What's wrong in this EPD?

Discussion of chess software programming and technical issues.

Moderator: Ras

lucasart
Posts: 3243
Joined: Mon May 31, 2010 1:29 pm
Full name: lucasart

Re: What's wrong in this EPD?

Post by lucasart »

Aser Huerga wrote:
lucasart wrote: On the contrary, the extra blank is essential. Parameters must always be separated by a single space. I would say that

Code: Select all

3r2k1/6p1/p3p2p/4Pp2/2P3PP/3pB1n1/PP6/2KR4 w  - f6 0 1 
is correct, and

Code: Select all

3r2k1/6p1/p3p2p/4Pp2/2P3PP/3pB1n1/PP6/2KR4 w  -f6 0 1 
is incorrect.

If your GUI says otherwise then the bug is in the GUI !

Here is the standard definition of a FEN
http://en.wikipedia.org/wiki/Forsyth%E2 ... s_Notation
It works for me either way coping the FEN to the clipboard and paste in Arena. I simply had the idea that it could have to do with the dash, sorry ...
Arena probably tries to improve the user experience, byaccepting an "almost correct" user input, and translating it internally into a correct FEN. I imagine if you don't put the "0 1" at the end, Arena will accept, and send some default values to the engine. Arena probably also translates sequences of spaces into a single space, for the same reason. It's all good programming practice!

Conclusion Arena 1 Chessbase 0 :D
Sven
Posts: 4052
Joined: Thu May 15, 2008 9:57 pm
Location: Berlin, Germany
Full name: Sven Schüle

Re: What's wrong in this EPD?

Post by Sven »

lucasart wrote:Conclusion Arena 1 Chessbase 0 :D
No, as Jouni already noted it's: Komodo 0 :-)

Code: Select all

Komodo version 3 32bit
uci
id name Komodo32 3 32bit
id author Don Dailey, Larry Kaufman
option name Ponder type check default true
option name Hash type spin default 64 min 0 max 4096
option name OwnBook type check default true
option name MultiPV type spin default 1 min 1 max 99
option name Book File type string default komodo.bin
option name use lmr type check default true
option name null move pruning type check default true
option name deterministic type check default false
option name Best Line type check default true
option name King Safety Aggressiveness type spin default 6 min 0 max 10
option name move overhead milliseconds type spin default 20 min 0 max 30000
option name Drawscore type spin default -5 min -50 max 50
uciok
position fen 3r2k1/6p1/p3p2p/4Pp2/2P3PP/3pB1n1/PP6/2KR4 w - f6 0 1 moves e5f6
go infinite
info depth 1
info depth 1 time 0 nodes 2 score mate 1 nps 19999 pv e1g1
info nodes 2
bestmove e1g1
So in the given position, after 1.exf6 e.p. Komodo instantly plays an illegal move "e1g1" and even thinks it is mate in 1 ply.

Furthermore, according to UCI protocol it should not send "bestmove" in infinite search until receiving "stop", but that's another story.

Sven
User avatar
hgm
Posts: 28461
Joined: Fri Mar 10, 2006 10:06 am
Location: Amsterdam
Full name: H G Muller

Re: What's wrong in this EPD?

Post by hgm »

Well, of course ruling 'illegal move' without telling what that move was should also be counted as a pretty severe GUI bug, so the 0 for Chessbase (and Arena) is quite deserved as well... This whole thread would not have existed when the GUI would have said "{Illegal move by Komodo: e1g1} 1-0"...
Sven
Posts: 4052
Joined: Thu May 15, 2008 9:57 pm
Location: Berlin, Germany
Full name: Sven Schüle

Re: What's wrong in this EPD?

Post by Sven »

hgm wrote:Well, of course ruling 'illegal move' without telling what that move was should also be counted as a pretty severe GUI bug, so the 0 for Chessbase (and Arena) is quite deserved as well... This whole thread would not have existed when the GUI would have said "{Illegal move by Komodo: e1g1} 1-0"...
Agreed. So here is the final outcome of our "illegal move after e.p. move" tourney:

WinBoard 1
ChessBase 0
Arena 0
Komodo 0

8-)

Sven
Jouni
Posts: 3819
Joined: Wed Mar 08, 2006 8:15 pm
Full name: Jouni Uski

Re: What's wrong in this EPD?

