Help with Glaurung

[Ponti]

I'm using Glaurung (MacBook pro), imported a FEN position - black winning (Glaurung gives +7), but I can't play it against the engine because it resigns, even if this option in "preferences is set to disabled.
Is there a way to force it to play?

This is the kind of exercise called "win a winning position" my coach gives me very often. It's frustrating, Glaurung just don't let me win by checkmate, it resigns!

[Nimzovik]

If this is trully the case just download another free engine such as Rybka 1 .

[Milos]

If this is trully the case just download another free engine such as Rybka 1 .

You mean Fruit?

[michiguel]

If this is trully the case just download another free engine such as Rybka 1 .

You mean Fruit?

For goodness sake... is it necessary to hijack every single thread with Rybka and or Ippos??????

Miguel

[Milos]

For goodness sake... is it necessary to hijack every single thread with Rybka and or Ippos??????

Seams it's impossible here on CCC to write anything that has even the slightest negative connotation towards Rybka without being attacked by a bunch of Rybka lovers...

[Tord Romstad]

If this is trully the case just download another free engine such as Rybka 1 .

That won't help at all, for two different reasons: There is no Mac OS X version of Rybka 1, and UCI engines don't resign. Resigning is the responsibility of the GUI, what engine you are using makes no difference.

That Glaurung doesn't resign sounds like a clear bug. I don't recall having seen or fixed this bug, but when I tried it just now with the public Stockfish, I couldn't reproduce the problem. Stockfish plays on until checkmate when the "resign in hopeless positions" preference is unchecked. If you're running OS X 10.6, please upgrade to Stockfish (available from http://www.stockfishchess.com and try again. If you're still running 10.5 or earlier, I'm afraid there isn't much I can do for you right now, except advising you to try some other GUI.

[alpha123]

If this is trully the case just download another free engine such as Rybka 1 .

You mean Fruit?

For goodness sake... is it necessary to hijack every single thread with Rybka and or Ippos??????

Miguel

I do have to agree with Miguel on this one, but I wouldn't recommend downloading Rybka 1 (or 2, or 3, or 4) due to its questionable legality.

Peter

[Dann Corbit]

For goodness sake... is it necessary to hijack every single thread with Rybka and or Ippos??????

Seams it's impossible here on CCC to write anything that has even the slightest negative connotation towards Rybka without being attacked by a bunch of Rybka lovers...

Miguel's point was this:

Joe: "It's a lovely day, isn't it?"
Fred: "Unless the horrible Rybka has snatched away your joy in it forever!"
===========================================
Sally: "I can't get Toga to work in Winboard. What should I do?"
Randy: "Consider the evil crimes of Vasik and thank your lucky stars it wasn't Rybka you were trying to use!"

[bob]

If this is trully the case just download another free engine such as Rybka 1 .

That won't help at all, for two different reasons: There is no Mac OS X version of Rybka 1, and UCI engines don't resign. Resigning is the responsibility of the GUI, what engine you are using makes no difference.

That Glaurung doesn't resign sounds like a clear bug. I don't recall having seen or fixed this bug, but when I tried it just now with the public Stockfish, I couldn't reproduce the problem. Stockfish plays on until checkmate when the "resign in hopeless positions" preference is unchecked. If you're running OS X 10.6, please upgrade to Stockfish (available from http://www.stockfishchess.com and try again. If you're still running 10.5 or earlier, I'm afraid there isn't much I can do for you right now, except advising you to try some other GUI.

Isn't that just another flaw in this protocol? Why would the "GUI" resign. I wonder if the xorg developers (X11 windows system) will step up and say "wait, the GUI can't resign, that is the responsibility of the windows system itself" until the linux kernel guys step forward and say "no, xwindows should not do that, that's a kernel function.

The poor engine, which knows what is going on in the game gets to sit idly by where everyone else argues over who gets to resign.

Meanwhile, back at the ranch, the Lone Ranger, disguised as a pool table, gets his balls racked.

[michiguel]

If this is trully the case just download another free engine such as Rybka 1 .

That won't help at all, for two different reasons: There is no Mac OS X version of Rybka 1, and UCI engines don't resign. Resigning is the responsibility of the GUI, what engine you are using makes no difference.

That Glaurung doesn't resign sounds like a clear bug. I don't recall having seen or fixed this bug, but when I tried it just now with the public Stockfish, I couldn't reproduce the problem. Stockfish plays on until checkmate when the "resign in hopeless positions" preference is unchecked. If you're running OS X 10.6, please upgrade to Stockfish (available from http://www.stockfishchess.com and try again. If you're still running 10.5 or earlier, I'm afraid there isn't much I can do for you right now, except advising you to try some other GUI.

Isn't that just another flaw in this protocol? Why would the "GUI" resign. I wonder if the xorg developers (X11 windows system) will step up and say "wait, the GUI can't resign, that is the responsibility of the windows system itself" until the linux kernel guys step forward and say "no, xwindows should not do that, that's a kernel function.

The poor engine, which knows what is going on in the game gets to sit idly by where everyone else argues over who gets to resign.

Meanwhile, back at the ranch, the Lone Ranger, disguised as a pool table, gets his balls racked.

It is not flawed for SMK and Shredder, which split roles between the interface and the engine. It is flawed for anyone who plug its UCI engine into shredder and cannot control this. In other words, in theory, the protocol it is not flawed per se. For instance, Fruit connects through polyglot with WB. There, the decision to resign etc. is done by polyglot. From the point of view of Fabien L., he controls the whole thing and UCI serves its purpose. There is no flaw in his design. But, anybody else that uses polyglot, surrenders their engine to Fabien decisions.

However... this is what I consider a flaw: In UCI, time management is in the engine's hand, but it shouldn't to be consistent. If the interface (polyglot or whatever) is the one that knows and control books etc., the interface is the one that should tell the engine: "Time remaining x seconds, I "recommend" to think y seconds".

My point is, if the interface (not necessarily the "graphic" interface, e.g. polyglot) is going to control certain things, it should also control whatever is related to those, or nothing at all. UCI has not split the tasks appropriately, IMHO. It is half baked.

It may not be a bad design to have a "polyglot-like" adapter that controls, book, TBs, when to clear hash, when to learn, what type o TM to have etc. and one engine file that is told just to search. If the same programmer do both, the adapter and the engine, things could be better organized. But UCI did not reach this point, with every detail. That is why it needs to "guess" when it is out of book and other things. I found it is awkward to be controlled but you need to guess what to do sometimes. Sounds like UCI engines have crazy bosses

Miguel