PHCEL

Discussion of computer chess matches and engine tournaments.

Moderators: hgm, Harvey Williamson, bob

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
User avatar
Rubinus
Posts: 949
Joined: Thu Jan 18, 2007 3:05 pm
Location: Prague
Contact:

Re: PHCEL

Post by Rubinus » Sun Aug 26, 2018 6:43 pm

finale_H_ph_2018_005, 2m+2s

Code: Select all

1   Crafty 25.3 WB         ** 10 1˝ 1˝ ˝1 ˝1 01 10   9.0/14  61.75
2   Zappa Mexico II        01 ** 10 10 10 11 11 01   9.0/14  57.50
3   Deep Onno 1-2-70       0˝ 01 ** 1˝ 1˝ ˝˝ 1˝ 11   9.0/14  56.00
4   Toga II 4.0            0˝ 01 0˝ ** ˝1 01 1˝ 11   8.0/14
5   The Baron 3.42 64-bit  ˝0 01 0˝ ˝0 ** ˝1 ˝˝ 1˝   6.5/14
6   Loop M1-T              ˝0 00 ˝˝ 10 ˝0 ** ˝1 1˝   6.0/14
7   Scorpio 2.8.6          10 00 0˝ 0˝ ˝˝ ˝0 ** ˝˝   4.5/14
8   Chess Tiger 2007.1     01 10 00 00 0˝ 0˝ ˝˝ **   4.0/14
games
Very balanced, except the last two weak. The question is, would Scorpio be better on the machine (i7 / 4cp) better with the MCTS turned off. Forehead equalized to much, it does not give Crafty and Zappa much chance in the final G.
Next G (RR 2x, today night): Nirvanachess 2.4 (3180), Protector 1.9.0 (3173), Cheng 4.39 (3059), Crafty 25.3 WB (3042), Spike 1.4 (3041), ExChess 7.97bx64 (3035), Zappa Mexico II (3023), Rybka 2.3 LK (3012).

User avatar
Rubinus
Posts: 949
Joined: Thu Jan 18, 2007 3:05 pm
Location: Prague
Contact:

Re: PHCEL

Post by Rubinus » Mon Aug 27, 2018 1:48 pm

finale_G_ph_2018_005, 2m+2s

Code: Select all

1   Nirvanachess 2.4x64popcnt  ** ˝1 1˝ 11 11 10 11 ˝˝  11.0/14
2   Protector 1.9.0            ˝0 ** ˝1 ˝˝ ˝1 0˝ ˝1 11   8.5/14
3   Zappa Mexico II            0˝ ˝0 ** ˝1 ˝0 1˝ ˝1 ˝1   7.5/14  46.25
4   Crafty 25.3 WB             00 ˝˝ ˝0 ** ˝1 1˝ ˝1 1˝   7.5/14  44.50
5   Rybka 2.3 LK               00 ˝0 ˝1 ˝0 ** 10 ˝1 11   7.0/14
6   Cheng 4.39                 01 1˝ 0˝ 0˝ 01 ** 1˝ 00   6.0/14
7   ExChess 7.97bx64           00 ˝0 ˝0 ˝0 ˝0 0˝ ** 11   4.5/14
8   Spike 1.4                  ˝˝ 00 ˝0 0˝ 00 11 00 **   4.0/14
games
Both favorites advanced, Nirvanachess with a lead, which also shows good chances for a higher floor. Somehow Spike did not succeed, yet ExChess does not have more threads support and it ended up better.Next F (RR 4x, probably tomorrow night): Nirvanachess 2.4x64popcnt (3191), Protector 1.9.0 (3171), Pedone 1.7 (3123), Vajolet2 2.5 (3112), Ivanhoe version 1945a (3105), Deep HIARCS 14 WCSC (3092), Deuterium v2018.1.35.514 (3040), Deep Junior Yokohama (2999).

