"Pattern recognition" DEBUNKED
Sort:
Pattern recognition is only part of the story but it is a significant part. It is less about recognizing positions, my Homo semi-erectus friend, than it is about recognizing particular motifs that might appear in any number of positions.
Dec 20, 2016
0
#3
When you memorize and fully comprehend a single 35 move chess game, you have absorbed 200+ patterns. This process of learning a single game can be accomplished in two hours if the opening is wholly new. A good teacher can accelerate the process.
Dec 20, 2016
0
#4
urk wrote:
This time-honored but hokey explanation for strong chess play just doesn't hold water when you really look at it.
"On average, with 10 positions learned a day, it takes 27 years to acquire 100,000 patterns which, in turn, makes it hard to explain how young super-grandmasters, like Magnus Carlsen (grandmaster at 13, learned chess at age 8, which suggests 20,000 patterns a year and 55 patterns a day) can at such a tender age can be much stronger than older grandmasters who have had much more time to acquire far more patterns...."
https://www.chess.com/article/view/pattern-recognition-fact-or-fiction
"On average, with 10 positions learned a day, it takes 27 years to acquire 100,000 patterns which, in turn, makes it hard to explain how young super-grandmasters, like Magnus Carlsen (grandmaster at 13, learned chess at age 8, which suggests 20,000 patterns a year and 55 patterns a day) can at such a tender age can be much stronger than older grandmasters who have had much more time to acquire far more patterns...."
https://www.chess.com/article/view/pattern-recognition-fact-or-fiction
There are two flaws in this logic.
(1) What makes you think that 100,000 is the magic number of patterns learned that enables one to be a grandmaster?
(2) What makes you think that chess skill is only about pattern recognition? Surely, it partly is--but there are other factors.
Don't agree with much of the debunking. Carlsen can identify far more positions from specific games and ideas than other GM's - despite having had far fewer years to assimilate these: Carlsen Wins Chess Trivia in Iceland
Dec 20, 2016
0
#6
You know that a pin is a pattern right. Without patterns 1 minute chess would near impossible to play for GM's.. your referencing how many moves there are on the board is ridiculous as usual.
Dec 20, 2016
0
#7
Grumblesmurf, you are aware that different people have different mental abilities, aren't you?
Dec 20, 2016
0
#8
I don't think it's true that people play well because they're recognizing patterns they've seen before. They're playing well because they understand what to do with the patterns. Which means what? It just means they're better at playing the game, which leaves us back where we started.
Yes, familiarity with your structure and the tactics that arise helps a lot, but GMs are always getting into uncomfortable positions and still playing like GMs. These youngsters who are playing so strong haven't memorized chess history - a lot of them probably couldn't even tell you who Paul Keres was or seen any of his games.
Yes, familiarity with your structure and the tactics that arise helps a lot, but GMs are always getting into uncomfortable positions and still playing like GMs. These youngsters who are playing so strong haven't memorized chess history - a lot of them probably couldn't even tell you who Paul Keres was or seen any of his games.
In the movie Rainman, in the scene where the toothpicks were scattered on the diner floor, the savant , Raymond (Dustin Hoffman) blurted out 82, 82, 82, 246. The waitress confirmed that there were only 4 toothpicks left in a box that originally contained 250. The savant's brain is wired to group the toothpicks in patterns of 3 groups of 82 each. Our average brains just see a jumble of toothpicks scattered on the floor.
Dec 20, 2016
0
#11
urk wrote:
I don't think it's true that people play well because they're recognizing patterns they've seen before. They're playing well because they understand what to do with the patterns.
As pullin already noted, patterns are dynamic relationships between the pieces, as well as static positions. As I pointed out in http://chessskill.blogspot.com/2016/12/pattern-training.html, these two positions feature the same pattern:
White to move
The position comes from Horvath -- Vigus, Haarlem 1998.
The second position is from a game I played last week in an online blitz game.
White to move
Dec 20, 2016
0
#12
Ziryab, so what?
Did you need to have seen Horvath-Vigus before you pulled that off? Of course you didn't. You just needed to have a good understanding of the dynamics of the game.
Where's the pattern recognition?
Did you need to have seen Horvath-Vigus before you pulled that off? Of course you didn't. You just needed to have a good understanding of the dynamics of the game.
Where's the pattern recognition?
Dec 20, 2016
0
#13
And pullin, nobody referenced how many moves there are on the board. It was simply doing some math on learning patterns. The whole idea behind "pattern recognition" is that you need to have seen (and memorized) thousands and thousands of patterns to become a good player.
Which is simply not true.
