Hard test set talkchesscom-2012

[Vinvin]

Here's the complete PGN (still 2-3% to investigate because too easy or unsure).

Only the PGN until now : http://home.scarlet.be/vincentlejeune/h ... a-july.pgn

There are some very very hard, probably not found by strong playing programs before 2030.

Firsts are easy, latest are hard but it's still some position that are not sorted.

A lot of position come from already known set (hardCCC, MGV77, Arasan,
100-brillante, Nightmare, Albillo). Easiest has been removed.

It probably still some double, sorry for that, It will be corrected later.

Enjoy !

Vincent.

[beachknight]

Thank you Vincent. best,

[Paloma]

Vincent, thanks for this.

I didn't know about albillo.epd

[gleperlier]

Very nice job

[jdart]

Nice collection. I'd like to have info about the positions' source games (if any) though. I think most of the ones in the first 60 labeled "Test Positon" are from the Arasan test suite but the comment fields that indicate the source are not preserved. (Same for a lot of the other positions)

--Jon

[Vinvin]

Nice collection. I'd like to have info about the positions' source games (if any) though. I think most of the ones in the first 60 labeled "Test Positon" are from the Arasan test suite but the comment fields that indicate the source are not preserved. (Same for a lot of the other positions)

--Jon

All position from Arasan suite are tagged like this: [Comment "bm Nf6+; id "arasan14.197"; c0 "Serban-Schuster, World Corr Ch 26 Final, 2010";"]

I've to reorganize all of this, I found some bugs today already.

I'll publish another intermediate beta version in 2 weeks and the finale version in 4 weeks (I hope... ).

[jdart]

> [Comment "bm Nf6+; id \"arasan14.197\"; c0 \"Serban-Schuster, World Corr Ch 26 Final, 2010\";"]

The problem is, this is a non-standard PGN tag so it is almost invisible in most UIs (ChessBase for example will show it only if you select the game and do "Edit Game Data"). Better to put Serban and Schuster in the White and Black fields and fill in the Event tag with World Corr Ch Final. That's a pain I know but it means viewers will show the source info not just "Test position".

--Jon

[jdart]

This one is interesting:

[Event "Panow-Taimanow, Moskau 1952"]
[Site "?"]
[Date "????.??.??"]
[Round "?"]
[White "Der brillante Schachzug 040"]
[Black "Freibauer"]
[Result "*"]
[Annotator "Walter Eigenmann"]
[SetUp "1"]
[FEN "3r2k1/p6p/b2r2p1/2qPQp2/2P2P2/8/6BP/R4R1K w - - 0 1"]
[PlyCount "7"]
[Source "Walter Eigenman"]
[SourceDate "2011.11.01"]

1. Rxa6 $3 (1. Ra4 $6 Bc8 $13) (1. Rfb1 $6 Bc8 $13) 1... Rxa6 2. d6 Qxe5 (2...
Qxd6 3. Bd5+ $18) 3. fxe5 Ra5 (3... Kf7 4. Re1 $18) 4. Bd5+ $18 *

In the line after 3. .. Kf7, Houdini thinks Bd5+ is better (Black really can't avoid this move):

118: Der brillante Schachzug 040 - Freibauer, Panow-Taimanow, Moskau 1952
3r4/p4k1p/r2P2p1/4Pp2/2P5/8/6BP/5R1K w - - 0 1

Code: Select all

Analysis by Houdini 1.5a w32 _15a_w32:

1. +-  (2.31): 4.Bd5+ Ke8 5.Re1 Rd7 6.c5 Ra5 7.Bc6 Rxc5 8.Bxd7+ Kxd7 9.e6+ Kxd6 10.e7 Rc8 11.e8Q Rxe8 12.Rxe8 Kc5 13.Ra8 Kb6 14.Kg2 a5 15.Kf3 Kb5 16.Ra7 h6 17.Rb7+ Kc5 18.Rh7 a4 19.Rxh6 g5 20.Ra6 Kb4 21.Rb6+ Kc4 22.Rc6+ Kb4 

