Wikipedia and some other websites are showing the so called 'Game of the century', played by Donald Byrne vs Robert James Fischer in 1956 as 'Gruenfeld Defence' with ECO as D92 whereas after importing the game in Lichess analysis board the opening is displayed as English opening: Anglo Indian Defense, King's Indian Formation with ECO A15.
Is it a bug? or is it that the Lichess database cannot detect the move combinations properly?
The game is given below:
Something to work Lichess!
What's in a name? It starts as English Opening A15 and later transposes into Grünfeld Defence D92. So both are right in some way.
In Wikipedia a human noticed the transposition and simply wrote it was a Grunfeld, the thing is that a Grunfeld is commonly played against 1.d4, the game started with 1.c4 which made it an English opening for the computer.
An opening can transpose quite easily, for instance c4 nf6 d4 g6 nc3 d5 is a grunfeld, just with a different move order, you would not call this an English, it would be called a Grunfeld. Take for example the moves d4 c6 e4 d5, this may have started out as a slav invitation (i.e. if white followed up with c4,) but because white plays e4 it is now a Caro-Kann, not a slav
As far as the computer is concerned, ECO D92 is defined by the position arising from 1. d4 Nf6 2. c4 g6 3. Nc3 d5 4. Nf3 Bg7 5. Bf4.
Black delayed playing d5 for long enough that the game never reached this position - 5...d5 only transposes to the position following 5...O-O. Had white played 6. e3, the game would have been a D93; but he didn't, so the game avoided every Grünfeld stem position defined in the ECO.
To a human, the game became a King's Indian (with the option for Grünfeld) with 4. d4 and a Grünfeld, Hungarian Attack with 5...d5. A computer could perform a search and find that the position after 5...d5 most commonly comes from D92, and therefore classify the game as D92 as well, but this is overly complicated and likely error-prone anyway.