Thermopylay Marathon 2011 (live!)

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.
Daniel Shawul
Posts: 3724
Joined: Tue Mar 14, 2006 10:34 am
Location: Ethiopia
Contact:

Re: Thermopylay Marathon 2011 (live!)

Post by Daniel Shawul » Tue Feb 08, 2011 1:47 pm

Second game of Nebiyu-Spartacus was a clear case of passed pawn runners on both wings. And dumb nebiyu doesn't understand it :( His hoplites at the centre were not advanced enough. His pieces were not developed at all and was defending passers all game long.. No wonder HG has +2 for developing spartans. I will get the position when the game finishes.

[D]2k1k2q/2r1p3/2pnn1P1/1Pp1p1PP/P1Ppp3/8/Q7/R1B1KB1R b KQ - 1 27 [/D]
You would never see such a terrible game

[Event "Thermopylae Marathon 2011"]
[Site "SCHAAKPC"]
[Date "2011.02.08"]
[Round "7.3"]
[White "Spartacus 0.23 / 6"]
[Black "Nebiyu 1.1 / 6"]
[Result "1-0"]
[TimeControl "40/1440"]
[Variant "spartan"]
[Annotator "1. -1.88 1... +0.40"]
[Number "42"]

1. Nc3 {-1.88/15 28} Lc6 {+0.40/19 27} 2. e4 {-1.51/15 33} Hge5
{+0.50/18 31} 3. d4 {-1.52/16 34} Lf6 {+0.50/19 36} 4. Nf3 {-1.13/16 30}
Hfe6 {+0.70/19 18} 5. g4 {-1.03/15 31} Lxe4 {+1.00/20 26} 6. Nxe4
{-0.89/16 9} Hxe4 {+1.20/16 42} 7. g5 {-0.88/16 36} Lg6 {+1.30/18 37} 8. h4
{-0.93/15 36} Wf7 {+1.30/16 32} 9. Nd2 {-0.79/15 36} Hed5 {+1.20/18 29} 10.
h5 {-0.60/16 36} Lf5 {+0.50/17 21} 11. g6 {-0.59/15 36} Wh8 {+0.40/20 33}
12. c3 {-0.24/15 36} Hxd4 {+0.40/20 28} 13. Qf3 {-0.34/15 31} Le6
{+0.50/19 23} 14. Nxe4 {-0.31/14 36} Hhf5 {+0.60/19 29} 15. Qg2
{-0.14/14 36} Hg4 {+0.40/17 39} 16. f3 {+0.52/14 36} Hd5 {+0.40/19 39} 17.
fxg4 {+0.59/14 36} He5 {+0.50/17 43} 18. Nc5 {+1.10/15 36} Cd6
{+0.30/18 38} 19. b4 {+1.15/14 33} Lf6 {+0.60/16 40} 20. g5 {+1.21/14 34}
Le6 {+0.10/17 38} 21. a4 {+1.39/14 37} Hc6 {+0.40/16 38} 22. Nxe6
{+1.66/15 37} Cexe6 {+0.50/18 40} 23. c4 {+1.86/14 34} He4 {+0.80/18 38}
24. Qf3 {+0.89/15 37} Ke8 {+0.90/17 38} 25. Qa3 {+0.74/15 35} Gc7
{+1.20/18 21} 26. b5 {+0.88/14 36} Hac5 {+1.40/16 20} 27. Qa2 {+1.06/14 38}
Hd3 {+1.30/15 1:09} 28. a5 {+0.52/15 38} Kcd7 {+1.00/16 38} 29. Rb1
{+1.34/15 38} Ce4 {-0.20/16 2:10} 30. b6 {+2.09/17 39} Cxb6 {-0.70/17 30}
31. axb6 {+3.66/16 30} Gb8 {-2.60/20 20} 32. Bh3 {+5.47/15 34} He2+
{-3.50/18 31} 33. Kf2 {+5.55/15 41} Ce6 {-3.50/18 29} 34. Bxe6
{+7.22/16 41} Hxe6 {-3.60/19 19} 35. Qa7 {+9.69/16 40} Wg7 {-7.40/20 18}
36. Qxb8 {+11.40/15 42} Wf5 {-7.40/19 41} 37. Qxe8+ {+15.23/16 43} Kxe8
{-13.10/22 1:22} 38. b7 {+17.06/16 34} Wd3+ {-13.30/21 26} 39. Kf3
{+17.64/16 49} Kd7 {-16.20/21 10} 40. b8=Q {+16.81/14 54} We4+
{-17.70/21 55} 41. Kxe2 {+22.77/16 31} Wd3+ {-18.40/20 28} 42. Ke3
{+22.83/15 32} Hb4 {-20.50/19 21} 43. Rxb4 {+27.02/14 34} Wc2+
{-292.80/21 36} 44. Kf2 {+319.92/15 26} Wd3+ {-293.60/22 42} 45. Kg1
{+319.93/16 35} Wxb4 {-295.20/23 18} 46. Qb7+ {+319.94/17 35} Kd6
{-296.00/25 21} 47. Qxb4+ {+319.95/26 31} Kc7 {-296.80/26 1.7} 48. g7
{+319.96/100 35} Hf4 {-297.60/24 0.1} 49. g8=Q {+319.97/100 32} Hde3
{-298.40/22 0.1} 50. Qa5+ {+319.98/100 35} Kd6 {-299.20/21 0.1} 51. Qgd8#
{+319.99/100}
{Xboard adjudication: Checkmate} 1-0