2. +/-  (1.37): 4.Re1 Kf8 5.c5 Ra5 6.c6 Rc5 7.Ra1 Ke8 8.h4 Ra8 9.c7 Rc8 10.Rxa7 Kd7 11.Bb7 R8xc7 12.dxc7 Rc1+ 13.Kg2 Rxc7 14.Kh3 Ke7 15.Kg3 h6 16.Kf4 Rc4+ 17.Ke3 Ke6 18.Ra5 Rc7 19.Bg2 g5 20.hxg5 hxg5 

In the main line Bd5+ is not as strong. I think White winds up with a large positional gain but I don't see a clear win. One line is Bd5+ Kg7 Re1 Rc5 e6 Rxd6 e7 Rc8 e8=Q Rxe8 Rxe8 and White has a Bishop vs. 3 pawns, with a better endgame.

--Jon

[zullil]

This one is interesting:

[Event "Panow-Taimanow, Moskau 1952"]
[Site "?"]
[Date "????.??.??"]
[Round "?"]
[White "Der brillante Schachzug 040"]
[Black "Freibauer"]
[Result "*"]
[Annotator "Walter Eigenmann"]
[SetUp "1"]
[FEN "3r2k1/p6p/b2r2p1/2qPQp2/2P2P2/8/6BP/R4R1K w - - 0 1"]
[PlyCount "7"]
[Source "Walter Eigenman"]
[SourceDate "2011.11.01"]

[d]3r2k1/p6p/b2r2p1/2qPQp2/2P2P2/8/6BP/R4R1K w - - 0 1

1. Rxa6 $3 (1. Ra4 $6 Bc8 $13) (1. Rfb1 $6 Bc8 $13) 1... Rxa6 2. d6 Qxe5 (2...
Qxd6 3. Bd5+ $18) 3. fxe5 Ra5 (3... Kf7 4. Re1 $18) 4. Bd5+ $18 *

In the line after 3. .. Kf7, Houdini thinks Bd5+ is better (Black really can't avoid this move):

118: Der brillante Schachzug 040 - Freibauer, Panow-Taimanow, Moskau 1952
[d]3r4/p4k1p/r2P2p1/4Pp2/2P5/8/6BP/5R1K w - - 0 1

Code: Select all

Analysis by Houdini 1.5a w32 _15a_w32:

1. +-  (2.31): 4.Bd5+ Ke8 5.Re1 Rd7 6.c5 Ra5 7.Bc6 Rxc5 8.Bxd7+ Kxd7 9.e6+ Kxd6 10.e7 Rc8 11.e8Q Rxe8 12.Rxe8 Kc5 13.Ra8 Kb6 14.Kg2 a5 15.Kf3 Kb5 16.Ra7 h6 17.Rb7+ Kc5 18.Rh7 a4 19.Rxh6 g5 20.Ra6 Kb4 21.Rb6+ Kc4 22.Rc6+ Kb4 

2. +/-  (1.37): 4.Re1 Kf8 5.c5 Ra5 6.c6 Rc5 7.Ra1 Ke8 8.h4 Ra8 9.c7 Rc8 10.Rxa7 Kd7 11.Bb7 R8xc7 12.dxc7 Rc1+ 13.Kg2 Rxc7 14.Kh3 Ke7 15.Kg3 h6 16.Kf4 Rc4+ 17.Ke3 Ke6 18.Ra5 Rc7 19.Bg2 g5 20.hxg5 hxg5 

In the main line Bd5+ is not as strong. I think White winds up with a large positional gain but I don't see a clear win. One line is Bd5+ Kg7 Re1 Rc5 e6 Rxd6 e7 Rc8 e8=Q Rxe8 Rxe8 and White has a Bishop vs. 3 pawns, with a better endgame.

--Jon

[jdart]

I see this was discussed when it first came out:

http://talkchess.com/forum/viewtopic.ph ... 41&t=28552

and it does look like Rxa6 gives a big + score eventually but it takes some programs a long time.

--Jon