Hello,
25 years ago when Chess Informator published the most important chess books
and Chessbase started there was no other way to find games than to have a
classification system like ECO or an opening tree like Chessbase. Today the
help pages of Chessbase tells us not to use them anymore and use the
position search mechanism instead. Performance of computers grew much faster
than the number of games in databases and so searching the games for a
position is much faster than searching in opening trees.
I am sure no Profi is using ECO anymore.
Gerd Lorscheid
-Ursprüngliche Nachricht-
Von: Gregor Cramer [mailto:rema...@gmx.net]
Gesendet: Samstag, 19. April 2014 14:43
An: f...@libero.it; scid-users@lists.sourceforge.net
Betreff: Re: [Scid-users] Scid.eco
Hi Fulvio,
I've very dumb about the eco codes.
The ECO codes are not very useful, but some people are believing in this
code, especially the so called Profis, and so it seems that it is
unavoidable to display these unfortunate ECO codes.
For example, the A26 code in the link you posted:
every game that starts with 1. c4 e5 2. Nc3 Nc6 3. g3 g61. c4 e5 2.
Nc3 Nc6 3. g3 g6 is ECO A26?
Every game starting with 1.c4 e5 2.Nc3 Nc6 3.g3 g6 4.Bg2 Bg7 5.d3 should
be assigned to A26, and also all games derived from this move sequence.
And if so, what is the meaning of the second line 4. Bg2 Bg7 5.d3?
1.c4 e5 2.Nc3 Nc6 3.g3 g6 4.Bg2 Bg7 5.d3 is one line, I think that your
email reader is not showing the line breaks correctly. Please have a look
into the file scid.eco. Note that a two-liner
A25 English: Closed, 5.d3
1.c4 e5 2.Nc3 Nc6 3.g3 g6 4.Bg2 Bg7 5.d3 *
has to be read as one line (1. ECO code; 2. name of this line; 3.
'normalized' move order of this line; 4. result *).
About the A00 - A06 code: how transpositions are managed?
That's the unclear thing in ECO. So ChessBase assigns 1.f4 e5 2.d4 to A02,
and 1.d4 e5 2.f4 to A40, but Scidb assigns both lines to A02. I think that
the handling of ChessBase is confusing, Scidb is going a different way. The
inventor of this ECO code, Sahovski Informator, does not give any rules
about the handling of move transpositions. Therefore it's important that the
user can overrule the automatic assignment of the code. Scid, and also
Scidb, allows this.
To classify a game it's used the longest matching line, or it's
required to build the final positions of ECO lines and check if the
game reached that position?
This is unclear, ChessBase is using both methods, it depends on the line,
but Scidb is unexceptional using a position match (the move transposition
independent way). The current algorithm of Scid is a position match. Scid is
traversing the main line of the game backwards until a position match with
an ECO position occurs.
I would like to have consistent ECO classification between scid and
scidb
Shane has invented his own ECO classifications like A00a, A00b, and so on,
but this is unique to Scid, and Scidb does not use this. Scidb is using a
private ECO classification table. Furthermore Scidb's handling is to search
for the ECO forward (not backward), because the classification will be used
for the position search acceleration, a backward search is not possible in
this case. And currently I'm overworking the ECO table in a way that a
forward search will find all move transpositions (played in practice). So I
think that a 100% match between Scid and Scidb about the ECO classification
is not possible. But I think that it should be possible to use common names
for the chess variants, for example Scidb is using Zukertort Opening
for the opening 1.Nf3, not Reti, because the Reti system is the move
order 1.Nf3 d5 2.c4. But the name handling in Scidb is not yet the last
word, currently a volunteer is working on Scidb's ECO table, his goal is
that Scidb will use the best known names for each named variation. But
this task is time confusing, I don't think that he will finish before end of
year.
(and possibibly with other chess open source software too)
And ChessBase is going his own (confusing) way.
could you write an email to chess informant asking for clarifications?
Not necessary, the classification of 1.c4 e5 2.Nc3 Nc6 3.g3 g6 4.Bg2 Bg7
5.d3
to A26 is clear - see
http://www.chessinformant.rs/content/eco/eco_a/eco_a2.html.
And the decision about the move transposition handling is private to every
application. In fact the ECO code of Chess Informant is a private invention
of the company Sahovski Informator, but everybody is using this code for any
reason. (In Scidb it is planned to provide information which is more related
to the well known chess variants - like Spanish: Berlin Defence - and the
ECO code is only an unavoidable extra information.)
Hope that this information is useful.
Happy Easter,
Gregor
--
Learn Graph Databases - Download FREE O'Reilly Book Graph Databases is the
