A good description of reading in latest Lex Fridman podcast with Magnus Carlsen, about 20 mins in…
2 Likes
Are you referring to this question and answer?
I feel like his question touched on what we have been discussing, but actually the answer didn’t really illuminate the process of reading that Carlson goes through…
I always want to know whether we need to visulization whole board in our brain.
In myself condition, I just to visulize imaginary stone on board, it is difficult and sometimes I just forgot which color at a position, reading on a board is easier, because I can’t handle a board in my head.
1 Like
One thing I do know about advanced chess player visualisation is that they don’t visualise the board by the placement of individual pieces, and tracking each, they see “whole positions”.
So for them a position is like “Ah, this game is at the Fred’s Gambit, except the kings rook pawn is advanced, so I will apply the Swardovsky Maneuver” … etc.
I can imagine that Go players also chunk positions to a certain extent, but I’m not clear how well the technique really transfers. One reason for believing that it doesn’t transfer well is that outside josekis and fusekis (where it does apply exactly: I can “read” some basic joseki, because I have seen them so often) we don’t read about positions with names.
For this reason, I suspect that advanced Go reading really is more like “reading” than “chunking”.
(And the Dwyrin example I posted shows this: he just “sees” the sequence, even though it is a unique one).
1 Like
I strongly disagree. Go is full of patterns and shapes and classic sequences. They appear all over the board and at all move numbers. Just because they don’t all have names doesn’t mean we don’t remember them.
Imagine I play a hane in the first line in the endgame. You respond on the first line with an atari on my stone, then I connect on the first line. Then either you need to connect on the second line, or you don’t, depending on the context. Well, you can read it out if the context is really unusual, but most of the time you’re just going to immediately recognise whether a connection is required or not without reading any sequence, just because this is a common pattern and you can intuitively recognise in which contexts you need to protect the cut and in which you don’t.
The concept of “vital point” also stems from this. And actually, many shapes and patterns do have names. Take a bamboo joint. If I have three of the four stones that would make a bamboo joint, then I’m naturally going to consider a play at the fourth intersection when I read sequences, either by me (to complete the bamboo joint) or by my opponent (to “break the bamboo”).
1 Like
Yes, there are local shapes we recognise, and this is certainly a form of chunking.
But you disagreed with a my statement as if you read “position” as “shape”.
The positions I was referring to a whole board things, in chess. We don’t have whole-board positions that we recognise in go, outside fuseki and joseki. Or at least, that is my experience/observation.
So it is true that we chunk shapes and sequences from them locally, but this is part of a reading exercise that requires the whole board to be held, for the most part, as distinct from chess position chunking, where an advanced chess player can literally look at a board for a moment, then correctly set up all the pieces on another board.
It is of course a trap to rule a black and white line between the games, when there is of course a grey zone of similarity. But if we return to the context of my comment: which was Carlsen’s response to the “what does it feel like to read” question … I think his answer is different to that from a go master, for the reasons I offered.
Probably feels like
(And many more Spider-Man gifs)
5 Likes