Testing Gothic Vortex.....Rh3 wins.

Discussion of anything and everything relating to chess playing software and machines.

Moderators: bob, hgm, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
User avatar
George Tsavdaris
Posts: 1614
Joined: Thu Mar 09, 2006 11:35 am

Re: Testing Gothic Vortex.....Rh3 wins.

Post by George Tsavdaris » Mon Dec 03, 2007 8:19 pm

GothicChessInventor wrote: Vortex Copper is ancient history compared to the new product line, called Vortex Nuclear.

After the 2007 Championship, I will make a version of Nuclear available for download. It will come with 4-piece tablebases and up to 30 seconds per move thinking time. But, it is a 200 MB download!

Anyone who defeats Nuclear will have their name on the website, forever, provided they send me their game and their claim is true.
Look for it in early 2008.
My name is already written as the only human that beat the old Vortex so it's time to write it again. :lol: :D

I have started playing some games with the new Nuclear 2.2.5 version and:

Conditions are 10 seconds per move for both, and 256 MB hash for Vortex and Vortex is playing with all his 3,4 endgames tablebases.
I lost the first 3 games. I won the 4th, lost another 3 and won the 8th.
So we have so far 2-6 in favor of Vortex.

My impressions so far is that:
-Vortex is much much much more persistent in my attacks. It defends tremendously well, deploying an amazing counterplay each time and finding amazing defensive recourses.
-New Vortex is much much more aggressive. It seeks for blood! Sacrifices a lot.
-New Vortex seems a lot better at tactics, although the depth from the last Vortex i had, is more or less the same(better or the same but not worse).
-Positionally is a bit worse i think.
-You can't have a quiet moment with it. You can't play quietly against it. It will attack and you have no chance so you should always be ready for tactics.
-It's evaluation is too high. In the game i will show you below(one of my 2 wins and the most impressive) it evaluated in the opening 6...Bg7 as +1.69 in favor of black, while of course reality is different and white stands perfectly fine. This is only an example and these too high scores happens
all the time.
So it's not the most suitable for analysis if you are not experienced with it. Of course if you are then you know a score of Vortex to what corresponds.

Of course i don't expect yet moves like Axa1? to be scored as a win for white, but to see scores like +5.70 in favor of black when i have all my pieces in front of its King is no the best i would expect.....

Of course the short time control is a factor here and having more time, Vortex would perform better and have better evaluation of the position, but the same exists for me too. :)

The 4th game(i sacrificed a Rook and a Knight and after Vortex accepted my gifts it saw the situation and desperately gave me its material. :lol: A pretty ending.):
1. f4 Nh6
2. d4 d5
3. g3 Nc6
4. c3 g6
5. Nj3 Af6
6. h3 Bg7
7. g4 O-O
8. Be3 Be6
9. Bf3 Ad7
10. Cg2 f5
11. g5 Ng4
12. Bd2 j6
13. Nh4 Na5
14. b3 Ni5
15. i4 Ab5
16. j4 Nj7
17. Ci2 c5
18. dxc5 Aa6
19. Kg2 Rf8
20. e3 Nh8
21. Nj5 Ad3
22. Be1 Ab2
23. Qe2 Axa1
24. Nxi7 Kxi7
25. Af2 Bi5
26. jxi5 Nf7
27. ixj6+ Ki8
28. Ai5 Cf6
29. gxf6 Rh8
30. Bj7+ Kj8
31. Ch4 Qg8
32. Ah6+ Nxh6
33. Cxh6 g5
34. fxe7 Nc6
35. Bi5 gxf4+
36. Kh2 Qg7
37. Bi6 Qg6
38. Cj7+ Ki8
39. Ci7+ Kj8
40. Bxh7 Qg4
41. hxg4 Ne5
42. Bh6 Nf3+
43. Qxf3 Rxh7
44. Cxh7+ Ki8
45. Cj8#

And the generated log file by Vortex:
http://crocodile13.angelfire.com/log1.txt
After his son's birth they've asked him:
"Is it a boy or girl?"
YES! He replied.....

GothicChessInventor

Re: Testing Gothic Vortex.....Rh3 wins.

Post by GothicChessInventor » Tue Dec 04, 2007 2:46 am

George Tsavdaris wrote:
GothicChessInventor wrote: Vortex Copper is ancient history compared to the new product line, called Vortex Nuclear.

