Thursday, May 9, 2013

Fluid Intelligence, Board Vision, Reading Speed and Tactcs

Intro / Background / Definitions

 What is Fluid Intelligence?

Cattell defined fluid intelligence as "…the ability to perceive relationships independent of previous specific practice or instruction concerning those relationships."
Fluid intelligence involves being able to think and reason abstractly and solve problems. This ability is considered independent of learning, experience, and education. Examples of the use of fluid intelligence include solving puzzles and coming up with problem-solving strategies.

While many people claim that their intelligence seems to decline as they age, research suggests that while fluid intelligence begins to decrease after adolescence, crystallized intelligence continues to increase throughout adulthood.
( http://psychology.about.com/od/cognitivepsychology/a/fluid-crystal.htm )



Working memor


is the system that actively holds multiple pieces of transitory information in the mind, where they can be manipulated

Working memory is generally considered to have limited capacity. The earliest quantification of the capacity limit associated with short-term memory was the "magical number seven" suggested by Miller in 1956.[21] He noticed that the memory span of young adults was around seven elements, called chunks, regardless whether the elements were digits, letters, words, or other units. 

Measures of working-memory capacity are strongly related to performance in other complex cognitive tasks such as reading comprehension, problem solving, and with any measures of the intelligence quotient.[26] Some researchers have argued[27] that working memory capacity reflects the efficiency of executive functions, most notably the ability to maintain a few task-relevant representations in the face of distracting irrelevant information. The tasks seem to reflect individual differences in ability to focus and maintain attention, particularly when other events are serving to capture attention. These effects seem to be a function of frontal brain areas.[28]

( http://en.wikipedia.org/wiki/Working_memory  )


Board Vision



The ability to quickly and accurately recognize where all
the pieces are and assess what they are doing in the present chess position


( http://www.chesscafe.com/text/heisman89.pdf  )




I was watching myself solving this tactical problem.


De Pauw, S.-Paci, Cedric 1997
1k4r1/1p2n3/p7/5r1p/1Qp2q2/P3B3/2P2PPP/R4RK1 b - - 0 30
 Black to move

I did recognise several elements, some of them:
  • Qb4-d6 +
  • Be3xQd4
  • Qf4xh2 is nonsense
  • Rooks at f2 and a1 are doing nothing and cant be won
  • Rg8xg2+
  • Ne7 is uninteresting
  • strong local majority of the black pieces  at the wing of the white King
  • black can hardly use the f- file
I did play around in my mind with these elements to find a useful combination of these elements ( you may call it "analysing" ;). The result: The Problem is about mate.
Now i did start to search the concrete solution ( calculation ) with series of checks.

1... Rxg2+
The white King HAS to take or Qh2#

and now Qf3+ ( usually you move the strongest piece as close as possible to the opponents King ) and then Rg5? OOps don't work there is the Bishop at e3. Hmmm..
Qg4+? ... don't work either. Insight: I need the Rook f5 at the g-file to mate. And so i found the solution

A more structured method to solve that problem would start with material count first and would focus on the king safety next.




Kramnik, Vladimir - Hertneck, Gerald 1994
6k1/5p1p/3p2pb/2pP4/4n3/7P/1PQN1PP1/r1Bn2K1 w - - 0 25
 White to move

Many tactical elements to see
  • weak white back rank
  • Ra1 already x-raying at the white king
  • Black king is save
  • Nd2 is pinned
  • Nd1 and Ne4 are forked by the Qc2
  • material: White has more material
  • black threatens to play Nxd2 and win a piece
  • ...
Playing around with these elements ( you may call it analysis ) you will see its necessary to move the Nd2 away but to protect the Bc1.



So what is necessary to solve chess problems?

