek MK 12 Trainer LV 5 90S/Move	52.6	65.75%	1973
13	Novag Constellation JR	46.2	57.75%	1733
14	Fidelity Sensory 9	45.2	56.50%	1695
15	Saitek MK 12 Trainer LV 4 15S/Move	40.4	50.50%	1515
Who would have thought that Novag Jade 2 would show up with a Grandmaster performance in this game. But unfortunately not quite good enough to beat either of the two players at the end of the day. Look for the Blunder on the last move. This loses the game. CXG 3000 also scored extremely well for Richard Lang. It just shows that even with the lesser rated computers you can have fun following a Grandmaster game.

Average Ranking and Rating after 5 Games

Code: Select all

1	Critter 1.6a 64 Bit - AMD Phenom 2 Core 2.8GHZ	373.7	93.43%	2803
2	Grandmaster Performance Standard  	339.6	84.90%	2547
3	Tasc CM 512K – 15 MHZ – KING 2.54	301.1	75.28%	2258
4	Saitek Travel Champion 2100	299.4	74.25%	2246
5	Mephisto TM Vancouver 68030 36 MHz	295.9	73.98%	2219
6	Nova Jade 2	294.2	73.55%	2207
7	Radioshack 2250XL Brute Force	289.6	72.40%	2172
8	Radioshack 2250XL Selective	288.5	72.13%	2164
9	MChess Pro 5 - P75	278.9	69.73%	2092
10	Saitek Corona	277.4	69.35%	2081
11	CXG 3000	249.8	62.45%	1874
12	Fidelity Sensory 9	231.8	57.95%	1739
13	Saitek MK 12 Trainer LV 5 90S/Move	219.5	54.88%	1646
14	Saitek MK 12 Trainer LV 4 15S/Move	204.5	51.13%	1534
15	Novag Constellation JR	203.4	50.85%	1526
The King 2.54 still sits on top of the Dedicated Chart 289 ELO points below Grandmaster Performance Standard.

Ok a lot of the ELO's look high, but are they really? If I recall CC9 had a USCF Rating of ELO 1721. Here after 5 Games it is ELO 1739. Perhaps the Manufacturers knew more than we care to admit :P Ok lets see what the next 11 games bring.

Below is the link for the updated spreadsheet. I will now try and add Atlanta, Magellan and Turbostar 540 to these 5 tests before I move on to Test Game 6.

http://spacious-mind.com/forum_reports/ ... _final.ods

Best regards,
Last edited by spacious_mind on Sun Jun 09, 2013 7:14 pm, edited 1 time in total.
Nick
User avatar
spacious_mind
Senior Member
Posts: 3999
Joined: Wed Aug 01, 2007 10:20 pm
Location: Alabama
Contact:

Post by spacious_mind »

Reinfeld wrote:For Nick's spreadsheet, here's the record of actual moves for RS 2150L, RS 2200X and Explorer Pro in the game 4 test (Unzicker v Sanchez):

RS 2150L

16. Ne3
17. d5
18. Ng5


- R.
Hi Reinfeld,

Thanks! I will add them a little later and repost the spreadsheet.

Regards
Nick
User avatar
Theo
Member
Posts: 132
Joined: Tue Mar 05, 2013 11:34 am

Post by Theo »

Good evening,

I just compared three Hiarcs versions on my 33Mhz Palm with GAME 1 from the test suite. Unregistered Hiarcs only plays 20 moves with normal strenth, so it was a rather short experience.

Embarassing: Hiarcs 9.6 scored better than Hiarcs 12.1 which scored better than Hiarcs 13.1 :-)


Hiarcs 13.1 Hiarcs 12.1 Hiarcs 9.6

7.Bb5+ 1,8 7.Bb5+ 1,8 7.Bb5+ 1,8
8.Be2 0,0 8.Bg5 1,6 8.Bf4 3,0
9.dxe6 0,0 9.Bxf6 1,0 9.Qc2 1,9
10.Qc2 4,0 10.Qd2 2,6 10.Qd2 2,6
11.Qc2 1,6 11.Bc4 3,0 11.Na4 0,0
12.0-0 3,0 12.0-0 3,0 12.Na4 0,0
13.dxe6 1,4 13.dxe6 1,4 13.Na4 0,0
14.Qc2 2,0 14.Qd3 1,2 14.Qc2 2,0
15.a5 1,4 15.dxe6 2,0 15.dxe6 2,0
16.dxe6 1,0 16.Rfd1 1,2 16.dxe6 1,0
17.Nd2 1,2 17.Nd2 1,2 17.Nxe5 1,5
18.g3 0,0 18.dxe6 1,0 18.h3 3,0
19.Na4 3,0 19.Dd2 0,0 19.Na4 3,0
20.Rxa4 2,1 20.Qxa4+ 3,0 20.Qxa4+ 3,0
22,5 24,0 24,8

*I apologize for the poor formatting of the table.*


No way I am gonna spend the whopping 29.95 Euro for the Hiarcs 13.1 for Palm licence! Its twice of what I spent for the Palm device itself :shock:

Chess Tiger 2010 for Palm is free Regards,
Theo
User avatar
spacious_mind
Senior Member
Posts: 3999
Joined: Wed Aug 01, 2007 10:20 pm
Location: Alabama
Contact:

Post by spacious_mind »

Radioshack 2200X scores the second highest dedicated score in Test Game 4 just behind Tasc CM 512K – 15 MHZ – KING 2.54. These Morschies :P really do seem to like these tests so far, can't wait to add Atlanta and Magellan. At the moment I am starting to wonder if the GK2100 and 2000's are better human match play opponents than Atlanta or Magellan :)



Test Game 4 Rankings

Code: Select all

1	Critter 1.6a 64 Bit - AMD Phenom 2 Core 2.8GHZ	76.2	95.25%	2858
2	Grandmaster Performance Standard	69.5	86.88%	2621
3	Tasc CM 512K – 15 MHZ – KING 2.54	61.4	76.75%	2303
4	Radioshack 2200X Selective	60.5	75.63%	2269
5	Saitek Travel Champion 2100	58.7	73.78%	2201
6	Radioshack 2250XL Selective	58.7	73.78%	2201
7	Saitek Explorer Pro Selective	57.5	71.88%	2156
8	Radioshack 2250XL Brute Force	56.6	70.63%	2119
9	Mephisto TM Vancouver 68030 36 Mhz	56.3	70.38%	2111
10	Novag Jade 2	53.1	66.38%	1991
11	Mchess Pro 5 – P75	52.9	66.13%	1984
12	Saitek MK 12 Trainer LV 5 90S/Move	52.4	65.50%	1965
13	Saitek Corona	51.6	64.50%	1935
14	CXG 3000	51.6	64.50%	1935
15	Saitek MK 12 Trainer LV 4 15S/Move	51.4	64.25%	1928
16	Radioshack 2150L Selective	51.0	63.75%	1913
17	Fidelity Sensory 9	46.7	58.38%	1751
18	Novag Constellation JR	42.3	52.88%	1586

Here is the update Spreadsheet:

http://spacious-mind.com/forum_reports/ ... _final.ods

@ Theo, Sorry I am not sure if I can rate your Hiarcs games for Test 1 if only 20 moves were played!

Best regards,
Nick
User avatar
Theo
Member
Posts: 132
Joined: Tue Mar 05, 2013 11:34 am

Post by Theo »

spacious_mind wrote:
@ Theo, Sorry I am not sure if I can rate your Hiarcs games for Test 1 if only 20 moves were played!

Best regards,
Hello Nick,

yes I know I can't rate them based on 15 positions really. From 15 positions its also impossible to check if programs are related:

Hiarcs 13.1 and 12.1 have only 4/15 moves in common. Hiarcs 12.1 and 9.4 have also only 4/15 moves in common. Thats 26.7% and they are *supposedly* from the same author!

I will continue with the Mephisto Nigel Short now. Back in 1992 my first computer was the Mephisto Milano and now I am eager to find out what the difference between them is.

Also, testing the free Chess Tiger 2010 (= same engine as 14.9) for Palm is of interest for me. I'd prefer Hiarcs, but getting a 400Mhz Tungsten instead of my 33Mhz Clie costs less than a Hiarcs licence :-)

Kind Regards,
Theo
User avatar
Steve B
Site Admin
Posts: 10140
Joined: Sun Jul 29, 2007 10:02 am
Location: New York City USofA
Contact:

Post by Steve B »

Theo wrote:
yes I know I can't rate them based on 15 positions really. From 15 positions its also impossible to check if programs are related:

Hiarcs 13.1 and 12.1 have only 4/15 moves in common. Hiarcs 12.1 and 9.4 have also only 4/15 moves in common. Thats 26.7% and they are *supposedly* from the same author!
well a 26.7% match rate
clearly these two engines do not come from the same author
seems like a "reasonable conclusion" that folks are making around here these days

Bordering on the Absurd Regards
Steve
User avatar
spacious_mind
Senior Member
Posts: 3999
Joined: Wed Aug 01, 2007 10:20 pm
Location: Alabama
Contact:

Post by spacious_mind »

Theo wrote:
spacious_mind wrote:
@ Theo, Sorry I am not sure if I can rate your Hiarcs games for Test 1 if only 20 moves were played!

Best regards,
Hello Nick,

yes I know I can't rate them based on 15 positions really. From 15 positions its also impossible to check if programs are related:

Hiarcs 13.1 and 12.1 have only 4/15 moves in common. Hiarcs 12.1 and 9.4 have also only 4/15 moves in common. Thats 26.7% and they are *supposedly* from the same author!

I will continue with the Mephisto Nigel Short now. Back in 1992 my first computer was the Mephisto Milano and now I am eager to find out what the difference between them is.

Also, testing the free Chess Tiger 2010 (= same engine as 14.9) for Palm is of interest for me. I'd prefer Hiarcs, but getting a 400Mhz Tungsten instead of my 33Mhz Clie costs less than a Hiarcs licence :-)

Kind Regards,
Theo
Hi Theo,

Thanks, I look forward to you completing Nigel Short. I always found Nigel Short to be a great program for it's 5 Mhz. I have examples of when I played it against Saitek Risc 2500 of it beating the Risc. So it can have good days.

Also please remember this is a rating test not a clone test. Besides there is a big difference in software that fits in a small 32K ROM and software that can take up your whole computer :)

Regards

Nick
Nick
Post Reply