After the 2007 Championship, I will make a version of Nuclear available for download. It will come with 4-piece tablebases and up to 30 seconds per move thinking time. But, it is a 200 MB download!

Anyone who defeats Nuclear will have their name on the website, forever, provided they send me their game and their claim is true.
Look for it in early 2008.
My name is already written as the only human that beat the old Vortex so it's time to write it again. :lol: :D

I have started playing some games with the new Nuclear 2.2.5 version and:

Conditions are 10 seconds per move for both, and 256 MB hash for Vortex and Vortex is playing with all his 3,4 endgames tablebases.
I lost the first 3 games. I won the 4th, lost another 3 and won the 8th.
So we have so far 2-6 in favor of Vortex.

My impressions so far is that:
-Vortex is much much much more persistent in my attacks. It defends tremendously well, deploying an amazing counterplay each time and finding amazing defensive recourses.
-New Vortex is much much more aggressive. It seeks for blood! Sacrifices a lot.
-New Vortex seems a lot better at tactics, although the depth from the last Vortex i had, is more or less the same(better or the same but not worse).
-Positionally is a bit worse i think.
-You can't have a quiet moment with it. You can't play quietly against it. It will attack and you have no chance so you should always be ready for tactics.
-It's evaluation is too high. In the game i will show you below(one of my 2 wins and the most impressive) it evaluated in the opening 6...Bg7 as +1.69 in favor of black, while of course reality is different and white stands perfectly fine. This is only an example and these too high scores happens
all the time.
So it's not the most suitable for analysis if you are not experienced with it. Of course if you are then you know a score of Vortex to what corresponds.

Of course i don't expect yet moves like Axa1? to be scored as a win for white, but to see scores like +5.70 in favor of black when i have all my pieces in front of its King is no the best i would expect.....

Of course the short time control is a factor here and having more time, Vortex would perform better and have better evaluation of the position, but the same exists for me too. :)

The 4th game(i sacrificed a Rook and a Knight and after Vortex accepted my gifts it saw the situation and desperately gave me its material. :lol: A pretty ending.):
1. f4 Nh6
2. d4 d5
3. g3 Nc6
4. c3 g6
5. Nj3 Af6
6. h3 Bg7
7. g4 O-O
8. Be3 Be6
9. Bf3 Ad7
10. Cg2 f5
11. g5 Ng4
12. Bd2 j6
13. Nh4 Na5
14. b3 Ni5
15. i4 Ab5
16. j4 Nj7
17. Ci2 c5
18. dxc5 Aa6
19. Kg2 Rf8
20. e3 Nh8
21. Nj5 Ad3
22. Be1 Ab2
23. Qe2 Axa1
24. Nxi7 Kxi7
25. Af2 Bi5
26. jxi5 Nf7
27. ixj6+ Ki8
28. Ai5 Cf6
29. gxf6 Rh8
30. Bj7+ Kj8
31. Ch4 Qg8
32. Ah6+ Nxh6
33. Cxh6 g5
34. fxe7 Nc6
35. Bi5 gxf4+
36. Kh2 Qg7
37. Bi6 Qg6
38. Cj7+ Ki8
39. Ci7+ Kj8
40. Bxh7 Qg4
41. hxg4 Ne5
42. Bh6 Nf3+
43. Qxf3 Rxh7
44. Cxh7+ Ki8
45. Cj8#

And the generated log file by Vortex:
http://crocodile13.angelfire.com/log1.txt
I think it's time for a new topic George, maybe "Wins against Vortex" or something like this.

Recall this is a completely new version of Vortex. It has only been tested against older versions of itself and other Gothic Chess programs. I "threw this together" in the 3 weeks prior to the 2007 Gothic Chess Championship, after the bad lost to SMIRF in the blog game.

I found many bugs, and many things not behaving as I expected. I decided the simple solution was to throw everything out and start from scratch :)

One of the "new concepts" I know you will like. It treats the positional score and the material score like two sides of a right triangle, with the "true score" being a strange sort of hypotnuse connecting the two.

I say "strange" because the result is not the square root of the sum of the squares, but something similar.

So, the odd scores are when Vortex believes it can optimize some parameter you, as the human, are apparently not minimizing. Usually when the positional score gets high, the opponent must sacrifice something to lower it. When the material score gets high, you can't damage vortex's position to lower that component -- they are truly mutually exclusive. It's positional score can only rise when its king is safe. If its king is not safe, it seeks to make it safe one of two ways: attacking you first and putting your king in greater danger, or getting it to a safe haven where the attacks disappear over the search horizon.

