It has an opening book of around 30,000 moves and the ability to announce a mate from a distance of 72 moves.
It does not support "countdown" time controls, only fixed time per move. It is based on the last version of Crafty with rotated bitboards, so it goes back to about 2004 or 2005 I think.
The tablebases can be probed in RAM during the search, so that if you set up a 6-piece position with a trade within the bounds of the horizon of search, it will be able to forecast the distance of the extended mate.
Note: This was an intermediate build. There is some "king safety" code that was in the process of being re-written that was not fully tested and also not completed. That is why you will see a text file "King Safety Values" being created and appended to when you launch the program. From what I can tell, this weakens the program considerably.
Thank you for sharing, this is a strong engine got a draw with my latest version with unbalance material in the ending using tc 1 minute per move, opening book is interesting too.
Interesting game. Here is the final position right before white's last move:
It's a strange sort of fortress position if white wants to keep things static. Black can't get the rook to the back rank since the knight holds d1 and b1 and and move of the rook to the a file white has Bb2. I set my version on Infinite Search to see how it would play if it could search deeper like your program does so well.
Notice it was able to hit the tablebases at ply 16. That is because I have the full 5-piece set loaded on my machine. That means it searched deeply enough with the extensions to encounter 1 5-piece position. It counts the "database hits" on the far right. It returns the "mate in x" information to the search tree also, so if it hits enough positions, it can forecast a win from a great distance.
Does Gothic Vortex able to read a log file for example and execute the move
based from the log? I am thinking of playing this gui vs other engines/winboard automatically.
It can play that way, yes, but I forget how to do it. It has a menu item called "Matches" where it will write its move to a file. If your program can read the file, and reply back to the file, the move will be made on the Vortex board.
That is an interesting feature. It would be easy to make a 'Vortex2WB' adapter that loads as a WinBoard engine, starts up Vortex, and then plays a game by reading from or writing to the file. It might still be needed to start Vortex by hand and set it to use the file and set the right time per move. But at least the game would then conduct itself after that.
Note that in case you can still recover the Vortex source code, my offer to hack a WinBoard interface in it that could run Vortex directly as a WB engine still stands. A few years ago I did manage to do that for ChessV, (which was developed on MSVC), and in the end it wasn't that difficult. What did the trick was compiling it as a console application rather than a GUI application, and include a function main() which would call WinMain() (which is the program that would be called when you link it as GUI application). The difference is that in that case it keeps stdin and stdout open.
Then I let main() just start another thread for reading WB commands from stdin, before the main thread launched ChessV's own GUI through calling WinMain. The WB commands would occasionally contain a move to be played, and I would just translate the from- and to-square to coordinates for a mouse click, and call ChessV's mouse event handler with those, so it would be thinking a user entered that move. I only had to locate the code where ChessV would play a move itself, to add a print statement there that would also send the move to stdout, so WinBoard would get it. ChessV would just be running with its own GUI, but at the same time communicate its moves to WinBoard, and allow the latter to operate its mouse and critical menus.
I could probably use the exact same code for main() and the input thread. The only thing to figure out is how to translate square coords to mouse coords and how to call the mouse event handler with them, and where to insert the print statement to print the move.
hgm wrote:That is an interesting feature. It would be easy to make a 'Vortex2WB' adapter that loads as a WinBoard engine, starts up Vortex, and then plays a game by reading from or writing to the file. It might still be needed to start Vortex by hand and set it to use the file and set the right time per move. But at least the game would then conduct itself after that.
Here is one way to test to see if it works. Launch Vortex, and a copy of Vortex, in the same directory.
Under the MATCHES menu from the second program to launch, select START MATCH BUT PLAY BACK.
From the first program, select CONFIGURE MATCH and just hit the OK button. Then select START MATCH AND SEND FIRST WHITE MOVE.
It should generate some debugging dialog boxes, and drawing moves on each others' boards.
hgm wrote:That is an interesting feature. It would be easy to make a 'Vortex2WB' adapter that loads as a WinBoard engine, starts up Vortex, and then plays a game by reading from or writing to the file. It might still be needed to start Vortex by hand and set it to use the file and set the right time per move. But at least the game would then conduct itself after that.
Here is one way to test to see if it works. Launch Vortex, and a copy of Vortex, in the same directory.
Under the MATCHES menu from the second program to launch, select START MATCH BUT PLAY BACK.
From the first program, select CONFIGURE MATCH and just hit the OK button. Then select START MATCH AND SEND FIRST WHITE MOVE.
It should generate some debugging dialog boxes, and drawing moves on each others' boards.
Thank you I was able to do the match. We hope HGM will be able to create an adaptor for it.