  • recognising the (possible) tactical elements
  • correct judgement of these elements
  • combining these elements to an Idea
  • "logical" reasoning with the elements to make the Idea work
The capacity of the working memory is of big importance here. The working memory is equivalent to the word-size of a CPU. Calculations with an 8 bit CPU are usually much slower than calculations with a 64bit CPU. If the number of important tactical elements exceed the capacity of the working memory of the tactician, then the tactician has a serious problem. This happens if the tactician don't "understand" the problem. A possible reason is, that some the important elements are not recognised or simply unknown ( not "chunked" ).

The number of ( possible/interesting/relevant ) tactical elements is high and will usually exceed ( by far ) the capacity of the working memory. So we have to load and unload these elements permanently. "Board Vision" is the Cache memory of the Tactician, the Tactician does not have to look at the Board and check, he "knows" where the pieces are and what they are doing. Things we don't have in our internal board have to be "read" from the board. That the slowest process, equivalent to the reading from the hard drive.

Wednesday, May 8, 2013

Leaving Chess.com because of Nazis

I found from time to time Nazis at Chess.com, for example : Adolf Hitler 1488.
1488 is a special sign to show the fascistic believes. You may find detailed info's about that number in the Internet. Usually these guys have been removed in no time, after a report of mine. Now a Nazi with the nick Hans_SS appeared and made the statements at the debate-chat-room that he is a Nazi, killing Jews by the SS was OK because Jews have been the enemy in that war and so on. I did report him, Hans_SS was not removed. I was told "that's an opinion".
So i left Chess.com, i cant accept that.

I did ask at http://www.chesscube.com about things of that nature and a moderator told me, that there is 0 tolerance with Nazis.

So i found a new home

Tuesday, April 16, 2013

Chess Thought Process and Speed Reading and Board Vision

Introduction

Surfing throug the internet i found a post where someone claimed an improvement of 150 elopoints in chess by doing the exercises of the speedreading program Eyeq ( cant find the link now though ). I cant believe that this would work with everyone, maybe it dont work at all. But it made me think:

 ( A generic )  Chess Thought Process


  1. Reading ( Orientation, Understanding , whats going on, who is to move, where are the pieces, what are the pieces doing.. )
  2. Searching ( what can i do, CCT .. )
  3. An idea ( weakness ) is found ( Pattern recognition )
  4. Reasoning  ( how to make the Idea work )
  5. Calculation ( is it ok )
  6. If ok then move else goto 2

Step one : Reading is "continous", you will improve the understanding of a position ( to read it ) more and more, even if you are already at step 2,4 or 5.
What i call "Reading" is called "Examination" by Beim ( The Enigma of Chess Intuition ) or "Analysis" by Heisman ( http://www.chesscafe.com/text/heisman89.pdf ). I like the Term "Read" because there are several similaritys between understanding a chess position and understanding a written text and in the english language "reading" can have the meaning of "understanding" too.


Reading with a highlighter pen

Here a chess tactic puzzle which is all about "reading" in my eyes. "No" searching, "no" calculation is needed:


 Wallner - Todor 1996
 3r1nk1/1q3nb1/r1pp2p1/3Np1P1/3BPP2/pP1Q1B1R/P1R5/2K5 w - - 

White to move

As soon as you know what every piece is doing you will know the solution. Maybe i should use a highlighter pen to highlight the most important sentences?





All other pieces are "uninteresting":



Last Hint: Pc6,Pf7,Pg7,Nf8 are blocking

Chess- and Text- reading

Munich did use the example of "Speedreading" very often in our discussions about chess improvement. Well i thought: reading is a "skill", others did compare chess with bike-riding. But the similaritys are bigger than i thought. Chess and written Text are 2 dimensional placed symbols ( letters at a paper or a pieces at a board ) which have to be "understood correctly".

An experiment at an elementary school in Germany

The pupils of an elemantary school in germany did get only 3 instead of 4 hours of Mathlesson a week. Therefor they had 1 hour chesslesson. The result of that lesson: The pupils did get better in Math than the averagr pupil with traditional lessons and : the pupils did get better in reading and writing

Now why Speedreading, why is speed important?