Now, that is a mouthful, and it is easier said than done, but that's what I decided to do. From this point forward, it's just a matter of adjusting the various weights and seeing what happens :)

By the way, there is a menu item under the SPECIAL menu you will like. Select RANDOM 4-PIECE CHECKMATE and it will select from among the longest wins in the 4-piece tablebase and then make the first move. Put it on AUTOPLAY if you want to see the mate played perfectly, or try it yourself.

Enjoy!

GothicChessInventor

Re: Testing Gothic Vortex.....Rh3 wins.

Post by GothicChessInventor » Tue Dec 04, 2007 1:58 pm

George Tsavdaris wrote: My name is already written as the only human that beat the old Vortex so it's time to write it again. :lol: :D
t
Actually, the Wall of Honor was for wins scored against the program at the tournament time controls, which were like game in 45 minutes + 15 seconds per move, not each move in 10 seconds, as your game was.
George Tsavdaris wrote: I have started playing some games with the new Nuclear 2.2.5 version and:

Conditions are 10 seconds per move for both, and 256 MB hash for Vortex and Vortex is playing with all his 3,4 endgames tablebases.
I lost the first 3 games. I won the 4th, lost another 3 and won the 8th.
So we have so far 2-6 in favor of Vortex.
At 10 seconds per move on your machine, your hash table setting is too large. You're never even examining 3 million nodes. Each hash table entry is about 16 bytes, so anything over a theoretical 48 MB is wasted. Set your hash to 32 MB if you are running a fast game. Otherwise, at the start of each move, the program is initializing its hash table with millions of unused entries, and, after all, this for loop does take some time.

George Tsavdaris wrote: Of course the short time control is a factor here and having more time, Vortex would perform better and have better evaluation of the position, but the same exists for me too. :)
When I replayed the game, I saw Vortex was basically just playing almost entirely on its heuristics. The program was not optimized for speed games. In fact, most moves that it made were improved upon before 15 seconds elapsed on my machine. So this game really gives me no information.

As a comparison: imagine playing a chess program and restricting it to less than 3,000,000 nodes per move. Some programs would be searching less than one second on modern hardware! Now consider how much larger the branching factor is for Gothic Chess, and you can understand this is a harsh handicap.
George Tsavdaris wrote: And the generated log file by Vortex:
http://crocodile13.angelfire.com/log1.txt
I saw some log entries where only 4 or 5 plies were searched in 0 seconds. Did you force the program to move at all before its time was up? Or was this because of some forced play? I will have to look at the game a little more closely.

In any event, a new version of the Gothic Vortex opening book is out, and it will not play into the positions where it miscued due to extremely short search depths. Email me if you are interested in receiving it.

User avatar
George Tsavdaris
Posts: 1614
Joined: Thu Mar 09, 2006 11:35 am

Re: Testing Gothic Vortex.....Rh3 wins.

Post by George Tsavdaris » Tue Dec 04, 2007 5:30 pm

GothicChessInventor wrote: At 10 seconds per move on your machine, your hash table setting is too large. You're never even examining 3 million nodes. Each hash table entry is about 16 bytes, so anything over a theoretical 48 MB is wasted. Set your hash to 32 MB if you are running a fast game. Otherwise, at the start of each move, the program is initializing its hash table with millions of unused entries, and, after all, this for loop does take some time.
In my opinion and knowing from the Chess programs, a hash table of 256MB in comparison with one with 32 MB in a 10 seconds per move game would make no difference at all! Perhaps 1 ELO or less....
Of course there is the possibility that Vortex works differently and 32 MB would give it more strength....

As a comparison: imagine playing a chess program and restricting it to less than 3,000,000 nodes per move. Some programs would be searching less than one second on modern hardware! Now consider how much larger the branching factor is for Gothic Chess, and you can understand this is a harsh handicap.
Yes but it is still machine on 10 s/move Versus human on 10 s/move.
Fair is fair! For both.
Machine was not handicapped more than me and the opposite. :D

The way you make it sound is that you propose that before 1995 for example where ancient hardware existed, computers should not play against humans blitz games because of the slow hardware....
And actually in 10 s/move that is blitz time control, machine always has the advantage.

Fact is at 10 seconds per move, in a AMD3500+ 2.2GHz i won twice against Vortex and i lost 7 times(i had another loss today)....

