Your Country Needs You!

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
Bryan Whitby
Senior Member
Posts: 1001
Joined: Wed Feb 18, 2009 9:57 pm
Location: England

Post by Bryan Whitby »

I now have Stockfish 8 on my Hudl2 but I really don't have the enthusiasm to do another test.
Josef
Member
Posts: 178
Joined: Mon Dec 12, 2016 12:56 pm

Post by Josef »

I bought Mephisto MM V HG550 one month ago. I love it, Ed Schröder made a great software, well done!! 8 bit CPU 65C02 beat 16 bit CPU 68000 and it is not far from Revelation II MM V version.

Image
Josef
Member
Posts: 178
Joined: Mon Dec 12, 2016 12:56 pm

Post by Josef »

Chessmaster Ireland wrote:I now have Stockfish 8 on my Hudl2 but I really don't have the enthusiasm to do another test.
I understand, it's a matter of the hardware only, better hardware = better result.
User avatar
spacious_mind
Senior Member
Posts: 4000
Joined: Wed Aug 01, 2007 10:20 pm
Location: Alabama
Contact:

Post by spacious_mind »

Josef wrote:Yes, but I find out, Gavon played sometimes another move than in the last game. This is the best result from several games. For example, Game Test 1 played 4. ... c5, but sometimes played 4. ... Nf6. Anyway, move 19. ... Qe4+ is designated like the best move, but Komodo 9 and Stockfish 8 on the PC say 19. ... Qxg2 is much better and Gavon played Qxg2.

Image
Hi Josef,

When I did these tests a couple of years ago, they were done on a fast 8 core desktop. Every move was analyzed between a minimum of 20 minutes and maximum 24 hours, depending on the move's complexity. Therefore even on today's I7's, I would be surprised if you were to find many moves in these tests that would vary a lot from the test scores, if you were to allow your computer to run the position deeply enough. Especially since these tests are meant to be run at either 30 seconds per move or 3 min per move.

If you run the 19th move to for example 30+ plies you should see that the move 19. ... Qe4+ is better than 19. ... Qg2. Hence the score difference.

Best regards
Nick
Josef
Member
Posts: 178
Joined: Mon Dec 12, 2016 12:56 pm

Post by Josef »

Hi

I tested several engines for 35 plies. Houdini 5 display Qe4+ but Komodo 9 or Stockfish 8 display Qxg2 (OSX Lion, 2x Xeon 3.0 Ghz, total 8 core, 32GB RAM).
User avatar
spacious_mind
Senior Member
Posts: 4000
Joined: Wed Aug 01, 2007 10:20 pm
Location: Alabama
Contact:

Post by spacious_mind »

Josef wrote:Hi

I tested several engines for 35 plies. Houdini 5 display Qe4+ but Komodo 9 or Stockfish 8 display Qxg2 (OSX Lion, 2x Xeon 3.0 Ghz, total 8 core, 32GB RAM).
Interesting that you show a difference. I just ran the program I used back then with Arena and Qxg2 would not have scored any points after 35 ply so the small score it receives in the test was being generous. :)

I also ran quickly Stockfish 8 to the 36th ply and Qe4 came out as the best move after completing 35 ply.

I have now configured engine and arena to play exactly the same as in my old tests. I will let you know shortly the result of Stockfish 8.

ps. after 32 Ply Qe4 = -2.56 Qxg2 = -2.92 for SF8 and still running.

Best regards
Nick
User avatar
spacious_mind
Senior Member
Posts: 4000
Joined: Wed Aug 01, 2007 10:20 pm
Location: Alabama
Contact:

Post by spacious_mind »

Hi Josef,

After 34 Ply Qxg2 -2.59 and Qe4+ -2,61. So Qxg2 indeed made a comeback.

However after 35 it is Qe4+ -2.63 and Qg2 -2.90 so it looks like it is back to the same delta of difference that overall favors the Qe4+ move.