User avatar
Rubinus
Posts: 949
Joined: Thu Jan 18, 2007 3:05 pm
Location: Prague
Contact:

Re: PHCEL

Post by Rubinus » Thu Aug 30, 2018 1:14 pm

finale_F_ph_2018_005, 2m+2s

Code: Select all

1   Nirvanachess 2.4x64popcnt  **** 0˝˝˝ 0˝01 1˝˝1 01˝1 1˝˝1 1111 ˝111  19.0/28
2   Pedone 1.7                 1˝˝˝ **** 011˝ 1˝00 01˝˝ 01˝1 11˝˝ 11˝1  17.5/28
3   Protector 1.9.0            1˝10 100˝ **** 0˝˝˝ 1011 1˝˝0 ˝1˝1 111˝  17.0/28
4   Ivanhoe version 1945a      0˝˝0 0˝11 1˝˝˝ **** ˝˝˝0 ˝˝10 ˝111 11˝1  16.5/28
5   Deep HIARCS 14 WCSC        10˝0 10˝˝ 0100 ˝˝˝1 **** 01˝˝ 1101 ˝˝11  15.0/28
6   Vajolet2 2.5               0˝˝0 10˝0 0˝˝1 ˝˝01 10˝˝ **** 11˝1 1˝01  14.5/28
7   Deuterium v2018.1.35.514   0000 00˝˝ ˝0˝0 ˝000 0010 00˝0 **** 1˝˝˝   6.5/28
8   Deep Junior Yokohama       ˝000 00˝0 000˝ 00˝0 ˝˝00 0˝10 0˝˝˝ ****   6.0/28
games
In the past series, Pedone played under that rating, now on the contrary, so he overcame the favored Protector. Nirvanachess confirmed his commitment and went on for the second time in a row. The next will probably be too difficult. Ivanhoe 1945a is probably not the best version of this clone. Hiarcs did not lose so much, but we will not stop the time. It is unfortunate to develop it, and today, in the name of the author, we will find ourselves in discussions about the originality of today's engines. My point is that if the clone is better than the original, nobody is interested in the origin. Whether Rybka came into being anyway, it has become a powerful impetus and the shift of today's peak to such a 700 ELO against the reality of 2004 is at least a third of its merit. A similar case as Hiarcs is Junior. Occasionally he plays an interesting part, but this once one of the fastest engines is now being computed by competitors, some of the modern selective techniques are missing there.
Next E (RR 4x, probably today night): Texel 1.08a11 (3251), Critter 1.6a 64-bit (3223), Chiron 4 64bit (3219), Bouquet 1.8 x64 (3218), Nirvanachess 2.4x64popcnt (3196), Naum 4.6 (3164), Nemorino 5.00 (3158), Pedone 1.7 (3137). Last PH-ELO.

User avatar
Rubinus
Posts: 949
Joined: Thu Jan 18, 2007 3:05 pm
Location: Prague
Contact:

Re: PHCEL

Post by Rubinus » Mon Sep 03, 2018 1:11 pm

finale_E_ph_2018_005, 2m+2s

Code: Select all

1   Critter 1.6a 64-bit        **** ˝0˝0 1˝˝1 ˝10˝ ˝1˝1 ˝˝˝1 ˝111 1˝1˝  18.0/28
2   Texel 1.08a11              ˝1˝1 **** 0˝˝˝ 10˝˝ 0˝1˝ 01˝1 11˝˝ ˝111  17.5/28
3   Chiron 4 64bit             0˝˝0 1˝˝˝ **** 101˝ 11˝˝ ˝˝1˝ 11˝1 ˝˝˝˝  17.0/28
4   Nirvanachess 2.4x64popcnt  ˝01˝ 01˝˝ 010˝ **** 1˝01 001˝ ˝˝˝1 111˝  15.5/28
5   Nemorino 5.00              ˝0˝0 1˝0˝ 00˝˝ 0˝10 **** ˝1˝0 111˝ 1˝11  14.5/28
6   Bouquet 1.8 x64            ˝˝˝0 10˝0 ˝˝0˝ 110˝ ˝0˝1 **** 1˝0˝ ˝˝1˝  13.5/28
7   Pedone 1.7                 ˝000 00˝˝ 00˝0 ˝˝˝0 000˝ 0˝1˝ **** 1˝˝˝   8.5/28
8   Naum 4.6                   0˝0˝ ˝000 ˝˝˝˝ 000˝ 0˝00 ˝˝0˝ 0˝˝˝ ****   7.5/28
games
The favorites went up, even though Texel had a weaker start. And, according to the list, they have a good chance of up one floor.
Next D (RR 6x, probably today night): Texel 1.08a11 (3255), Critter 1.6a 64-bit (3234), Equinox 3.30 x64mp (3211), Laser 1.6 (3199), Senpai 2.0 (3155), Wasp 3.0 (3153), Deep Fritz 13 (3093), Alfil 15.8.16 (3055). Laser is probably undervalued and it's probably complicated by the favorites.
Last PH-ELO.

