Myracle GUI

[Max]

Thank you Martin for the lightning-fast change and a wonderful holiday!

[Wilhelm]

I wish you a relaxed holyday and don't dream about weird users and ideas...
Wilhelm

[tapio]

Wish you nice holidays - you probably read this - admit it, nobody can resist checking the Internet stuff

[Andres Valverde]

I wish you a relaxed holyday and don't dream about weird users and ideas...
Wilhelm

How did you dettach the tourney tab ?

[Wilhelm]

Oh, I didn't. Martin is only dreaming

[mar]

I see you guys were having some fun here...


yes I couldn't resist - mostly because VT had several false positive "detections" by anon AVs sharing the same detection engine

so I've uploaded build 114 (rebuilt) that finalizes 3d support (Layout/3D board) as a bonus:
- multipv arrows with numbers in 3d mode
- piece color tint can be set for 3d pieces
- added directional shadow (default on) - note that this is a bit demanding so may cause some lag
in 4k since I can only use a single core to render 3D

[Frank Quisinsky]

A quick update for the captain!

Yes, ESET Internet Security Ultimate do it with your build 113.
Build 114 is clear for take-off, wind is good, on runway "must have" without that ESET problem.

Best and thanks!
Frank

[mar]

good to know, Frank! ESET here too btw
but I guess a local build is always fine, so didn't spot any problem

ok, I made some additional improvements to 3d mode, so... one more...

build 115 is up:
- parallel 3d rendering if core budget permits
- 3d mode arrows and target hints now xray through 3d pieces
- new 3d mode default: 50% tint and no blob shadows

seems VT clean, so I don't expect any problems

[PHILOCHESS]

Hi Martin,

I would like to ask how Myracle handles the following situation. If an engine recognizes a dead position in which legal moves are still available and responds with bestmove 0000, would Myracle declare and record the game as a draw? Or would it treat bestmove 0000 as an engine error, an illegal move, a resignation, or a loss?

I would also like to know:

- Does Myracle independently detect dead positions, including those caused by permanent blockades rather than only elementary material cases?
- If the opposing engine does not recognize the dead position and would otherwise continue playing, would Myracle still terminate the game?
- If bestmove 0000 is not the appropriate way for an engine to report this terminal state, what behavior would you recommend?
- How would Myracle record the result and termination reason in the PGN?

Thank you very much for your help.

[mar]

hi Robert,

first - I'm not quite sure what you mean by "dead" position. I assume you mean a dead draw?

I would like to ask how Myracle handles the following situation. If an engine recognizes a dead position in which legal moves are still available and responds with bestmove 0000, would Myracle declare and record the game as a draw? Or would it treat bestmove 0000 as an engine error, an illegal move, a resignation, or a loss?

I'm not sure what you mean by dead position - the GUI will never let the engine search in a terminal position where no moves are possible
bestmove 0000 (or null move) is treated as an illegal move and whoever plays it loses the game

that being said, an engine should always search and output a valid move whenever it's asked to, assuming there is at least one valid move in that position

- Does Myracle independently detect dead positions, including those caused by permanent blockades rather than only elementary material cases?

no blockades, only trivial material configurations. if syzygy is setup, it also detects and adjudicates syzygy win/draw if enabled in adjudication settings

- If the opposing engine does not recognize the dead position and would otherwise continue playing, would Myracle still terminate the game?

once myracle adjudicates a game, it unloads the engines so technically the engine cannot continue playing, if that's what you are asking.

- If bestmove 0000 is not the appropriate way for an engine to report this terminal state, what behavior would you recommend?

bestmove 0000 should ideally never be sent by the engine, like I stated above. an (uci) engine should only send bestmove 0000 if no bestmove was found, i.e. there are 0 legal moves. some engines also send bestmove 0000 if time is up while doing depth 1 search, which I think is actually a bug and it's better to potentially lose on time. not that it matters much, of course.
note that myracle will not ask the engine to search any such position with 0 legal moves

I'm not sure what you mean with terminal state, if by dead position you mean a dead draw, then the engine should simply evaluate the position as 0.0 (along with a move)

- How would Myracle record the result and termination reason in the PGN?

like I said, this depends on adjudication settings. could be draw by insufficient material, a checkmate or a stalemate or draw by tablebase adjudication.
if an engine sends bestmove 0000 in a valid position that has at least 1 legal move, it'll lose on the spot by illegal move