User avatar
hgm
Posts: 23496
Joined: Fri Mar 10, 2006 9:06 am
Location: Amsterdam
Full name: H G Muller
Contact:

Re: Thermopylay Marathon 2011 (live!)

Post by hgm » Tue Feb 08, 2011 2:12 pm

This game went very typical for games between Spartacus and Nebiyu when the latter plays the Spartans. (Both in a style that could be described as extremely opportunistic.) But centralyzing the Hoplites is not always a losing strategy. Quite often the flank attacks peter out and Nebiyu can force a winning break-through in the center. Overall the 40/5 games between the two had a score that is very close to 50%.

I have the feeling both engines are a bit unbalanced. For instance, with the Persians Spartactus almost never castles.In stead it stampedes its Pawn forward like mad. I don't think this is good strategy. But because it does work so well against Nebiyu, I am reluctant to change it. I have given Spartacus a PST for Hoplites that hardly encourages centering them, and I give sizable bonuses for Persian 'passers' (i.e. no Hoplites that could step in its way. This produced more balanced play with the Spartans, as it puts a break on creating holes in the wings. But when the opponent does crate such holes,Spartacus just cannot resist jumping into them.

I think the most fatal decision by Nebiyu was to park its Warlord on h8.

Standings

After the first cycle, Spartacus and Nebiyu are tied for the first place, with a sizable lead on numbers 3/4 (also tied) Fairy-Max / Oberon. One couldconsider this an indication I did not make their time-odds handicap large enough. But what is the point of being better, if you are not allowed to win a bit? :wink: Nevertheless, for the next (=last) qualifier cycle I will increase their time-odds factor to 8. It seems they should still be able to handle that; now they scored 75% and 85% out of their non-mutual games. Now I have to install the updates, and find a suitable starting position.

Code: Select all

Cross table, sorted by score percentage, Buchholz, SB

                              Sp Ne Ob Fa Ch Sj Ca
 1. Spartacus 0.23 / 6        ## =1 10 =1 11 01 11   75%   9.0 ( 66.0,  47.0)
 2. Nebiyu 1.1 / 6            =0 ## 11 11 0= 11 11   75%   9.0 ( 66.0,  44.3)
 3. Oberon                    01 00 ## 01 10 01 11   50%   6.0 ( 72.0,  28.0)
 4. Fairy-Max  4.8R           =0 00 10 ## 11 0= 11   50%   6.0 ( 72.0,  26.3)
 5. ChessV (Spartan)          00 1= 01 00 ## 11 10   46%   5.5 ( 73.0,  31.5)
 6. Sjaak 92                  10 00 10 1= 00 ## 11   46%   5.5 ( 73.0,  26.0)
 7. Catalyst 3                00 00 00 00 01 00 ##    8%   1.0 ( 82.0,   5.5)

Daniel Shawul
Posts: 3724
Joined: Tue Mar 14, 2006 10:34 am
Location: Ethiopia
Contact:

Re: Thermopylay Marathon 2011 (live!)

Post by Daniel Shawul » Tue Feb 08, 2011 2:26 pm

I agree. 8x handicap is better because with the longer time control Nebiyu & spartacus are less bound to make tactical mistakes due to time trouble.
I would say place 3-6 is up for grabs and with 8x handicap they might join the the top. Who knows with the new catalyst too? ChessV actually beat Nebiyu. It seems I can not overlook eval anymore.
Can you send me Nebiyu 16:xx if you have it, I need it please :) dshawul at yahoo.com

User avatar
Evert
Posts: 2923
Joined: Fri Jan 21, 2011 11:42 pm
Location: NL
Contact:

Re: Thermopylay Marathon 2011 (live!)