In order to evaluate the worth of every single possible move so that I can correctly score each move, I have to widen the parameters in Arena so that I can see all the possible moves. Therefore it takes a little longer once you get higher in the ply's.

I will let it run for another hour.

Best regards

Nick
Nick
User avatar
spacious_mind
Senior Member
Posts: 4000
Joined: Wed Aug 01, 2007 10:20 pm
Location: Alabama
Contact:

Post by spacious_mind »

Below is the result after 36 ply as you can see Qe4+ is -2.5718 and Qxg2 is -2.9318 a difference of 36 points.

36/77 3:00:05 87,740,887,126 8,120,318 -M1 18. ... Bf8 19.Qxd7+
36/77 3:00:05 87,740,887,126 8,120,318 -M2 18. ... Bxf4 19.Qxd7+ Kf8 20.Qd8+
36/77 3:00:05 87,740,887,126 8,120,318 -M2 18. ... Be5 19.Qxd7+ Kf8 20.Qd8+
36/77 3:00:05 87,740,887,126 8,120,318 -M2 18. ... Ba3 19.Qxd7+ Kf8 20.Qd8+
36/77 3:00:05 87,740,887,126 8,120,318 -M2 18. ... Bc5 19.Qxd7+ Kf8 20.Qd8+
36/77 3:00:05 87,740,887,126 8,120,318 -M3 18. ... Be7 19.Qxd7+ Kf8 20.Qc8+ Bd8 21.Qxd8+
36/77 3:00:05 87,740,887,126 8,120,318 -M11 18. ... Qg4 19.hxg4 0-0 20.Bxd6 c6 21.Qxc6 Re8+ 22.Be5 Nf6 23.g5 Rf8 24.gxf6 g6 25.Qd6 Re8 26.Qe7 h5 27.Qxe8+ Kh7 28.Qf8 g5 29.Qg7+
36/77 3:00:05 87,740,887,126 8,120,318 -M13 18. ... Qd3 19.Rxd3 0-0 20.Bxd6 cxd6 21.Qxd7 d5 22.Rxd5 h6 23.Rd6 g6 24.Qe7 Kg7 25.0-0 Rg8 26.Qf6+ Kh7 27.Rd7 Rf8 28.Rxf7+ Rxf7 29.Qxf7+ Kh8 30.Re1 g5 31.Re8+
36/77 3:00:05 87,740,887,126 8,120,318 -M14 18. ... Qg3 19.Bxg3 c6 20.Qxc6 Bb4+ 21.c3 0-0 22.Qxd7 Kh8 23.cxb4 f6 24.Qe7 Rg8 25.Rd8 h6 26.Rxg8+ Kxg8 27.Qe8+ Kh7 28.Kd2 h5 29.Qxh5+ Kg8 30.Re1 g5 31.Re7 f5 32.Qe8+