I saw some log entries where only 4 or 5 plies were searched in 0 seconds. Did you force the program to move at all before its time was up? Or was this because of some forced play? I will have to look at the game a little more closely.
You will have to look the game more closely. :D
Needless to say that i didn't force the computer to move in any move. It played the last moves in zero seconds because they were forced moves....

In any event, a new version of the Gothic Vortex opening book is out, and it will not play into the positions where it miscued due to extremely short search depths. Email me if you are interested in receiving it.
I will. :D
I will play today late at night 2 games with 30 seconds per move for both. What hash do you suggest for 30 s/move?
After his son's birth they've asked him:
"Is it a boy or girl?"
YES! He replied.....

User avatar
George Tsavdaris
Posts: 1614
Joined: Thu Mar 09, 2006 11:35 am

Re: Testing Gothic Vortex.....Rh3 wins.

Post by George Tsavdaris » Tue Dec 04, 2007 11:17 pm

Hi Ed again,

I intended to play 2 games with 30 s/move against Vortex, but i played one, and it was one of the toughest games in Gothic Chess in my life and i'm very tired. I feel exhausted after a damn intense fight of 100 moves!
I used real clocks for me and since game lasted so long my brain is in big pain from the complete concentration. :D

I won the game finally! It was one of the most wonderful games and endgames specifically i have played in Gothic Chess.

I attacked, Vortex found amazing defensive resources, and turned my attacking position into a defensive one, thankfully not very difficult to defend and we went to an mini endgame where Vortex made my brain explode since i couldn't use its time while it was thinking to think for my own, since it played mostly surprising moves that i never got a correct ponder hit. :D
We went to a clear endgame and Vortex had a draw i think but i offered 2 Pawns for taking advantage of my advanced Passed Pawn and it worked and after an amazing endgame i outplayed it i think, and won.

The game with a few short comments i've made right now:
Me - G.Vortex
1. f4 d5
2. d4 Nc6
3. c3 Nh6
4. g3 g6
5. h3 Bg7
6. Nj3 e6
7. Bf3 Ae7
8. Ah2 Bd7
9. Cg2 O-O
10. e3 j5
11. Nd2 Rf8
12. Be2 Kh8
13. Nf3 e5
14. Nh4 i5
15. dxe5 ixh4
16. Nxh4 g5
17. Nf3 g4
18. Ng1 Ag6
19. hxg4 Ni4
20. Ai3 Qxi3! WOW! Nice one.
21. jxi3 Ae4
22. Bf3 Bxg4
23. Rxj5 Bh6
24. Bxe4 Bxd1
25. Bxd5 Cd8
26. Bf3 Bxf3
27. Nxf3 Cd1+
28. Ke2! Ch1
29. g4 Rg8
30. Kd3! Cf1
31. Nd4 Nxd4
32. exd4 c5
33. d5 Rd8
34. Kc4 Bi7
35. a4 b6
36. a5 Rd7
37. Rh5 Ng3
38. Rh4 Nf5
39. Rh3 b5+!! Vortex rocks! I never expected this and spent much time on this....
40. Kxc5 Cd1
41. d6 Rc8+
42. Kb4 Rc4+
43. Kb3 Nd4+!
44. cxd4 Cxd4+
45. Ka2 Rc2
46. Ce1 Bxf4
47. Rb1! My head was already one step from exploding since there were many variations and threats i had to deal with.
47...Bxc1
48. Rxc1 Ca4+
49. Ra3 Rxb2+
50. Ka1 Cxa3+
51. Kxb2 Ca4+
52. Kc2 Cd4+
53. Kc3 Cxg4
54. Cd3! f6
55. exf6 Cxf6
56. Rd1 Ce4+
57. Kc2 a6
58. Cc3 Ce8
59. Cf3! Cc8+
60. Kb1 Ki7
61. Cd4 Cc3+
62. Kb2 Cxi3
63. Rc1!! Cxi2+?! Hmm i don't know if this is actually good or bad.
64. Rc2 Ci3
65. Rc7 Cg2+
66. Kb3 Cg3+
67. Kb4 Rd8
68. d7! Cg2
69. Kc5 Cg6
70. Cd6 Cf8
71. Kb6 Cg8
72. Ra7!! Ce7
73. Kc7 Rg8
74. Rxa6 Rf8
75. Rb6 Ra8
76. a6 Ra7+
77. Rb7 Ra8
78. a7 b4
79. Kb6 Rd8
80. Cc6 Cd5+
81. Ka6 Cf5!! An amazing trap for me by Vortex. If i capture the Rook it will be a repetition draw i think, since i can't avoid the checks. I didn't spend even a second about analyzing the capture since i didn't want to risk anything.
82. Rxb4! Simple.... Cf7
83. Cb6 Cg5
84. Rc4 Ce5
85. Rc8 Cxd7
86. Cxd7 Rxd7
87. a8=Q Rd6+
88. Kb5 Rd3
89. Ri8+ Kh6
90. Qf8+ Kg6
91. Rg8+ Kh5
92. Qf5+ Ki6
93. Qxd3 h6
94. Qf5 h5
95. Rg6+ Ki5
96. Qh3 h4
97. Rg5+ Kh6
98. Qxh4+ Ki6
99. Ri5+ Kj6
100. Qj4#
1-0