The main difference between a strong and a less strong player is the time they need to find a move. Empirical Rabbit did show in his analysis ( here and at many other posts ) that a less good player is "just" slower. A player who is thinking twice as long will find moves which are 100-200 elopoints better. Some tacticians at Chesstempo are thinking 1 hour and more about one puzzle and they solve puzzles which are rated several hundred elopoints higher than their usuall rating. 




Chunking

Good chessplayer dont think quicker, they know by experience what to look for. But they can "read" a position quicker too. Several posts of this blog are about the superior board vision of Chess Masters. The cognitive research did show that our fluid intelligence is limited by the size of our working memory. Experts are bypassing the problems with our limited working memory by using chunks. That is like the difference between a 8 Bit CPU and say a 64 Bit CPU.

And there is the next similarity between reading and chess: Chunking. A beginner in reading is reading Letter by Letter. improved reader are reading complete words ( =  chunks ). Chess masters dont see a bishop and a lot of squares, they "see" the bishop with the whole controlled diagonals.



Further similaritys: Regression 

Regression is chess: if you repeat to calculate the same over and over again. Kotov's famous Book  "Thinking like a Grandmaster" is about it.
Regression in reading: https://www.youtube.com/watch?v=Qd6meu-nlQE



Further similaritys: Fixation

Watch the eyes of your opponent and see how many fixations are common in chess.
Speedreading: https://www.youtube.com/watch?v=w5IFvFJbNKs




A view other exercises and further information

https://www.youtube.com/watch?v=Sf02YOzTyI0
https://www.youtube.com/watch?v=0a1gCbAZfc0
https://www.youtube.com/watch?v=ZH4_36IxS7s
http://www.youtube.com/watch?v=2hQhBpTdhqs&feature=share&list=PLofaHHP7uYJKEOgRi53FRt1vkAnNuHktw
https://www.youtube.com/watch?v=MrAXtTSVstc

https://www.youtube.com/watch?v=egK86sPCi2M
https://www.youtube.com/watch?v=3g_yzoY8yjc
http://www.ireadfaster.com/peripheralApplet.html


Now i wonder if some of the exercises for speedreading, may influence the performance in chess too?

Thursday, April 4, 2013

Free spaced repetition chess training programs

Problem

You improve ( in chess ) if you learn more ( relevant ) things than you forgett.

            Improvement = Quantity * Quality learned - Quantity * Quality forgotten

Everything you dont repeat will be forgotten soner or later

            Use it or lose it.

We can try to learn more and more new things  but after a few days all this new information is lost. It is like running water in a watertank with holes all over.

Solution : Repetition

 New things will only stay for a short while in our Memory , but with every recall of this information, it will stay longer and longer in our memory. Every recall ( at the right moment ) will double the time till its forgotten. The best method to learn is called spaced repetition . That is the reason why i use Anki as a tool for my chess learning.





YATT - Yet Another Tactics Trainer

Using Spaced repetition







 Lucas Chess

Using Spaced repetition

 

 

Look here for other good and free chess training software



Sunday, March 31, 2013

Understanding Tactics Tit for Tat ( = UnCounterAttack )

 Under Construction:

One of our pieces is attacked. Now we have 5 possible ways of defending ( Yuri Averbakh : Chess Middle games, Essential Knowledge )

  1. Withdraw
  2. Covering
  3. Capture the attacking piece
  4. Support
  5. counter attack
Often we prefer a (strong) counter attack, to keep, or to get, the initiative.
If this counter attack is not sound, then we have a tactical puzzle. I did call this situation : Tit for Tat ( =UnCounterAttack ).