36/77 3:00:05 87,740,887,126 8,120,318 -M14 18. ... Qh6 19.Bxh6 Kd8 20.0-0 Re8 21.Bf4 Re6 22.Bxd6 cxd6 23.Rfd1 Ke7 24.Rxd6 Rxd6 25.Qb4 f6 26.Qxd6+ Kf7 27.Qxd7+ Kg6 28.Rd5 h6 29.Qg4+ Kf7 30.Rd7+ Kf8 31.Qxg7+ Ke8 32.Qe7+
36/77 3:00:05 87,740,887,126 8,120,318 -M15 18. ... Qf5 19.Rxf5 Be7 20.Rd5 0-0 21.Rxd7 Bd8 22.Bxc7 Bxc7 23.Rxc7 h6 24.Qd7 Kh7 25.Ke2 g5 26.Rc8 Rxc8 27.Qxc8 Kg6 28.Rd1 h5 29.Rd7 Kf6 30.Qg8 Ke5 31.Qxf7 h4 32.Rd5+ Ke4 33.f3+
36/77 3:00:05 87,740,887,126 8,120,318 -M16 18. ... Qg5 19.Bxg5 0-0 20.Qxd7 Bb4+ 21.Ke2 c6 22.Qxc6 h6 23.Be3 Rb8 24.Qc7 Re8 25.Rd8 Rxd8 26.Qxd8+ Kh7 27.Qe8 f6 28.Qe4+ Kh8 29.Qxb4 h5 30.Rd1 Kh7 31.Qe4+ f5 32.Qxf5+ g6 33.Qf7+ Kh8 34.Bd4+
36/77 3:00:05 87,740,887,126 8,120,318 -M29 18. ... Qh5 19.Bxd6 Qxd5 20.Qxd5 cxd6 21.Qxd6 h5 22.0-0 Rh6 23.Re1+ Re6 24.Rxe6+ fxe6 25.Qxe6+ Kd8 26.Qg8+ Kc7 27.Qxg7 Kc8 28.Qh8+ Kb7 29.a4 Nb6 30.a5 Nd5 31.Qxh5 Kc6 32.Qe8+ Kd6 33.a6 Kc7 34.Qf7+ Kd6 35.Qxd5+ Kxd5 36.a7 Kd6 37.a8Q Kd7 Qa4 Ke7 Qc6 Kf7 Qe4 Kg7 41.f4 Kh6 Qe5 Kg6 Qe7 Kh6
36/77 3:00:05 87,740,887,126 8,120,318 -54.57 18. ... Rg8 19.0-0 Ke7 20.Bxd6+ cxd6 21.Re1+ Qe6 22.Rxe6+ fxe6 23.Rd1 g5 24.Qc6 d5 25.Re1 Nf8 26.Qxd5 Rg7 27.Qb7+ Kf6 28.Qf3+ Ke7 29.Re3 g4 30.hxg4 Rg5 31.Qb7+ Nd7 32.Qc6 Re5 33.Rxe5 Nxe5 34.Qc5+ Kf6 35.Qd4 h5 36.Qf4+ Ke7 37.Qxe5 hxg4 38.Qg7+ Kd6 39.Qxg4 Kd5 40.Qh5+ Ke4 41.Qe8 Kd5 42.Qf7 Ke5 43.Qe7 Kf5 44.g4+ Ke5 45.Qg5+ Kd6 46.Qd2+ Ke5
36/77 3:00:05 87,740,887,126 8,120,318 -29.31 18. ... c5 19.0-0 0-0 20.Bxd6 Nf6 21.Re5 Rc8 22.Qb7 h6 23.Qxc8+ Kh7 24.Qf5 h5 25.Bxc5 Qxf5 26.Rxf5 Ng8 27.Rxh5+ Kg6 28.Rh4 Nf6 29.a4 Kf5 30.Rd4 Kg6 31.a5 Ne8 32.a6 Nc7 33.a7 Kh7 34.Rd7 Na8 35.Rd8 Nc7 36.Bd6 Ne6 37.a8Q Ng5 Qd5 Kg6 39.f4 Nh7 40.f5+ Kg5 Qxf7