The log of the game:
http://crocodile13.angelfire.com/log2.txt
After his son's birth they've asked him:
"Is it a boy or girl?"
YES! He replied.....

GothicChessInventor

Re: Testing Gothic Vortex.....Rh3 wins.

Post by GothicChessInventor » Wed Dec 05, 2007 10:39 am

George Tsavdaris wrote:
Me - G.Vortex

1. f4 d5
2. d4 Nc6
3. c3 Nh6
4. g3 g6
5. h3 Bg7
6. Nj3 e6
7. Bf3 Ae7
8. Ah2 Bd7
9. Cg2 O-O
10. e3 j5
11. Nd2 Rf8
12. Be2 Kh8
With the new King Safety code in place, this looks more comfortable based on the pawn structure. A wise move.
George Tsavdaris wrote: 13. Nf3 e5
14. Nh4 i5
Black has just played a Steinitz flank attack move!
George Tsavdaris wrote: 15. dxe5 ixh4
White gives up a knight on move 15 for the purpose of wrecking the black king position, but Vortex has brand new King Safety code and it knows how to reconstitue the position. This game will feature both players being "loose" with material for the purpose of "smacking around" the other player. It's just a matter of who can get to who first. This is a very different approach to the game, more like Greco in the 1600's of the Italian Era of chess.
George Tsavdaris wrote: 16. Nxh4 g5
17. Nf3 g4
18. Ng1 Ag6
19. hxg4 Ni4
20. Ai3 Qxi3!  WOW! Nice one.
All for the sake of turbulence, sweet turbulence! If 21. Cxi3? Nh2+ forks the Rook on j1.
George Tsavdaris wrote: 21. jxi3 Ae4
22. Bf3 Bxg4
How about giving 22...Bxg4! like it deserves! :)
George Tsavdaris wrote: 23. Rxj5 Bh6
I liked 23...Nxg3+ personally.
George Tsavdaris wrote: 24. Bxe4 Bxd1
25. Bxd5 Cd8
26. Bf3 Bxf3
27. Nxf3 Cd1+
28. Ke2! Ch1
29. g4 Rg8
30. Kd3! Cf1
31. Nd4 Nxd4
32. exd4 c5
33. d5 Rad8
34. Kc4 Bi7
35. a4 b6
36. a5 Rd7
The black chancellor is positioned perfectly to allow for the knight's safe passage via 36...Nh2 then ...Nf3 next, creeping ever closer to the white king. But that's the human's way of thinking, the machine is out for something completely different.
George Tsavdaris wrote: 37. Rh5 Ng3
38. Rh4 Nf5
39. Rh3 b5+!! Vortex rocks! I never expected this and spent much time on this....
40. Kxc5 Cd1
41. d6 Rc8+
42. Kb4 Rc4+
43. Kb3 Nd4+!
Black has just played a Fischer move!

Vortex was originally looking at 43...Nd4+ 44. cxd4 Rxc1 but just one move later it changed its mind. The program has some serious guts to be playing moves like this without recovering material right away!
George Tsavdaris wrote: 44. cxd4 Cxd4+
45. Ka2 Rc2
46. Ce1 Bxf4
47. Rb1! My head was already one step from exploding since there were many variations and threats i had to deal with.
47...Bxc1
I thought there was at least draw by repetition here with 47...Ca4+ 48. Kb3 49. Cd4+ etc., with check possible with ...Ca4+ no matter what, but your Rook on the 3rd rank can interpose. Trading Bishops was the correct course of action, and this is still a draw at this point.
George Tsavdaris wrote: 48. Rxc1 Ca4+
49. Ra3 Rxb2+
Amazing that there is not some kind of mate here! There must be something?!?
George Tsavdaris wrote: 50. Ka1 Cxa3+
51. Kxb2 Ca4+
52. Kc2 Cd4+
53. Kc3 Cxg4
54. Cd3! f6
So far all of the above moves by white are predicted by Vortex in its principal variation. Personally I would have opted for 54...Cc4+ and a draw since black is still down a pawn at this point, the pawn can be recovered, and the game becomes balanced in short order. This only requires a 7 ply search so this is a mystery to me why Vortex played 54...f6 here instead. Example 54...Cc4+ 55. Kb3 {forced} Cxa5+ and now:

