Banksia GUI released

[Wilhelm]

How do I get the opening classification in the notation? I have only "end of opening" here.

BSG auto matches all opening positions with ones from ECO list and it keeps the longest one to show. The ECO list is quite limited on depth thus it typically can’t cover longer openings (out of range). Since it is an auto process, users don’t have to do anything nor change it.

I can't find any ECO list and it never shows any classification. I tried many different book and pgn files.
Do I need a ECO list?

A bit ambiguous since I am not sure what you need:

- If you mean about BSG: no, it is auto and you can't interfere with the process of finding ECO codes for openings, you don't need to do any thing
- If you want to learn, collect, just Google it, there are many websites, listed a few:

https://en.wikipedia.org/wiki/Encyclopa ... s_Openings
https://www.chessgames.com/chessecohelp.html

Download:
https://github.com/jdart1/arasan-chess/ ... r/book/eco

I mean I always see {end of opening} in the notation and nothing else. Does it only work in human games?
In all screenshots I can see the eco codes also when two engines are playing.

[Nigel Robson]

The ERET test position suite contains 5 alternative move tests which Banksia skips, perhaps because they are not understood. For example
8/6pp/5k2/1p1r4/4R3/7P/5PP1/5K2 w - - am Ke2; id "ERET 098 - Endgame";
where the correct solution is any move other than Ke2. Would it be possible to add the am capability?
Thanks, Nigel

I have added it to my to-do list. Thanks.

BTW, how should we count the result if an engine hits an "am" move then change to other moves at the end? A good or a bad result?

A good result.

[Wilhelm]

Here's a screenshot to see what I mean. As I said I tried different books and pgn files and also all posible settings.
A new clean install also shows no eco codes, always end of opening.

[Torom]

Did you see any problem/bug when BSG plays Polyglot book moves?

BSG has understood deeply about that kind of books. Some book parameters such as top percentage is designed for Polyglot (as well as similar ones, such as online books), worked based on weights of moves.

Furthermore, BSG can generate Polyglot books from game databases.

Do I see any problems? BSG just doesn't play the best moves, although it's supposed to. The problem is the assumption that the moves are always correctly sorted. You wrote here that BSG assumes this:

[...]
However, when probing books for playing, BSG supposes those moves are in the right order thus it won't sort, just pick up as the conditions. In this case, it picks Bc4 as the top one, not Bb5 since Bc4 is the first.
[...]

This assumption is not correct in reality. And I honestly can't find anyone else on the internet who assumes this. I have shown the implementation of python-chess, for example, where the move that is saved first with the highest weight is considered the best.

I have shown here the example of the M13.4 opening book, the book is sorted so that BSG plays the worst and not the best moves. This book cannot be used with BSG, but it can be used with any other polyglot implementation I know.

With M13.4 it is extreme, but the problem occurs with pretty much every book, first we found it with the Perfect2021 book, but it also occurs with Cerebellum and Chucaro. There, however, it is not as extreme, but rather on a smaller scale, for example, Italian is played instead of Spanish or the other way round.

[AlexChess]

Torom I understand what you are explaining to Pham, but strangely MK13.4 on BSG is doing really well on my Superblitz... (I'll publish all the games tonight, last updated games almost ended)

[Cornfed]

@Alexchess - The crosstable is so small on the website...is there a way to view it better or even download it?

Nevermind - found the pgn file, downloaded it into Chessbase then used the crosstable function.
Suprized to find Dragon 2.5.1 way down at the 22nd spot.

[BrendanJNorman]

Here's a screenshot to see what I mean. As I said I tried different books and pgn files and also all posible settings.
A new clean install also shows no eco codes, always end of opening.

That's a BEAUTIFUL theme. Did you make it?

Would you share it? Makes the GUI very attractive imo.

[Wilhelm]

Here's a screenshot to see what I mean. As I said I tried different books and pgn files and also all posible settings.
A new clean install also shows no eco codes, always end of opening.

That's a BEAUTIFUL theme. Did you make it?

Would you share it? Makes the GUI very attractive imo.

Yes, the bitmaps was made by me.
http://descent.at/files/Chess/Bitmaps.zip
Layout and design seems to be saved inside Banksia so I don't know how to share.

[AlexChess]

@Alexchess - The crosstable is so small on the website...is there a way to view it better or even download it?

Nevermind - found the pgn file, downloaded it into Chessbase then used the crosstable function.
Surprized to find Dragon 2.5.1 way down at the 22nd spot.

You can enlarge the browser view. On Chrome full screen 200% is perfectly readable. 73 engines is the biggest BSG crosstable I can publish.

Yes is a big challenge for all engines to stay on top 20.

[huanhuan]

Hi Pham

This is very unsightly.
=> https://prnt.sc/1xm4dx3

Hope you can fix it in the next release.

Thank you,