36/77 3:00:05 87,740,887,126 8,120,318 -29.31 18. ... Rf8 19.Kd1 Kd8 20.Qc6 Re8 21.Bxd6 cxd6 22.Rxd6 Qxd6+ 23.Qxd6 Re6 24.Qd4 f6 25.Kc1 Ke8 26.Rd1 Nf8 27.a4 Ke7 28.a5 Kf7 29.Qc4 Kg6 30.a6 Re7 31.Qa4 Kh6 32.a7 Ng6 33.a8Q Re5 34.Qd7 Re2 Qg8 Ne7 Qc4 Rxc2+ Qxc2+ Ng8+ Qc7+ Ne7+ 39.Qxe7
36/77 3:00:05 87,740,887,126 8,120,318 -27.93 18. ... f5 19.Bxd6 cxd6 20.0-0 0-0 21.Qxd7 h6 22.Rxd6 Qh5 23.Rfd1 Kh7 24.a4 Ra8 25.Qc6 Ra5 26.Rd8 Re5 27.Qc4 Re8 28.Qd5 Re7 29.a5 Qf7 30.Qxf7 Rxf7 31.R8d7 Rf6 32.R1d6 Rf8 33.Re6 Kg8 34.Re1 Ra8 35.Ra1 Ra6 36.b4 f4 37.b5 Rf6 38.a6 Rf7 39.Rxf7 f3 40.Rxf3 g5 41.a7 Kh7 42.a8Q Kg6 43.b6
36/77 3:00:05 87,740,887,126 8,120,318 -24.95 18. ... Kf8 19.0-0 Kg8 20.Qxd7 Bf8 21.Qxc7 h6 22.Rd8 Kh7 23.Qb8 Qf5 24.Bd6 Qe6 25.Rxf8 Rxf8 26.Qxf8 Kg6 27.a4 h5 28.a5 Qa2 29.Re1 Qd5 30.b4 Qd2 31.Qe7 Qxc2 32.a6 Qa2 33.a7 Kh6 34.Bf4+ Kh7 35.Qe4+ g6 36.a8Q Qxa8 37.Qxa8 g5 38.Qe4+ Kg7 39.Bxg5 f5 40.Qxf5
36/77 3:00:05 87,740,887,126 8,120,318 -23.21 18. ... f6 19.0-0 0-0 20.Bxd6 cxd6 21.Qxd7 h6 22.a4 Qe4 23.Rxd6 Qe5 24.Rfd1 Kh7 25.b4 Re8 26.a5 Re7 27.Qc8 Qe4 28.a6 Ra7 29.R6d5 Re7 30.b5 f5 31.Qc5 f4 32.a7 Re6 33.a8Q Rg6 34.g3 fxg3 35.f3 Qxf3 Qe8 Qf6 37.Rd8 Rg5 38.Qe3 Qf2+ 39.Qxf2 gxf2+ 40.Kxf2 Rf5+ 41.Ke3 Rg5+ Qh8 Kg6 Qg8 Re5+ 44.Kd4 Rxb5
36/77 3:00:05 87,740,887,126 8,120,318 -18.05 18. ... h6 19.Bxd6 cxd6 20.0-0 Ke7 21.Re1+ Qe6 22.Rxe6+ fxe6 23.Rd1 Rd8 24.Qb7 d5 25.Re1 Kf7 26.Qc7 Ke7 27.Qg3 Rg8 28.Qc3 Nf8 29.Qf3 g5 30.Qxd5 Rg7 31.a4 Kf7 32.Qd8 h5 33.a5 g4 34.hxg4 Rxg4 35.a6 Ra4 36.Qc7+ Kg8 37.a7 Ra2 38.Qb7 Ra4 39.a8R Rxa8 40.Qxa8 Kf7 41.Qb7+ Kf6 42.Qa7 h4 43.Qd4+ Ke7
36/77 3:00:05 87,740,887,126 8,120,318 -16.38 18. ... Kd8 19.0-0 Re8 20.Bxd6 cxd6 21.Qa5+ Ke7 22.Qc7 Kf8 23.Qxd7 Qe6 24.Rxd6 Qxd7 25.Rxd7 Re4 26.Rc7 g5 27.Rd1 Kg7 28.Rdd7 Rf4 29.b3 Kg6 30.a4 h5 31.Rxf7 Rb4 32.a5 Rb8 33.a6 Ra8 34.a7 h4 35.Rfd7 Kf5 36.Rc6 Ke4 37.Rb7 Kd5 38.Ra6 Kd4 39.Rb8 Rxa7 40.Rxa7 Kc3 41.Rc7+ Kb2 42.Rg8 Kc1 43.Rxg5