Which is simply not true.
Dec 20, 2016
0
#14
I'd say you are throwing the baby out with the bathwater here. You need to see patterns to understand concepts and you need concepts to see patterns. And most importantly, you need to understand what concepts are higher priority than others in any position, which is how GM's can play like GM's when outside their preparations and familiar positions- they can identify what concepts are a priority in the position then look for the patterns of those concepts to help calculate the details. To say you don't need to know patterns to be a strong player is just inaccurate, without patterns in our head we'd minimally be forced to calculate like engines, and we'd be far poorer as a result (not to mention slow and mentally exhausting).
Its not as simple as memorizing patterns, no, but it's a definite part of what helps a player play better.
Dec 20, 2016
0
#15
Chess is all about judgment - setting priorities and timing - UNDERSTANDING HOW THE PIECES INTERACT.
I don't agree that we're "forced to calculate like engines" without relying on familiar patterns everywhere. Good players understand the underlying dynamics and they're relying on that understanding to make good moves.
As humans we see patterns everywhere but to simplistically say GMs win because they know their patterns is malarkey.
I don't agree that we're "forced to calculate like engines" without relying on familiar patterns everywhere. Good players understand the underlying dynamics and they're relying on that understanding to make good moves.
As humans we see patterns everywhere but to simplistically say GMs win because they know their patterns is malarkey.
Dec 20, 2016
0
#16
Nobody is saying they win because they know their patterns- chess is obviously more than pattern recognition, and yes without patterns we literally have to calculate "I go here, they go here, I take that- I'm up a pawn." Which again without pattern familiarity we'd be susceptible to horizon effect where 6 moves down the road a pawn begins to race down the board that without pattern recognition you'd have no way of understanding the severity of it until you saw it in action as an unstoppable passed pawn- and that'd be a pattern you'd remember along with the concept of unstoppable passed pawn!
You can't have understanding of concepts and ability to see dynamics in positions when you have no pattern from which to base it on. Here's an example: one person has a book that explains the various mechanisms of trains, one person has seen many trains, and one person has seen many trains and has the book explaining the mechanisms of trains. Which one is most likely to successfully build a train?
fieldsofforce wrote:
In the movie Rainman, in the scene where the toothpicks were scattered on the diner floor, the savant , Raymond (Dustin Hoffman) blurted out 82, 82, 82, 246. The waitress confirmed that there were only 4 toothpicks left in a box that originally contained 250. The savant's brain is wired to group the toothpicks in patterns of 3 groups of 82 each. Our average brains just see a jumble of toothpicks scattered on the floor.
Dec 20, 2016
0
#18
urk wrote:
You just needed to have a good understanding of the dynamics of the game.
Where's the pattern recognition?
Where's the pattern recognition?
In your previous sentence. In "the dynamics of the game". The "good understanding" you note stems from pattern recognition, in part. Having seen many hundreds of back rank checkmates and having presented Horvath -- Vigus to many hundreds of students, I needed neither calculation not time to see the checkmate in three. Recognition was instantaneous.
Dec 20, 2016
0
#19
Candidate I see your analogy but we're talking about chess here, not train construction.
There are only 64 squares on the board and 32 pieces. That's it, and it's a complete information game. In weird positions a strong player is seeking to understand the dynamics and setting priorities. He's not desperately looking for a familiar pattern.
There are only 64 squares on the board and 32 pieces. That's it, and it's a complete information game. In weird positions a strong player is seeking to understand the dynamics and setting priorities. He's not desperately looking for a familiar pattern.
I've heard people say that chess is a game of memorizing the openings (particularly a 960 opponent, trying to claim that 960 is a purer form of chess). I agree with some combination of Urk and Candidate35. Pattern recognition is a tool in a game of judgment. The problem, semantically is, what do we consider pattern recognition? If it's rote memorization, then the quote which generated the thread is likely correct- there is no way a 13 year old could have memorized the number of positions and their correct move to know how to be a grandmaster. If we take pattern recognition to be knowledge of the several repeating elements of games which one should look for, then pattern recognition was attainable for the 13 year old. Is there a hard and soft theory of pattern recognition, like with AI?
"On average, with 10 positions learned a day, it takes 27 years to acquire 100,000 patterns which, in turn, makes it hard to explain how young super-grandmasters, like Magnus Carlsen (grandmaster at 13, learned chess at age 8, which suggests 20,000 patterns a year and 55 patterns a day) can at such a tender age can be much stronger than older grandmasters who have had much more time to acquire far more patterns...."
https://www.chess.com/article/view/pattern-recognition-fact-or-fiction