User avatar
Rubinus
Posts: 949
Joined: Thu Jan 18, 2007 3:05 pm
Location: Prague
Contact:

Re: PHCEL

Post by Rubinus » Fri Sep 07, 2018 7:58 pm

finale_D_ph_2018_005, 2m+2s

Code: Select all

1   Laser 1.6            ****** ˝˝010˝ 1˝101˝ ˝˝0˝˝0 ˝1˝˝˝1 ˝11˝11 11111˝ ˝˝˝111  27.5/42
2   Critter 1.6a 64-bit  ˝˝101˝ ****** 1˝˝0˝0 ˝˝˝1˝1 ˝˝0˝1˝ 1˝˝111 ˝˝˝˝11 ˝1011˝  26.0/42
3   Texel 1.08a11        0˝010˝ 0˝˝1˝1 ****** ˝˝˝˝˝˝ 11˝˝00 10100˝ 1˝˝˝11 111111  24.5/42
4   Equinox 3.30 x64mp   ˝˝1˝˝1 ˝˝˝0˝0 ˝˝˝˝˝˝ ****** 1˝00˝1 1˝1˝˝1 ˝001˝1 ˝011˝1  23.5/42
5   Senpai 2.0           ˝0˝˝˝0 ˝˝1˝0˝ 00˝˝11 0˝11˝0 ****** 011˝0˝ ˝1˝˝˝˝ 1˝˝0˝1  21.0/42
6   Deep Fritz 13        ˝00˝00 0˝˝000 01011˝ 0˝0˝˝0 100˝1˝ ****** ˝001˝˝ 1010˝˝  15.5/42
7   Wasp 3.0             00000˝ ˝˝˝˝00 0˝˝˝00 ˝110˝0 ˝0˝˝˝˝ ˝110˝˝ ****** ˝0˝˝˝0  15.0/42  309.75
8   Alfil 15.8.16        ˝˝˝000 ˝0100˝ 000000 ˝100˝0 0˝˝1˝0 0101˝˝ ˝1˝˝˝1 ******  15.0/42  299.25
games
New laser, unlike the swiss, has confirmed strength and won. Similarly, Critter finally proved he was a bit better than his older 1.4.
Next C (RR 6x, probably today night): Andscacs 0.94 (3297), Xiphos 0.3.11 BMI2 (3255), Laser 1.6 (3250), Mars 3.35 x64 (3249), Fritz 16-phcel (3246), Critter 1.6a 64-bit (3237), Hannibal 1.7 x64 (3228), DeepSaros_eXp rev.6 (3212). The clear favorite is the Andsacs, whose current rating is undervalued. On the other hand, Xiphos is probably revalued and perhaps a little Laser. I would advise "RybkaFritz".
Last PH-ELO.

User avatar
Rubinus
Posts: 949
Joined: Thu Jan 18, 2007 3:05 pm
Location: Prague
Contact:

Re: PHCEL

Post by Rubinus » Mon Sep 10, 2018 9:32 pm

finale_C_ph_2018_005, 2m+2s

Code: Select all

1   Andscacs 0.94        ****** ˝0˝01˝ 11˝˝˝˝ ˝˝1˝˝˝ ˝1˝˝˝˝ 110111 ˝11˝1˝ 1˝1˝1˝  27.5/42
2   Laser 1.6            ˝1˝10˝ ****** 00˝˝˝˝ ˝10˝˝1 ˝˝1˝1˝ ˝˝˝˝˝1 10˝11˝ 11˝˝11  25.5/42
3   Fritz 16-phcel       00˝˝˝˝ 11˝˝˝˝ ****** ˝0˝˝˝˝ 0˝1˝1˝ 1˝˝˝1˝ ˝0˝˝˝˝ ˝0111˝  22.5/42  457.25
4   Critter 1.6a 64-bit  ˝˝0˝˝˝ ˝01˝˝0 ˝1˝˝˝˝ ****** ˝1˝10˝ 1˝˝1˝0 ˝01˝1˝ 1˝˝0˝1  22.5/42  456.25
5   Xiphos 0.3.11 BMI2   ˝0˝˝˝˝ ˝˝0˝0˝ 1˝0˝0˝ ˝0˝01˝ ****** ˝˝1˝˝0 ˝˝˝˝˝˝ 0˝11˝1  19.5/42
6   Mars 3.35 x64        001000 ˝˝˝˝˝0 0˝˝˝0˝ 0˝˝0˝1 ˝˝0˝˝1 ****** 0˝˝11˝ ˝˝˝˝1˝  18.0/42
7   DeepSaros_eXp rev.6  ˝00˝0˝ 01˝00˝ ˝1˝˝˝˝ ˝10˝0˝ ˝˝˝˝˝˝ 1˝˝00˝ ****** ˝˝˝˝0˝  17.5/42
8   Hannibal 1.7 x64     0˝0˝0˝ 00˝˝00 ˝1000˝ 0˝˝1˝0 1˝00˝0 ˝˝˝˝0˝ ˝˝˝˝1˝ ******  15.0/42
games
Andsacs as expected and surprisingly the second Laser, with a fairly large lead. Rybka Fritz hesitated at the start when after the third was the seventh and catch up the loss failed. Critter kept the chances long, but at the end of the second period he failed.
Next B (RR 8x, probably today night): Booot 6.3.1_x64_popcnt (3359), Deep Shredder 13 x64 (3343), Andscacs 0.94 (3334), Gull_18062311_x64_bmi2 (3313), Fizbo 2 (3301), Arasan 21.0 (3297), Laser 1.6 (3281), Deep Rybka 4.1 x64 (3199). The only thing that seems certain is the last place of the old Rybka. And it's only ten years since her top position seemed so unshakeable that some authors resigned to further development. It will be an interesting struggle. Shredder ages and did not go straight to Group A for the first time, but he has nothing. Nominated versions of Booot, Andsacs, Gull and actually Arasan have played so few games so far, so the estimate of their strength is only approximate. On Fizbo, Laser and probably Arasan will be too strong, but the top four can rank in any way.
Last PH-ELO.

User avatar
Rubinus
Posts: 949
Joined: Thu Jan 18, 2007 3:05 pm
Location: Prague
Contact:

Re: PHCEL

Post by Rubinus » Mon Sep 17, 2018 7:00 pm

finale_B_ph_2018_005, 2m+2s

Code: Select all

1   Deep Shredder 13 x64    ******** ˝˝˝˝˝˝˝˝ ˝101˝˝˝0 ˝˝˝˝11˝˝ 11˝˝˝01˝ 1˝˝˝˝˝˝˝ ˝1˝111˝˝ 1˝1111˝1  35.5/56
2   Booot 6.3.1_x64_popcnt  ˝˝˝˝˝˝˝˝ ******** 1˝1010˝˝ ˝˝˝˝˝1˝0 ˝˝˝˝˝˝˝1 ˝˝101˝˝˝ 10111˝11 11111˝˝1  35.0/56
3   Fizbo 2                 ˝010˝˝˝1 0˝0101˝˝ ******** ˝1˝˝˝˝˝1 ˝˝˝˝˝˝11 ˝11˝˝˝˝˝ 1˝11˝˝˝0 ˝11111˝1  34.5/56
4   Andscacs 0.94           ˝˝˝˝00˝˝ ˝˝˝˝˝0˝1 ˝0˝˝˝˝˝0 ******** ˝˝˝0˝1˝1 ˝˝˝1˝˝˝1 ˝1˝˝˝1˝0 ˝1˝0˝11˝  29.0/56  775.25
5   Gull_18062311_x64_bmi2  00˝˝˝10˝ ˝˝˝˝˝˝˝0 ˝˝˝˝˝˝00 ˝˝˝1˝0˝0 ******** ˝˝1˝˝˝0˝ ˝1˝11˝11 ˝˝1˝1˝˝1  29.0/56  751.75
6   Laser 1.6               0˝˝˝˝˝˝˝ ˝˝010˝˝˝ ˝00˝˝˝˝˝ ˝˝˝0˝˝˝0 ˝˝0˝˝˝1˝ ******** ˝˝˝˝˝˝01 ˝˝0˝˝˝11  25.5/56
7   Deep Rybka 4.1 x64      ˝0˝000˝˝ 01000˝00 0˝00˝˝˝1 ˝0˝˝˝0˝1 ˝0˝00˝00 ˝˝˝˝˝˝10 ******** 11˝˝˝˝01  20.5/56
8   Arasan 21.0             0˝0000˝0 00000˝˝0 ˝00000˝0 ˝0˝1˝00˝ ˝˝0˝0˝˝0 ˝˝1˝˝˝00 00˝˝˝˝10 ********  15.0/56
games
The Andsacs for the second time in a row, Sredder and Boot advance. Surprisingly, Fizbo was holding it to the end.
Next A (RR 12x, start 14.9.): Stockfish 010818 64 BMI2 (3727), Raubfisch X38d3b (3552), Houdini 6.02 Pro x64 (3529), Komodo 12.1.1 64-bit (3447), Ethereal 10.55 (PEXT) (3373), Fire 7.1 x64 popcnt (3371), Booot 6.3.1_x64_popcnt (3357), Deep Shredder 13 x64 (3345). Stockfisch was a good Swiss, but it is not enough to have it against the updated Raubfisch clone. The position of Houdini and Komodo is clear, the only question is the loss at the top. Etheral and possibly Booot are probably overestimated.
Last PH-ELO.

Finals And about halfway through.

Code: Select all

1   Raubfisch X38d3b          ****** ˝˝˝˝˝˝ 1˝˝101 101˝˝0 1˝11˝1 111111 ˝111˝1 11˝1˝1  31.5/43
2   Stockfish 010818 64 BMI2  ˝˝˝˝˝˝ ****** 1˝˝˝˝˝ ˝˝˝11˝ 11˝1˝˝ 1˝111˝ 1˝˝11˝ 11111˝  30.5/43
3   Houdini 6.02 Pro x64      0˝˝010 0˝˝˝˝˝ ****** ˝˝˝1˝˝ 1˝11˝˝ ˝1˝1˝1 111111 ˝˝1111  28.5/43
4   Komodo 12.1.1 64-bit      010˝˝1 ˝˝˝00˝ ˝˝˝0˝˝ ****** 1˝˝˝˝˝ 1˝11˝1 ˝˝˝˝11 11˝1˝1  26.0/43
5   Fire 7.1 x64 popcnt       0˝00˝0 00˝0˝˝ 0˝00˝˝ 0˝˝˝˝˝ ****** 1˝1˝˝1 ˝˝˝1˝1 ˝˝˝1˝1  19.5/43
6   Deep Shredder 13 x64      000000 0˝000˝ ˝0˝0˝0 0˝00˝0 0˝0˝˝0 ****** 1˝˝1˝1 ˝˝0˝˝1  13.0/43
7   Booot 6.3.1_x64_popcnt    ˝000˝0 0˝˝00˝ 000000 ˝˝˝˝00 ˝˝˝0˝0 0˝˝0˝0 ****** ˝˝1˝˝0  11.5/43  238.00
8   Ethereal 10.55 (PEXT)     00˝0˝0 00000˝ ˝˝0000 00˝0˝0 ˝˝˝0˝0 ˝˝1˝˝0 ˝˝0˝˝1 ******  11.5/43  213.75

Post Reply