36/77 3:00:05 87,740,887,126 8,120,318 -14.61 18. ... Ke7 19.0-0 Qe6 20.Bxd6+ cxd6 21.Qa5 Rc8 22.Re1 Rxc2 23.Rxe6+ fxe6 24.Rd3 Rc1+ 25.Kh2 Rc5 26.Qd2 Nf6 27.Rxd6 Rc7 28.Qf4 Rd7 29.Rxd7+ Nxd7 30.Qb4+ Ke8 31.Qd6 e5 32.Qe6+ Kd8 33.Qg8+ Kc7 34.Qxh7 g5 35.Qg6 Nc5 36.Qxg5 Kd6 37.Qg6+ Kd5 38.Qf7+ Kd6 39.Qc4 e4 40.Qd4+ Kc6 41.b4 Nd7 42.Qxe4+ Kd6 43.Qd4+ Ke6 44.a4 Nf6 45.a5 Kf5 46.a6
36/77 3:00:05 87,740,887,126 8,120,318 -13.12 18. ... h5 19.0-0 0-0 20.Bxd6 c6 21.Qxc6 Qe6 22.a4 Nf6 23.Re5 Rc8 24.Qb7 Qd7 25.Qxd7 Nxd7 26.Re7 Nb6 27.a5 Nc4 28.Bb4 Rb8 29.Bc3 Nxa5 30.Bxa5 Rxb2 31.Rd1 Rxc2 32.Rd8+ Kh7 33.Rxf7 Kg6 34.Rdd7 Rc5 35.Rxg7+ Kf5 36.Bd8 Rc1+ 37.Kh2 Ra1 38.Rd5+ Ke6 39.Rxh5 Ra2 40.Rh6+ Kd5 41.Rg5+ Kc4 42.Rh4+ Kc3 43.Bf6+ Kd3 44.Rg3+ Kd2 45.Rd4+ Kc1
36/77 3:00:05 87,740,887,126 8,120,318 -11.74 18. ... Qxc2 19.0-0 0-0 20.Bxd6 c6 21.Qd3 Qxd3 22.Rxd3 Ra8 23.Rc1 h5 24.Ra3 Rd8 25.b4 Kh7 26.Ra7 Nf8 27.Rxc6 Ne6 28.Rxf7 Nd4 29.Ra6 Kg8 30.Rfa7 Re8 31.Bc5 Ne6 32.Be3 Kh7 33.b5 Kg6 34.a4 Kf5 35.a5 g6 36.Rc6 g5 37.Rd7 Ke5 38.a6
36/77 3:00:05 87,740,887,126 8,120,318 -11.45 18. ... Qf6 19.Bxd6 cxd6 20.Rd3 0-0 21.Qxd7 Qxb2 22.0-0 Qxa2 23.Qc7 d5 24.Rfd1 Qc4 25.Qe7 Qe4 26.Qc5 Qc4 27.Rxd5 Qxc5 28.Rxc5 f5 29.Rd7 h5 30.Rcc7 Ra8 31.Rxg7+ Kh8 32.Rh7+ Kg8 33.Rxh5 Ra5 34.Rg5+ Kf8 35.c4 Re5 36.Kh2 Ra5 37.c5 f4 38.Rf5+ Kg8 39.Rxf4 Rb5 40.Rd4 Rb8 41.Rg4+ Kh8 42.c6 Rb2 43.Rc8+ Kh7
36/77 3:00:05 87,740,887,126 8,120,318 -10.23 18. ... Bb4+ 19.Kd1 0-0 20.Rxd7 Qxg2 21.Qd5 Qxd5+ 22.Rxd5 c6 23.Rd7 Ra8 24.a3 Bc5 25.Be3 Bxe3 26.fxe3 c5 27.Ke2 h5 28.Rc7 f6 29.Rg1 g5 30.Rf1 Rf8 31.Rxc5 g4 32.hxg4 hxg4 33.Rc6 Kg7 34.a4 g3 35.a5 Rb8 36.a6 Rxb2 37.Ra1 Rb8 38.a7 Ra8 39.Kf3 g2 40.Kxg2 f5