Post by Jouni »

Actually in Chessbase this resulted as some hours of lost testing time, but in Arena tournament continues nornally and it's was difficult to note any special (Arena restarts engine after each game). BTW is Arena only GUI, that has nice shut-down PC after finished tournament -feature?

Jouni
User avatar
hgm
Posts: 28461
Joined: Fri Mar 10, 2006 10:06 am
Location: Amsterdam
Full name: H G Muller

Re: What's wrong in this EPD?

Post by hgm »

Jouni wrote:BTW is Arena only GUI, that has nice shut-down PC after finished tournament -feature?
WinBoard automatically shuts down after a match or tourney when you started the tourney from the command line (which historically was the only way to do matches), if you suppressed the exit popup to report the result (a persistent setting). That includes startups through clicking the tournament file, which are in fact invoking WinBoard with command-line arguments through a shortcut.

For interactively started tournaments (i.e. through the menu), I assumed users would want to remain interactive, and I added code to keep WinBoard up in that case. But that normally only holds for the instance you used to start the tourney. Any helper processes used for concurrently playing the tourney games would terminate automatically, as these are started by clicking the tourney file. When a user would want auto-close he can simply quit the WinBoard instance used to start the tourney, and start another one by double-clicking the tourney file.

You think it would be better to include an 'auto-close' checkbox (unticked by default) in the Tournament dialog? It is always a tricky trade-off between simplicity and flexibility, and I don't want to make the dialog too user unfriendly by overwhelming the user with all kinds of options almost no one ever uses...

Image
User avatar
Giorgio Medeot
Posts: 52
Joined: Fri Jan 29, 2010 2:01 pm
Location: Ivrea, Italy

Re: What's wrong in this EPD?

Post by Giorgio Medeot »

hgm wrote:
Jouni wrote:BTW is Arena only GUI, that has nice shut-down PC after finished tournament -feature?
WinBoard automatically shuts down after a match or tourney
I think he was talking about having WB shut down the PC, not itself. Could be useful e.g. when you launch a lengthy tournament which happens to end during night time, and you don't want your PC to waste electric power sitting there idle.
It could be generalized if WB could run an "after tournament" batch script.

Giorgio
User avatar
hgm
Posts: 28461
Joined: Fri Mar 10, 2006 10:06 am
Location: Amsterdam
Full name: H G Muller

Re: What's wrong in this EPD?

Post by hgm »

Oh, I missed that! :oops:

I did not even know itwas possile for user programs to swith off the PC. :shock:

You are right: an -afterTourney option would be the logical and most flexible way to implement this. I already provided an -afterGame option, to satisfy my immediate need for updatng the cross table on a broadcastig website. Adding similar -afterRound, -afterCycle and -afterTourney options would be a logical next step, but so far I had no real application for them.
User avatar
Don
Posts: 5106
Joined: Tue Apr 29, 2008 4:27 pm

Re: What's wrong in this EPD?

Post by Don »

Jouni wrote:OK I found it. I first was sure this is Chessbage bug, then Rybka bug, but it's Komodo bug! It crashes after e.p. move! For to do list Don :)

Jouni
There is indeed a bug - it will be fixed before the next release. As you can probably guess it is only a fen translation bug.

Don
User avatar
abik
Posts: 824
Joined: Fri Dec 01, 2006 10:46 pm
Location: Mountain View, CA, USA
Full name: Aart Bik

Re: What's wrong in this EPD?

Post by abik »

hgm wrote:Well, of course ruling 'illegal move' without telling what that move was should also be counted as a pretty severe GUI bug
Do I get a point? :-)

Code: Select all

[Event "Chess for Android Tournament"]
[Site "United States"]
[Date "2011.12.20"]
[Round "1"]
[White "BikJump v2.1P (32-bit)"]
[Black "Komodo32 3 AB"]
[Result "1-0"]
[PlyCount "1"]
[TimeControl "-"]
[Device "Galaxy Nexus"]
[Setup "1"]
[FEN "3r2k1/6p1/p3p2p/4Pp2/2P3PP/3pB1n1/PP6/2KR4 w - f6 0 1"]

1. exf6 {[7/16]67/e6e5 0.694} 1-0 {black resigned:illegal move e1g1}