Terms:
             AO   ::= Attacked Object, 
             cAO ::=CounterAttacked Object
             TFT  ::= Tit for Tat - Situation


I)  cAO is moving At once

a)  cAO is getting lost

Hajnal -  Vanczac 2002
5rk1/2p2r1p/3b4/1Q2P1p1/1PbR2q1/4B1P1/5PNP/5RK1 b - - 0 33
Black to move
xQb6 - xQg4
With QxRd4 we have a new TFT
Method: desperado of cAO / CounterCounterAttack
! : The rook at f1

                                        

Los - Bertholee 1995
4r1k1/1b3ppp/1p1Nr2q/p1p1n3/8/PP4P1/2P2PBP/R3RQK1 b - - 0 22
 Black to move
xNd6 - xBb7
With Bxg2 we have a new TFT 
Method : Desperado / CounterCounterAttack


b)  cAO is not getting lost

5rk1/p4p2/2pr4/3p2qp/Q3R3/2P5/PP3PP1/3R2K1 b - - 
Black to move 
xRe4 - xRd6

With 1... Rg6 black has a double blow at g2 and e4
Method: remove cAO with Mate Threat

                                       

Konopka - Tibensky 1997
4q3/r5k1/1pp1rpNp/p2p1Qp1/Pb1P2P1/3R1R2/1P3PKP/8 b - - 0 40

Black to move
QxN - QxR

But after 1...Rd6 the Knight has no escape
Method: remove cAO

                                       

5r1k/1p4pp/3Q4/2p3q1/3P4/7b/1P3PP1/2R1N1K1 b - - 

Black to move
QxR - QxR

Rd8 dont work because of Qxc5, but the mate-threat Re8 is enough
Method: Remove cAO with Mate Thread

                                       

2rr3k/pb4pB/1p2p3/2p1q3/Q7/4P3/P4PPP/2RR2K1 b - - 

Black to Move
Kxh7 - Qh4+ Forking d8

But after 1...Rxd1+ Qh4 dont Fork the Rook anymore
Method: Remove cAO with Check

                                        

Pavasovic - Rotstein
Austria 2006
8/q4k2/p3p1p1/3RPn1p/4r3/PB3Q2/1PP5/2K5 b - - 0 41

Black to move
xRd5 - Bxd5+ and BxRe4

But Re3..


II)  cAO is not moving At once

  a)  cAO is saved

Schebler -  Rotstein 2001
r4Bk1/p3Qp1p/1pqp2p1/8/2P5/1P4P1/P3P2P/R2n2K1 b - -
Black to move
xBf8 - xNd1
After Qc5+ the Knight can be rescued by tempo with Nf2+ or Ne3+


Not Sorted Examples:

r4rk1/1p4b1/3q2B1/p5Bn/P2PR1nN/8/1P1Q4/R5K1 b - - 1 1
Black to move
xBg6 - xNg4

With 1...Rf2 black has a double ( triple ) blow Dh2#, RxQ, QxBg6
Method: Moving a third piece to create a strong threat
                                       

6rk/6pp/1Qb3r1/P4p2/2p3q1/7P/5PP1/R3RBK1 b - - 0 1


Black to move
Qxh2 - Qxc6

With 1...Qf3 the Queen protects the Bc6 and the Mate-threat is enough

                                       

r5k1/b2brp1p/p1pN1pn1/q1N5/1p2P3/1Q4P1/PP2B1KP/2RR4 b - - 1 1

Black to move
BxNc5 - Nb7 Fork and Nxc5

But after 1... Bxc5 2.Nb7 Bh3+ und 3...Rxb7

                                       



2r1rbk1/1R1Q1p1p/3p2p1/3P4/8/4N2P/P4qP1/3R3K b - - 1 1

Black to move
QxN - Qxf7 and mate
RxN - QxR

But 1... Red8 forces the Queen to retreat

                                       


Jirovsky-Hennigan 1992
2k4r/pp6/2p1p1b1/P2pn1P1/1P1P2Nq/2P3N1/7Q/R5K1 w - - 0 35

White to move
QxQ - Nf3+ and NxQ

But after dxe5 the Knight is gone
( I love this type of puzzle: If i take it, then i have it :)

                                       

8/p4pk1/1p2p1p1/7R/3P1Q2/P5P1/5PKP/1qr5 b - - 0 1
Black to move
xRh5 - Qg5+ and draw by repetition or xRc1