A. 56. Kb2 Ca4+ 57. Kb1 Ki7 and I see no white advantage now that checking chances for white are exhausted while this is not the case for black.

B. 56. Kb4 Ca2+ 57. Kb3 {not 57. Kxb5?? Rb7+ mate in 8!} Ca5+ 58. Kb4 Ca2+ etc. draw, or transposing into line A.

C. 56. Kc3 and now f6! and 57. exf6? loses to ...Ca3+ exchanging chancellors on d3, then ...Rxd6+ and ...Rxf6 picking up the pawn with a stronger position. 57. Cc5 is better and ...Rd8 or ...Ca2+ should draw.

D. 56. Kc2 and black waits with ...Ki7, wanting white to move the chancellor so that ...Ca2+ can be played with ...Cd2 coming (d2 is controlled by the white chancellor on d3 currently.)
George Tsavdaris wrote: 55. exf6 Cxf6
56. Rd1 Ce4+
57. Kc2 a6
58. Cc3 Ce8
More white moves, all predicted by Vortex in every principal variation along the way. Vortex Nuclear needed both the 5-piece tablebases probed in RAM and more time to see 55...Cc4+! 56. Kb3 Rxd6!! and white can just barely draw the pawn ending even though up a full pawn.

Example: 57. Cxc4 Rxd3+ 58. Rc3 Rxc3+ 59. Kxc3 a6!! and now:

60. i4 Kg8 61. Kd4 Kf7 62. Kf7 Ke5

A. 63. i3 loses (the backwards doubled i-pawn) to 63...b4! 64. Kd4 Kxf6 65. Kc4 Ke5 66. Kxb4 Kd4 67. Kb3 Kc5 68. Kc3 Kb5 then ...Kxa5 and the black a-pawn now promotes.

B. 63. Kf5! the only draw! Starting the b-pawn now does not matter: 63...b4 64. Ke4 Kxf6 65. Kd4 Kf5! {65...Kg5 66. Kc4 Kh4 67. Kxb4 Kxi4 68. Kc5 Ki3 69. Kb6 h5!! 70. Kxa6 Kxi2 71. Kb7 h4 72. a6 h3 73. a7 h2 74. a8=Q h1=Q+ 75. Kb8 Qxa8+ 76. Kxa8 and a bare king draw!} 66. Kc4 Ke4 67. Kxb4 Kd5! and now the white king has been "shouldered off" by the black king. 68. Kc3 Kc5 69. Kd3 Kb4 70. Ke4 and it appears white will now head to the other side of the board and promote its i-pawn, but the tablebases again become the dominant factor and say draw! 70...Kxa5 71. Kf5 Kb4 72. Kg6 a5 73. Kxh6 and black is down 2 pawns to 1 but the 5-piece tablebases are unbeatable form here 73...a4 74. i5 a3 75. i6 a2 76. a7 a1=Q 77. i8=Q Qh1+ draw!

Probing the 5-piece tablebases in RAM is the only way to see all of this from move 55!
George Tsavdaris wrote:
59. Cf3! Cc8+
60. Kb1 Ki7
Again both white moves above predicted by Vortex...
George Tsavdaris wrote: 61. Cd4 Cc3+
62. Kb2 Cxi3
63. Rc1!! Cxi2+?! Hmm i don't know if this is actually good or bad.
Now we see some unanticipated play made by George as white. We see white sacrifice a pawn way down on i2 in order to hit the chancellor with the Rook and make strong overtures with the passed d-pawn coming into sharp focus. With the rooks still on the board, the Chancellor + Pawn vs. Chancellor tablebase cannot be probed to offer long distance insight. Black should decline the pawn and play 63...Ci6 to draw more easily.
George Tsavdaris wrote: 64. Rc2 Ci3
65. Rc7 Cg2+
Hidden tactical flares are revealed by the 5-piece tablebases after 65...Rd8 66. Rxh7+?! {tossing away the rook for a pawn} Kxh7 67. Ch4+!! Kg7 68. Cf5+! Kg6 69. Cf4+! Kh6 70. Cf7+ and the Rook will be recovered with Cxd8 next.
George Tsavdaris wrote: 66. Kb3 Cg3+
67. Kb4 Rd8
68. d7! Cg2
69. Kc5 Cg6
69...b4!? is a human way to deal with the issue, offsetting the white chancellor so that after 70. Cxb4?? Ce6+ wins or 70. Kxb4 Ce6 kicks the king out and the pawn on d7 has no monarch guiding it to promotion.