Post by Evert » Tue Feb 08, 2011 3:08 pm

Daniel Shawul wrote:I agree. 8x handicap is better because with the longer time control Nebiyu & spartacus are less bound to make tactical mistakes due to time trouble.
I would say place 3-6 is up for grabs and with 8x handicap they might join the the top. Who knows with the new catalyst too? ChessV actually beat Nebiyu. It seems I can not overlook eval anymore.
Yes, very exciting after one round. Almost anything could still happen for the next round.
I'm still annoyed at the second game Sjaak played against Spartacus (it lost on time), it didn't look as though it was doing too badly. :(
Oh well, hopefully it won't happen again.

User avatar
hgm
Posts: 23496
Joined: Fri Mar 10, 2006 9:06 am
Location: Amsterdam
Full name: H G Muller
Contact:

Re: Thermopylay Marathon 2011 (live!)

Post by hgm » Tue Feb 08, 2011 3:39 pm

I have a problem with Nebiyu now. When I start a test game with a forced opening line ("/lpf=startpos.pgn") against Oberon, it produces a false illegal-move claim. Looking in the debug, I find it chokes on the commands

computer
name Oberon 0.04


to which it replies

setup ....
hello Oberon!
Error (illegal move): 0.04


I am not sure why we did not have that problem before. It seems to only occur when forcing in openng moves.

Can you fix that quickly, Daniel?

Daniel Shawul
Posts: 3724
Joined: Tue Mar 14, 2006 10:34 am
Location: Ethiopia
Contact:

Re: Thermopylay Marathon 2011 (live!)

Post by Daniel Shawul » Tue Feb 08, 2011 3:47 pm

Yes I can. I used to have it as "Error (unknown command) " but changed it to "Error (Illegal move) " after your sugestion if you remeber. Nebiyu takes only the first string as name and discards the rest as unknown commands. It is hard to figure out how long a name of an engine would be, may be until the next known command. But this is complicated. Anyway this is changed to "unknown command" now and I am sending it now.

Edit: Uploaded now
https://sites.google.com/site/dshawul/N ... ects=0&d=1

User avatar
hgm
Posts: 23496
Joined: Fri Mar 10, 2006 9:06 am
Location: Amsterdam
Full name: H G Muller
Contact:

Re: Thermopylay Marathon 2011 (live!)

Post by hgm » Tue Feb 08, 2011 3:58 pm

You should read until the end of the line!

Catalyst has the infamous WinBoard bug: it fails to count the forced moves. I had an opening line of 8 ply, so when Catalyst gets anothe minute added to its clock on move 40 (I was doing 40/1), it uses it all upon move 41-44. Then, when it has to do 40 more moves in 0.9 sec, because it does not get the 1 min it counted on, it forfeits because of the illegal move 'unknown'.

Can you fix that, Richard?

User avatar
hgm
Posts: 23496
Joined: Fri Mar 10, 2006 9:06 am
Location: Amsterdam
Full name: H G Muller
Contact:

Re: Thermopylay Marathon 2011 (live!)

Post by hgm » Tue Feb 08, 2011 4:11 pm

Sjaak has exactly the same bug as Catalyst. It also does not count the forced moves.

Can you fix that, Evert?


Or would it be better if we just abandon the idea of playing from a different position in the second cycle? Most engines randomize, I think.

Daniel Shawul
Posts: 3724
Joined: Tue Mar 14, 2006 10:34 am
Location: Ethiopia
Contact:

Re: Thermopylay Marathon 2011 (live!)

Post by Daniel Shawul » Tue Feb 08, 2011 4:25 pm

That may take longer for them to fix. I can add a randomization on the root move ordering just in case it is needed. Quick too :)

User avatar
hgm
Posts: 23496
Joined: Fri Mar 10, 2006 9:06 am
Location: Amsterdam
Full name: H G Muller
Contact:

Re: Thermopylay Marathon 2011 (live!)

Post by hgm » Tue Feb 08, 2011 4:34 pm

Are you sure? It is a bit tricky, with regard to window bounds. If you can do it, I think it would be a good idea.

Fairy-Max randomizes its first 4 moves, Spartacus randomizes 7 cP throughout the game, ChessV randomizes through its (small) book, and Pawel said he had built randomizing in Oberon too. Catalyst was replaced by a completely different version, so not much chance of repeat games there.

I don'tknow about Sjaak, but if Nebiyu randomizes tcould at worst be the only non-randomizing engine, and that would be no problem.

Post Reply