keoki010 wrote:Jeremy, both gran versions have problems with this game fraqment. They seem to get stuck on d=1 for some reason. If you go back and forth sometimes it'll display a full pv.
[GameID "0"]
[SetUp "1"]
[WhiteElo "0"]
48... e4 49. Rxf7 Re6 50. Nc4 e3 51. Nxe3 {^13 ^10 } {i get mate in 43 with
rondo^13 ^10 } {^13 ^10 } {David Evans^13 ^10 } {^13 ^10 } {I see mate in 42
Rybka on I7 950.^13 ^10 } {^13 ^10 } {Rubén Cómes} 51... Rxe3 52. Rf6 a5
{Black resigns.2011.02.11.} 1-0
Observed only one problem: I place a position different from the initial position of a game, to study and on the first play does not return the first line. Once placed moves the problem disappears.
8/8/4k3/8/8/4K3/8/4BB2 w - - 0 1
snino64 wrote:Observed only one problem: I place a position different from the initial position of a game, to study and on the first play does not return the first line. Once placed moves the problem disappears.
8/8/4k3/8/8/4K3/8/4BB2 w - - 0 1
snino64 wrote:Observed only one problem: I place a position different from the initial position of a game, to study and on the first play does not return the first line. Once placed moves the problem disappears.
8/8/4k3/8/8/4K3/8/4BB2 w - - 0 1
OK, I see this in Scid. In ChessBase, it works properly. Weird...
jb
It does exactly the same thing with Houdini, by the way. Basically, it skips the first variation. If you pause and restart the engine, the variation appears. This appears to be a bug in Scid.
keoki010 wrote:Jeremy, both gran versions have problems with this game fraqment. They seem to get stuck on d=1 for some reason. If you go back and forth sometimes it'll display a full pv.
[GameID "0"]
[SetUp "1"]
[WhiteElo "0"]
48... e4 49. Rxf7 Re6 50. Nc4 e3 51. Nxe3 {^13 ^10 } {i get mate in 43 with
rondo^13 ^10 } {^13 ^10 } {David Evans^13 ^10 } {^13 ^10 } {I see mate in 42
Rybka on I7 950.^13 ^10 } {^13 ^10 } {Rubén Cómes} 51... Rxe3 52. Rf6 a5
{Black resigns.2011.02.11.} 1-0
What's the starting FEN here?
Jeremy
Sorry about that here it is:
7k/4rp1P/p2N4/4p3/8/8/KP6/5R2 b - - 0 48
keoki010 wrote:Jeremy, both gran versions have problems with this game fraqment. They seem to get stuck on d=1 for some reason. If you go back and forth sometimes it'll display a full pv.
[GameID "0"]
[SetUp "1"]
[WhiteElo "0"]
48... e4 49. Rxf7 Re6 50. Nc4 e3 51. Nxe3 {^13 ^10 } {i get mate in 43 with
rondo^13 ^10 } {^13 ^10 } {David Evans^13 ^10 } {^13 ^10 } {I see mate in 42
Rybka on I7 950.^13 ^10 } {^13 ^10 } {Rubén Cómes} 51... Rxe3 52. Rf6 a5
{Black resigns.2011.02.11.} 1-0
What's the starting FEN here?
Jeremy
Sorry about that here it is:
7k/4rp1P/p2N4/4p3/8/8/KP6/5R2 b - - 0 48
Thanks, I'm seeing this in ChessBase, too. I'll get this solved ASAP.
keoki010 wrote:Jeremy, both gran versions have problems with this game fraqment. They seem to get stuck on d=1 for some reason. If you go back and forth sometimes it'll display a full pv.
[GameID "0"]
[SetUp "1"]
[WhiteElo "0"]
48... e4 49. Rxf7 Re6 50. Nc4 e3 51. Nxe3 {^13 ^10 } {i get mate in 43 with
rondo^13 ^10 } {^13 ^10 } {David Evans^13 ^10 } {^13 ^10 } {I see mate in 42
Rybka on I7 950.^13 ^10 } {^13 ^10 } {Rubén Cómes} 51... Rxe3 52. Rf6 a5
{Black resigns.2011.02.11.} 1-0
What's the starting FEN here?
Jeremy
Sorry about that here it is:
7k/4rp1P/p2N4/4p3/8/8/KP6/5R2 b - - 0 48
Thanks, I'm seeing this in ChessBase, too. I'll get this solved ASAP.
Jeremy
Try this one.
What was happening: assuming that there are no tablebase hits in the root search (in this case, obviously so -- there are too many pieces on the board), a recursive search will be made of checks, captures and queen promotions. In some of these lines, tablebase hits occurred, causing the search to falsely shut down when it returned. So that was a bug.
However, it seems stupid to ignore the information in the tablebases that's reached over this extended root search. The attached version simply doesn't search the tablebases except at the base of the line. I would like to experiment a bit with using this information to potentially improve the accuracy and speed of searching (there is no reason to continue to search a line if its outcome is known). I'll drop another version on you when I've puzzled that out.