Clearly 69...Cg6 missed the mark in the game but 68...Cg2?? deserves to be tagged as a losing move. The natural 68...Cg7 is almost impossible to resist so Vortex must be put under the microscope to see what ails it.

The 5-piece tablebases are hit hard in this position, and 69. Rc6!! is seen as very strong, leading to a won Rook and pawn ending if 69...Ce5 70. Rxa6 Cxd7 71. Cxd7 Rxd7 72. Rb6 h5 73. a6 and white wins though down a pawn! These tablebases are amazing when probed in RAM!
George Tsavdaris wrote: 70. Cd6 Cf8
Could there be a move more passive that 70...Cf8?? I would think this has to be losing even. Almost anything is better, ...Cg7 waiting then ...b4 if the white king moves, or even the "obvious" ...Ce5+ looks ok at first glance... who wrote this stupid program! :)
George Tsavdaris wrote: 71. Kb6 Cg8
72. Ra7!! Ce7
More moves for white all predicted by Vortex but easy enough to see. And wy not 71...b4 to at least make some feigning of counterplay? I must have a chat with my program!
George Tsavdaris wrote: 73. Kc7 Rg8
74. Rxa6 Rf8
75. Rb6 Ra8
76. a6 Ra7+
76. Cf6!! and now ...Rxa5?? would be a tragic error since 77. Kb8!! creates an unstoppable promotion.
George Tsavdaris wrote: 77. Rb7 Ra8
78. a7 b4
79. Kb6 Rd8
80. Cc6 Cd5+
81. Ka6 Cf5!! An amazing trap for me by Vortex. If i capture the Rook it will be a repetition draw i think, since i can't avoid the checks. I didn't spend even a second about analyzing the capture since i didn't want to risk anything.
Correct. On 82. Cxd8?? Cc5+ 83. Kb6 Cc4+ 84. Kb5 Ce5+ 85. Kxb4 Ce4+ 86. Kb5 Cc3+87. Kb6 CC4+and the chancellor behind the king has perpetual check.
George Tsavdaris wrote: 82. Rxb4! Simple.... Cf7
Or Rb8! right away, both are winning. The rest of the game is easily won.
George Tsavdaris wrote: 83. Cb6 Cg5
84. Rc4 Ce5
85. Rc8 Cxd7
86. Cxd7 Rxd7
87. a8=Q Rd6+
88. Kb5 Rd3
89. Ri8+ Kh6
90. Qf8+ Kg6
91. Rg8+ Kh5
92. Qf5+ Ki6
93. Qxd3 h6
94. Qf5 h5
95. Rg6+ Ki5
96. Qh3 h4
97. Rg5+ Kh6
98. Qxh4+ Ki6
99. Ri5+ Kj6
100. Qj4#
1-0
Very nice game George. I will be adding my comments to it also. You have to admit it is much more fun to play against this version of Vortex than the older one! Although I still have some code tweaking to do.

GothicChessInventor

Re: Testing Gothic Vortex.....Rh3 wins.

Post by GothicChessInventor » Fri Dec 07, 2007 3:50 pm

rightrook wrote:OK..thanks...thanks good news..will be watching for it..

regards,

Robert
Robert,

You can join the Gothic Chess discussion board if you want me to see your posts a little quicker.

:)

http://s13.invisionfree.com/Gothic_Ches ... /index.php

supersharp77
Posts: 764
Joined: Sat Jul 05, 2014 5:54 am
Location: Southwest USA

Re: Testing Gothic Vortex.....Rh3 wins.

Post by supersharp77 » Wed Jul 10, 2019 5:05 am

rightrook wrote:
Fri Nov 30, 2007 5:10 pm
[D]r1bb1r1k/2qn2pp/ppp2p2/2P2P2/3PQ3/PB1N1R2/1P1B2PP/4R1K1 w - - 0 1

I was able to set up this same position on the 8X10 board, and

Gothic Vortex found ...Rh3....in 3 seconds!