36/77 3:00:05 87,740,887,126 8,120,318 -9.83 18. ... Qe6+ 19.Kd1 Bxf4 20.Re1 Bd6 21.Rd3 h6 22.Rxe6+ fxe6 23.Re3 Ke7 24.Qf5 e5 25.Qg6 Rf8 26.Qxg7+ Kd8 27.Ke2 Re8 28.Qxh6 Nc5 29.Qg5+ Re7 30.Qg4 c6 31.Ra3 Ne6 32.Ra8+ Kd7 33.Ra7+ Bc7 34.Kf1 Re8 35.Qd1+ Ke7 36.Qh5 Kd7 37.Qf7+ Re7 38.Qf5 Re8 39.a4 Kd6 40.Qd3+ Ke7 41.Qe4 Kd7 42.a5 Nd4 43.c3 Nb5 44.Rb7 Kd6 45.Qg6+ Re6 46.Qd3+ Kc5
36/77 3:00:05 87,740,887,126 8,120,318 -8.70 18. ... c6 19.Qe2+ Be7 20.Ra5 Qe6 21.Ra8+ Bd8 22.Qxe6+ fxe6 23.Bg5 Nf6 24.a4 Kf7 25.Ke2 Re8 26.a5 h6 27.Be3 Be7 28.Rxe8 Nxe8 29.a6 Nc7 30.a7 e5 31.Ra1 Ke6 32.a8Q Nxa8 33.Rxa8 Bf6 34.b3 e4 35.Re8+ Kd5 36.c4+ Kd6 37.Rxe4 Kd7 38.Kf3 Be7 39.Rg4 g5 40.Re4 Kd6 41.b4
36/77 3:00:05 87,740,887,126 8,120,318 -7.20 18. ... 0-0 19.Bxd6 cxd6 20.0-0 Nc5 21.Re1 h5 22.a4 Qxc2 23.a5 g6 24.a6 Nxa6 25.Qxa6 Qxb2 26.Qxd6 Kg7 27.Rd4 Qb8 28.Qe7 Qb2 29.Qe5+ Kg8 30.Qf6 Qc3 31.Re7 Qc1+ 32.Kh2 Qh6 33.Rdd7 Qg7 34.Qf4 Qh7 35.Rd6 Qg7 36.Rd5 Qh7 37.Qd6 Qg7 38.Kg1 Qh6 39.Qe5 Qg7 40.Qf4
36/77 3:00:05 87,740,887,126 8,120,318 -2.93 18. ... Qxg2 19.Rf1 Qe4+ 20.Be3 Qe6 21.Rd4 0-0 22.Qd5 Qe8 23.Re4 Qc8 24.Bd4 g6 25.f4 Rd8 26.Ke2 Qa6+ 27.Kf2 Qa4 28.Rc1 Qa8 29.Qxa8 Rxa8 30.Ra1 f5 31.Re2 Ra4 32.Be3 Kf7 33.b3 Ra3 34.Kf3 Nf8 35.Bc1 Ra8 36.a4 Ne6 37.Be3 g5 38.fxg5 f4
36/77 3:00:05 87,740,887,126 8,120,318 -2.57 18. ... Qe4+ 19.Be3 Qxg2 20.Rf1 Qxh3 21.Rh5 Qe6 22.Rfh1 h6 23.Qd5 Ke7 24.Qxe6+ Kxe6 25.a4 Ra8 26.b3 Nf6 27.R5h3 Be5 28.Ke2 Ne4 29.f3 Nc3+ 30.Kf2 Nd5 31.Rh4 Nxe3 32.Kxe3 Ra5 33.Rh5 f5 34.R5h4 Bd6 35.Kd3 Be5 36.Kc4 Kf7 37.Kb4 Ra8 38.Rc4 Bd6+ 39.Kb5