After Rc1-g1+ Black can take the Rh5 and protect the king by Qb1-g6
                                        


Edvardsson-Sigurpalsson
Reykjavik 1995
8/5pk1/1p1q2p1/5n1p/PpQ2r2/5PPP/8/3R1R1K w - - 0 40
White to move

Qxf4 - Nxg3+ and Qd6xQ
RxQd6 - RxQc4

But g3xRf4

                                       

Tsorbatzoglou - Petraki 1994
3Br1k1/bpq2ppp/2pp4/p3nP2/2P5/1P4PP/P1Q2nB1/3R1R1K w - - 0 22
White to move
Rxf2 - xBd8

But after Qxf2 we have a new TFT: QxB - QxB


Tuesday, March 19, 2013

Chess Atoms, Contacts, Board Vision

Intro:

Since a view weeks i am thinking about Board Vision. We need to see the pieces and what they are doing. This seems to be easy for someone with healthy eyes but .. we don't see with our eyes, we see with our memory. With Board Vision Exercises we create chunks = memories of pattern which enable us to see better whats going on. 

Why is Board Vision interesting?
It is possible to improve in Board Vision and it has a positive ( but maybe [very] small ) effect. No other form of training in chess should have that much "transfer".

Purpose of this Post:

Writing this post should help me to detect useful new exercises. I try to find "all" different (sub)types of Board Vision. Board Vision is the Atom of understanding in chess. The Atoms of tactics are the "contacts".

Method of this Post:

"Bottom up", it will start from the very beginning :)

And here we go:



I) Empty Board:

Has: Black and white squares, diagonals, files, ranks, center, borders.

Existing Training-Tools: Chess Eye ( and others ) but they work with Notation. This is useful for blindfold chess, but i think the "translation" board -> notation and back is not "the right thing".
I think ranks and diagonals are only interesting in combination with Rooks and Bishops. I don't see the need for a special trainingstool here.


II) A piece:

Has: Value ( Knight ~ 3.5 ). The Material Value of a piece is a measurement of the possible activities of the piece in the future.

Existing Training-Tools: My Material Trainer ( beta )

III) A single piece at the Board:

A piece at the board

Has: Mobility ( = the number of legal moves. Most pieces have more mobility at the center of the board. ), Piece->Location, Location->Piece.

Existing Tools: Mobility : My Mobility Trainer and ( somehow and indirect ) Chess Training Wheels of Josh Weihnacht. Location <-> Piece : My Memory Trainer, "Chess Memory", ChessGym Memory, My Trainingtools for Visiualisation, Chessity "Route Planner".

Special case: Pawn - Promotion_Square Contact
( Averbakh Contact #6 )
pawn close to promotion

Has: increasing value of the pawn (potential future activity as Queen? ) as better the chances to promote.

Existing Tools: (Pawn-) Endgame Training ( f.e. Chessimo ).
It would be possible to extend my Material Trainer for that ( increasing value of pawn ... )


IV)  2 pieces at Board

This are ( some of ) the "contacts" of Averbakh/Nimzowich

a) Attacking Contact:
( Averbakh contact #5 )
Bishop attacking pawn

Has: "isAttacked" ( f6 ) , "isAttacking" ( Bc3 )

Existing Tools: Chess Gym Attacked Training (isAttacked), Fritz Attack Training  (isAttacked) and ( somehow and indirect ) Fritz Check Training (~ may attack the king in the next move ). 
Missing: A Tool about "isAttacking" (  i will ( have to ) write it by myself  ) and a tool to find all pieces which are not overprotected. ( they are the targets for double blows.. )

b) Defending ( supporting ) Contact
( Averbakh contact #1 )

Bishop supporting pawn

Has: "isProtected" (f6), "isProtecting" (Bc3)
Existing Tools: ChessGym ("isDefended") and Fritz ( "isNotDefended")
Missing: isDefending (or isNotDefending? ), i am not sure if this is interesting, might be more interesting to find pieces which are "doing" nothing = not defending AND not attacking?.