That is faster than the Fidelity Mach 3 computer...that took 33 seconds to find Rh3.

RISC takes 70 sec.

and Novag Scorpio does not find it at all.... :roll:

If black plays......g7-g5

White mates with...Qe8!
Gothic Vortex Engine (Rare) :wink:
https://web.archive.org/web/20060330175 ... vortex.zip

zullil
Posts: 5682
Joined: Mon Jan 08, 2007 11:31 pm
Location: PA USA
Full name: Louis Zulli

Re: Testing Gothic Vortex.....Rh3 wins.

Post by zullil » Wed Jul 10, 2019 3:28 pm

GothicChessInventor wrote:
Sat Dec 01, 2007 4:30 am
rightrook wrote:Hi...Well, I was really comparing it to some of my test positions for my dedicated chess computerors...

And since I posted that position, I have found out that it is doing excellent...finding difficult moves that other programs require much time to find..

.thanks anyway... :D

Robert
The only thing that I did radically different to Crafty was to use my new piece values derived from my 80-square research:

http://www.GothicChess.com/80.pdf

I don't know what else could have Vortex outperform the software it was based on. This just seems counterintuitive to me.

Maybe "chess is easy" to Vortex now since it has been "working out" with an 80-square board with 2 new piece types

:)
Crafty finds mate-in-6 in less than one second. So I'm confused by "outperform".

Code: Select all

Crafty v25.2 (20 threads)

machine has 40 processors

White(1): mt = 0
parallel threads disabled.
White(1): hash = 8M
hash table memory = 8M bytes (512K entries).
White(1): phash = 1M
hash path table memory = 536K bytes (1K entries).
White(1): st = 1
search time set to 1.00.
White(1): setboard r1bb1r1k/2qn2pp/ppp2p2/2P2P2/3PQ3/PB1N1R2/1P1B2PP/4R1K1 w - - 0 1
White(1): disp

       +---+---+---+---+---+---+---+---+
    8  |<R>| . |<B>|<B>|   |<R>|   |<K>|
       +---+---+---+---+---+---+---+---+
    7  | . |   |<Q>|<N>| . |   |<P>|<P>|
       +---+---+---+---+---+---+---+---+
    6  |<P>|<P>|<P>| . |   |<P>|   | . |
       +---+---+---+---+---+---+---+---+
    5  | . |   |-P-|   | . |-P-| . |   |
       +---+---+---+---+---+---+---+---+
    4  |   | . |   |-P-|-Q-| . |   | . |
       +---+---+---+---+---+---+---+---+
    3  |-P-|-B-| . |-N-| . |-R-| . |   |
       +---+---+---+---+---+---+---+---+
    2  |   |-P-|   |-B-|   | . |-P-|-P-|
       +---+---+---+---+---+---+---+---+
    1  | . |   | . |   |-R-|   |-K-|   |
       +---+---+---+---+---+---+---+---+
         a   b   c   d   e   f   g   h

White(1): go
        time surplus   0.00  time limit 1.00 (1.00)
        depth     time       score   variation (1)
         19->   0.65/1.00    Mat06   1. Rh3 g5 2. Qe8 Qxh2+ 3. Rxh2 Rxe8
                                     4. Rxe8+ Nf8 5. Rxf8+ Kg7 6. Rg8#
        time=0.65(100%)  nodes=4900552(4.9M)  fh1=82%  pred=0  nps=7.5M
        chk=105.2K  qchk=149.4K  fp=2.3M  mcp=235.3K  50move=1
        LMReductions:  1/87.9K  2/54.5K  3/43.6K  4/31.3K  5/10.1K  6/608
        null-move (R):  3/167.1K  4/20.7K  5/763  6/3

mate in 6 moves.

White(1): Rh3
              time used:   0.65

User avatar
Nordlandia
Posts: 2476
Joined: Fri Sep 25, 2015 7:38 pm
Location: Sortland, Norway

Re: Testing Gothic Vortex.....Rh3 wins.

Post by Nordlandia » Wed Jul 10, 2019 3:39 pm

I bought a Gothic chess set from Ed Trice. I used the iron to flaten it out and it worked.

Image

Alternatively i can rotate the board and play 8x10 chess with standard piece set.

Image

Post Reply