Here is also the score for 37 ply. Qe4+ is -2.4518 and Qxg2 is -3.0318 which is a difference of 58 points. At 37 ply Qxg2 would also not have scored in my test. The 4.2 points Qxg2 is fair and reasonable considering I used a different program to SF8 for original the test a couple of years ago.

37/77 3:06:12 90,777,774,174 8,124,954 -8.93 18. ... c6 19.Qe2+ Be7 20.Ra5 Qe6 21.Ra8+ Bd8 22.Qxe6+ fxe6 23.Bg5 Nf6 24.a4 0-0 25.a5 Bc7 26.Rxf8+ Kxf8 27.a6 Bb6 28.Be3 Nd7 29.Ke2 c5 30.Ra1 Ba7 31.b4 Ke8 32.bxc5 Nb8 33.Rb1 Nxa6 34.Ra1 Bxc5 35.Rxa6 Be7 36.c4 Bf6 37.Rxe6+ Kd7 38.Rb6 Kd8 39.Rb7 Kc8 40.Ra7 Kd8 41.c5 h6 42.Kf3
37/77 3:06:12 90,777,774,174 8,124,954 -7.17 18. ... 0-0 19.Bxd6 Qe4+ 20.Kd1 cxd6 21.Rxd6 Qxg2 22.Qd5 Qxd5+ 23.Rxd5 Nb6 24.Rd4 f6 25.b3 Kf7 26.a4 Ra8 27.Rd6 Nxa4 28.bxa4 Rxa4 29.Kd2 Kg6 30.Rd7 Kh6 31.Rc7 Rf4 32.Ke3 g5 33.c4 Kg6 34.c5 h5 35.c6 Rc4 36.Kd3 Rc5 37.Rc8 Kf7 38.Re1 Rd5+ 39.Kc4 Rd2 40.Rc7+ Kg6 41.Rd7 Rxf2
37/77 3:06:12 90,777,774,174 8,124,954 -3.03 18. ... Qxg2 19.Rf1 Qe4+ 20.Be3 Qe6 21.Rd4 0-0 22.Qd5 Qe8 23.Re4 Qd8 24.Rg1 g6 25.Bh6 Qa8 26.Qxa8 Rxa8 27.a3 f5 28.Rd4 Ne5 29.Ke2 Re8 30.Be3 Kf7 31.Ra4 Re6 32.Rd1 g5 33.Kf1 Nf3 34.Ra5 Kg6 35.a4 Be7 36.Ke2 Ne5 37.Rb5 Nc4 38.Kd3 Nd6 39.Rc5 f4 40.Bd4 Nf5 41.Rxc7 Nxd4 42.Kxd4 Rd6+ 43.Kc4 Rxd1 44.Rxe7 h5 45.a5
37/77 3:06:12 90,777,774,174 8,124,954 -2.45 18. ... Qe4+ 19.Be3 Qxg2 20.Rf1 Qxh3 21.Rh5 Qe6 22.Rfh1 h6 23.Qd5 Ke7 24.Qxe6+ Kxe6 25.Bd4 f6 26.a4 Nb6 27.b3 Nd5 28.Be3 Rd8 29.Rg1 g5 30.Ke2 Bf8 31.Rgh1 Nb4 32.c3 Nd5 33.Bd2 f5 34.c4 Nf6 35.Rxh6 Bxh6 36.Rxh6 Kf7 37.Bxg5 Re8+ 38.Kf3 Ng4 39.Rh7+ Kg6 40.Re7 Rd8

