SCID vs PC

You can enable the main board bindings by editing scid.gui (or tcl/end.tcl) and find the

  "#### Bindings for main board"

and uncomment the six lines there. ... Perhaps these bindings should be enabled by default.

I'm glad you're enjoying this project Ed, and thanks for the feedback. Scid truly is a great resource.

 I'm not sure if i can easily enable the above texture wrapping. Wish (Tcl/Tk) is very powerful, but can also be limiting to do things like this.

Hi Steven and thanks for all your work!

Any chance you could add cbh to scid tool from scidb's project?

Yes, a thousand thanks for all of the effort you've put into the program.  I hope it's something you've enjoyed doing and will continue to do for awhile. 

It's a great program.  I find myself learning more about it almost every time I use it!  

I was finally able to make my own chessboard in Scid vs PC.  Thanks for the help.

Has anyone figured out how to use the "Mark Tactical Exercise" feature?  I've checked that box in the Configure Annotations window but I've never seen it produce any results.  Truth be told, I'm not sure what this feature is supposed to do.

Can anyone help?

Amy

Will chessbase support Scid support?

That would be the right question.

Ask chessbase. 

Yes, this is a great tool of Gregor's. When i get a chance i'll post a linux i386 binary in Scid vs. PC support files.

I'm still learning about this one too.  

Yes, it *is* huge.

I had a love of Tcl and Tk before discovering this project, and since it really needed some serious maintenance, it's been a good match. It's a shame the dockable wiindows feature isn't included, but it really does have unresolveable issues, and is a little half-baked anyhow. A project this size really needs a quality code base to work properly.

 My next release won't be for a year i'd guess, but plans are to fix the last few bugs, and start getting my hands dirty with the database and a "Find Move Combos" feature (eg b5 Bxb5).

I've been playing around with this feature a little since your post, and I think I've figured out some of how it works. First, it only works with UCI engines, and when you start the automatic annotation of a game, the engine will evaluate using 4 principal variations. (So it won't evaluate quite as deeply as if it were only using 1 PV.)

The criteria it seems to use to find a tactical move for marking is that the move must change the evaluation by at least 2 pawns (so it helps to be annotating bad games, ha ha), and there must be only one PV that produces this swing of more than 2 pawns. If it finds 2 or more PVs that produce this > 2 pawn swing, then it will not mark the move as tactical.

When it finds a move that fits these criteria, then it will annotate the move with a code starting with 4 asterisks and some other stuff. For example, a code of ****D1 1.5->4.7 means that the evaluation went from 1.5 to 4.7. I think the single digit after the letter D is some kind of difficulty level.

Once the engine has marked up a game or games with tactical shots, you can use the "Training: Find best move" feature to go through the tactics.

From skimming the scid.gui file, it looks like the criteria for choosing tactical shots may be more complicated than I thought. But maybe I'm close?

Yes, I don't think I have done a computer analysis on any game with a 2-pawn swing.  My own games should provide sufficient fodder, however.  Lol

Thanks for your help.  I will try again.

Amy

Are any of you aware that there is a computer language developed for chess called Chess Query Language or CQL which is supposed to pick out specific positions from a data base. I know a little Javascript and VB but am not very good at programming at all. If someone would like to make a database using CQL it would be very interesting!

Frank,

Yes, I've heard of CQL.  I don't know a thing about it, but I've heard of it.  It might be a fun hobby/topic to get into someday.  (I have some prior programming experience.)

As a test, just now I loaded a database I have that contains 11.3 million games in it, in In SCID vs. PC.)  ("Scid" for the rest of this reply.)

It only took about five seconds to load.  (It's not a pgn file.  It's a file saved in Scid's format.)

I then copy and pasted a position that keen readers might be familiar with.  This position:

That's the final position of the FIRST game EVER where Fischer beat Spassky.  (3rd round game in Reykjavic, in '72)

I then asked Scid to search this 11.3 M database, for this exact position.  I wasn't sure if it would be in there, but it was.  Scid found it in about 24 seconds.  (And it didn't seem that long, with the progress bar moving on the screen, while it was searching.) In fact, as I knew, this database contains many duplicate games.  Scid confirmed it... this game was found 6 times!

But what I find more interesting that just finding specific positions, is something that to me is a lot harder... finding specific material or specific patterns.  For example, I just asked Scid to pull up, from my 11.3 M database, all of the games where two bishops were up against two knights.

This took longer... 41 seconds according to the progress bar.  But it found them... exactly 395,584 games.  And when I pull up a dozen or two games to check, sure enough... in each one at one point there is two bishops against two knights.  And what is cool is when I pull up each game, Scid initally goes straight to the first position in that game when the two bishops against two knights becomes true.  (As soon as White takes one of Black's bishops or knights or vice versa.)

Scid can also search for games where White or Black has an isolated queen pawn, or a Rook sacrifice on c3, or O-O vs O-O-O, for example. 

Pretty cool.

I don't see why not.  If you checked out this topic on an earlier page, you noticed you can ADD a new piece font to Scid vs. PC, simply by adding the font data to the scid.gui file, the file where the font data is stored. 

So it should then be possible, with the right tools, to take this data and edit it, with a font editor or font graphics editor or whatever. 

After editing the font you probably just change it back to a base 64 string, which is what I believe that font data in scid.gui is.

That's another thing I've wanted to do but never done... create my own font.  (Although I was thinking of letters and not chess pieces.)  I know there are a lot of letter font editors out there, but I've never played with one.

Convekta's products like Aquarium, Chess Assistant support CQL.

Anyway, one can download a stand-alone tool from CQL's website. There're a bunch of sample scripts along with the manual too.

It's very tough. The implementation is a little crude. Here's a quick guide

http://sourceforge.net/projects/scidvspc/files/support%20files/scid_chess_pieces.tgz/download

cbh2si4 linux binaries

How does one evaluate a position, in SCIDvsPC? I don't want to analyse, I only want to know if white is winning (or not) in a certain position. How should I do that?

The program won't know who is winning.   No program can determine that.  Only an engine can tell you that, and only after it has analyzed the position.  So even though you don't want to analyze, that's the only way to know who is winning. 

To do that...

Paste the position in, by first copying the FEN from an outside source, and then pasting it into the program.  (EDIT - PASTE FEN)

If you don't have the FEN, then EDIT - SETUP BOARD might be the best way to enter the position.  Of course, if you have th entire game score you could load that and then just go straight to your desired position.

Click the ANALYSIS ENGINES button.  Your #1 engine will then load and start anaylyzing that position.  A window will pop up showing the analysis, the engine info, an option to increase or decrease the number of lines you want displayed, etc.

Alternately, you can click TOOLS - ANALYSIS ENGINES and then choose from any of the engines you've already installed.  (I currently have about 25 of my favorite engines installed.)

Note that you can have more than one engine analyze the same position at the same time.