c) Restricting , a piece is reducing the mobility of the other piece
( Averbakh contact #3 )

i) same color ( = often not "coordinated" pieces )

restricting each others mobility

Has: isRestricting ( Re5 ), isRestricted ( Qd5 [from going to f5,g5..] )
Existing Tools: My Mobility Trainer. Every piece which is defending is restricted in its Mobility by the defended piece so the defense training is a tool too? Maybe a training with the question: "which piece is blocking the mobility of the other pieces most ?" is interesting, but then whats about Batteries, connected rooks...?

ii) not i)

pieces restricting each others mobility
( to a different degree )

Has: isRestricting, isRestricted 
Existing Tools: None
A special tool has to be made, which concerns the "save mobility". The save_mobility of the knight is here 0 and the save_mobility of the bishop is 9. But a "proper" save_mobility is not easy to calculate ( see the "SEE-algorithm"). Maybe a "problem filter" ( if in doubt don't use example ) may help???

d) Threat of an attack
( Averbakh contact #4 )

may attack in the next move

Has: IsThreatendToBeAttacked ( Bg6 by Ne2 ( via f4) ; Ne2 by Bg6 ( via d3 or h5) ), isTreateningToAttack
ExistingTools: Fritz Check Training (only King is threatend ), Chess Fork Trainer.
Missing: Find all pieces which can be threatend to be attacked by  piece X. Find all pieces which can threat to attack piece Y
( to be written )

Edit: I fond an easy tactic puzzle illustrating a version of this type of bard vision:

 Luskutov - Polluljahov 1991
 Black to move

Black would like to take the Knight at c3 but what is to do after whites Rh3?
Hint: It's not "Thread of Attack" its "Threat of Defence"


V)  3 pieces at Board

a) Covering
( Averbakh contact #2 )
bishop covering king

Has: is(Xray)Attacking, is(Direct)Attacked,is(Xray)Attacked
Existing Tools: Tactics Training with Skewer, Pin, XrayAttack.
Missing: Tools to find all 3 roles/pieces in a given position ( to be written )

 b) Fork

King-fork

Has: isForked (isAttacked),isForking(isAttacking)
Existing Tools: Chess Fork Trainer, Tactics Training with Forks
Missing: Find all possible Forks with X isForked, Find all possible Forks with Y isForking ( ? interesting ? )



Monday, March 18, 2013

My Board Vision Training : Update

I use Megabase 10 as Database for my Fritz - Attack - Training. With my own Games as Database i would be ( estimatingly ) quicker:


Solovtsov, Alexander Vladimiro - Chigorin, Mikhail 
Moscow 1892

Find all white and black pieces which can be taken

Usually i find 40 pieces per minute ( still improving ), at this puzzle it did take me a little longer to find all pieces. I did not "see"  (=  instance awareness ) Bb8 x g3. 

Why? I "see" a move like Bc1 x h6 in "notime". This is the same type of move?

A bishop at b8 is very rare, i did not "learn" this move, i have to "calculate" it: i have to look at the piece Bb8 then i have to look along the diagonal till g3. If i would use my own games as Database for the Attack-Training then i would see "only" positions which are ( more or less ) common for me. As less common a position is for me , as longer it takes to find all pieces which can be taken.

I have the "impression", that my calculations are getting less exhausting thanks to this training. I can bypass these "move possible" check, i just know ( = recall ) it. Tactic puzzles calculation is  usually a bunch of take this, take that..

These Board Vision Exercises "connect" the pieces in the brain to a ( =1!) simple function. My hope is, that this might be a (the) foundation for more complex chunks. One single Vision will not improve me a lot, but several Visions in combination with other training might have a bigger effect on my performance.
The Attack Training  did not tell me anything new but the Mobility Trainig is seemingly opening a new form of "sight".