It does make you wonder though by not doing a full search of all positions if a chess program including SF8 or Komodo 9 always make the wisest and best decisions in a non fully expanded search of all available moves of a chess position within a game?

best regards

Nick
Nick
User avatar
spacious_mind
Senior Member
Posts: 4000
Joined: Wed Aug 01, 2007 10:20 pm
Location: Alabama
Contact:

Post by spacious_mind »

Hi Josef,

I did let Stockfish 8 calculate up to 40 ply yesterday and the results are as follows:

Code: Select all

STOCKFISH 8			
Search Depth (ply)	18. … Qe4+	18. … Qxg2	  Best
8			                 -2.57		-2.56		Qxg2
9			                 -2.52		-2.56		Qe4+
10			                -2.52		-2.73		Qe4+
11			                -2.59		-2.73		Qe4+
12			                -2.59		-2.76		Qe4+
13			                -2.59		-2.77		Qe4+
14			                -2.59		-2.80		Qe4+
15			                -2.59		-2.80		Qe4+
16			                -2.59		-2.78		Qe4+
17			                -2.59		-2.79		Qe4+
18			                -2.59		-2.81		Qe4+
19			                -2.59		-2.80		Qe4+
20			                -2.59		-2.47		Qxg2
21			                -2.59		-2.64		Qe4+
22			                -2.59		-2.80		Qe4+
23			                -2.59		-2.72		Qe4+
24			                -2.59		-2.61		Qe4+
25			                -2.59		-2.67		Qe4+
26			                -2.59		-2.75		Qe4+
27			                -2.56		-2.67		Qe4+
28			                -2.52		-2.60		Qe4+
29			                -2.64		-2.61		Qxg2
30			                -2.54		-2.84		Qe4+
31			                -2.49		-2.50		Qe4+
32			                -2.56		-2.92		Qe4+
33			                -2.56		-2.81		Qe4+
34			                -2.61		-2.59		Qxg2
35			                -2.63		-2.90		Qe4+
36			                -2.57		-2.93		Qe4+
37			                -2.45		-3.03		Qe4+
38			                -2.61		-3.12		Qe4+
39			                -2.60		-2.84		Qe4+
40			                -2.59		-3.06		Qe4+
Out the 33 ply's shown above Qxg2 only evaluates best 4 times. But every time it does, it seems that the better variation it found gets refuted in the next ply search. Qe4+ with its -2.59 at ply 40 seems to be a much more consistent move choice. It shows -2.59 at ply 11 and still calculates the same at ply 40.

I am curious though about which program you are using Stockfish 8 on and how it looks between the ply's. I don't use Komodo 9 therefore I cannot make a comment on it.

Interesting also is that if you take the last 5 or 6 search ply from 40 downwards Qxg2 is trending worse. Therefore I do still continue to feel confident that of the two moves Qe4+ is the better move and deserves the better rating score.

This exercise was interesting for me as I created the tests a couple of years ago and I just wanted to check if with later software and faster computers the tests still stand up. To me it seems that they still do.

Best regards
Nick
Josef
Member
Posts: 178
Joined: Mon Dec 12, 2016 12:56 pm

Post by Josef »

Yes, you have right, very nice exercise! I think older or weak software will choose Qe4+, best software is not so explicit.
User avatar
spacious_mind
Senior Member
Posts: 4000
Joined: Wed Aug 01, 2007 10:20 pm
Location: Alabama
Contact:

Post by spacious_mind »

Just completed test with Amiga Sargon 3 running at 68060:

Image

Not a bad score just above Berlin Pro London and just below Tasc R30 King 2.5 Offensive.

Best regards
Nick
Josef
Member
Posts: 178
Joined: Mon Dec 12, 2016 12:56 pm

Post by Josef »

I ran two tests (test1 and 2) on my new Revelation II, this is a very good result.

Image; Image
Josef
Member
Posts: 178
Joined: Mon Dec 12, 2016 12:56 pm

Post by Josef »

Test 3

Image
Josef
Member
Posts: 178
Joined: Mon Dec 12, 2016 12:56 pm

Post by Josef »

Test 4, Smallfish versus Revelation II

Image
User avatar
spacious_mind
Senior Member
Posts: 4000
Joined: Wed Aug 01, 2007 10:20 pm
Location: Alabama
Contact:

Post by spacious_mind »

Hi Josef,

That just shows you how strong Stockfish is. Obviously a beast with Revelation II.

Best regards
Nick
Post Reply