Copyright (c) 2013 John L. Jerz

Positional Chess by Shaun Taulbut

Home
A Proposed Heuristic for a Computer Chess Program (John L. Jerz)
Problem Solving and the Gathering of Diagnostic Information (John L. Jerz)
A Concept of Strategy (John L. Jerz)
Books/Articles I am Reading
Quotes from References of Interest
Satire/ Play
Viva La Vida
Quotes on Thinking
Quotes on Planning
Quotes on Strategy
Quotes Concerning Problem Solving
Computer Chess
Chess Analysis
Early Computers/ New Computers
Problem Solving/ Creativity
Game Theory
Favorite Links
About Me
Additional Notes
The Case for Using Probabilistic Knowledge in a Computer Chess Program (John L. Jerz)
Resilience in Man and Machine

Positional Chess by Shaun Taulbut

Before we write a computer chess program that attempts to play a strong positional game of chess, it would be a good idea to read as much as we can about how humans play positional chess. This particular book is aimed at the typical English or American club player of the 1980's.
 
p.3"The aim of this book is to show the club player aspects of positional play which he can use in his own games."
 
Perhaps we should check and see which of the attacked pawns can be possibly defended by other pawns. This would free our pieces for other action in the game. A piece tied down to defending a pawn cannot be good.
 
p.4"When your opponent attacks a pawn it is important to see if you can defend it with another pawn or only with a piece. If your pieces get tied down to defending pawns your enemy can switch his attack to a different part of the board and you will not be able to react in time."
 
Taulbut gives an example game where black gets compensation for doubled pawns in the form of potential outposts for his pieces. Additionally, certain squares are denied to the enemy pieces.
 
p.8"The doubled pawns have come into being [in the game being discussed] and here they control many important central squares. Black [the side with the doubled pawns] has outposts for his pieces on d4 and f4, whereas White cannot use d5 as an outpost because of the [black] pawn on e6. In addition Black will be able to use the half open f-file to mount an attack on the White king."
 
Another example game is presented where doubled pawns deny critical squares to the opponent's pieces. This means that having doubled pawns (especially ones that cannot be easily attacked) is not necessarily a bad thing.
 
It would seem that an evaluation function which looks at:
 1) the ability of a pawn to deny squares to pieces, 
 2) how attackable the pawns are and
 3) the mobility of the pawns for defensive purposes
 
 would be reflecting Taulbut's ideas for playing positional chess.
 
p.9"Another positive feature of doubled pawns is a possible cramping effect. Doubled pawns may exert pressure by denying natural development squares to the opponent's pieces"
 
Accurate assessment of  piece activity can help us to determine if we have compensation for pawn weaknesses, especially doubled pawns.
 
p.11"So we have seen that there are other factors which must be taken into account along with the doubled pawns. Piece mobility can provide compensation for the doubling."
 
Isolated pawns can be attacked or blockaded. An attacked isolated pawn obviously cannot be defended by another pawn, so it must require that a friendly piece be tied down to its defense. This again cannot be good.
 
p.14-16"The isolated pawn is weak because it has to be defended by pieces if attacked... Another drawback of an isolated pawn is that the square in front of it provides an outpost for an enemy piece. An outpost is a reasonably advanced square from which a piece cannot easily be dislodged."
 
The activity of the pieces can determine whether we have compensation for an isolated queen pawn.
 
p.18-20"Sometimes the activity of the pieces and attacking chances out weigh the positional weakness of the IQP [isolated queen pawn]... Despite their difficulty I think it is possible to play such positions by considering the activity of the pieces. The players without the IQP should concentrate on neutralising the activity of his opponent's pieces"
 
Taulbut believes that backward pawns must be assessed by determining who has control of the square in front of this particular piece. A traditional chess program might not do this.
 
p.24"Normally the most important factor in assessing a backward pawn is control of the square in front of it. If the square in front of it is controlled by the opposition then the backward pawn cannot advance. Its own forces are then required for its defence when it is attacked. There are many openings in which one side accepts a backward pawn for the sake of free play for its pieces. So again the activity of the pieces is very important."
 
Taulbut feels that the side not having the pawn weakness should try to restrict the activity of the opposing pieces. This is difficult to accomplish if we are not accurately measuring piece activity when performing our evaluation function.
 
p.28"In all the examples the activity of the pieces on either side is important in assessing the position. The side not having the defect in pawn structure should try to restrict the opposing army while keeping a close watch on the weak pawns. In general it is easier to play the side without a pawn weakness and the club player who learns this will score many points."
 
Pawn majorities and minorities are the next level of abstraction that Taulbut says we should consider if the position we are playing does not have weak pawns.
 
p.29-30"But what if there are no doubled, isolated or backward pawns? Then we must look at the number and size of the pawn groups on either side. This may seem very abstract, but many a victory can be gained by creating a passed pawn from a pawn majority... Another important point is the position of the king. Black [in the game being discussed] cannot create a majority without fatally exposing his K[ing] to White's army... If both sides have castled on the same side of the board then the player with a pawn majority on the opposite wing often has a decisive positional advantage when the endgame is reached."
 
In some cases, there might be a reason to advance the pawns on the side where we have a minority. The resulting passed pawn created for our opponent might be weak, not particularly far advanced, and perhaps blockadable or attackable.
 
p.30-31"The minority attack is when one side advances a pawn minority against a majority. This does not make sense at first, since it will only result in the creation of a passed pawn... The passed pawn will not be a strong one, but weak, often isolated or backward... So in a minority attack the aim is to smash up the pawn majority into weak configurations, such as isolated pawns and backward pawns. The weak pawns can then be attacked and destroyed."
 
Pawn chains might be attackable at the base or at the head.
 
p.33-34-38"In a pawn chain each pawn is protected by another except for the last pawn, called the base of the chain... If the base of the chain collapses then the other pawns in the chain can be attacked, starting from the back... It becomes important, therefore to find ways of breaking them up. A chain can be broken by an attack against its base or against its head... An attack on the head of a chain often leads... to a backward pawn."
 
Which pieces are better, knights or bishops? Certain rules of thumb might help us to determine whether or not to exchange a bishop for a knight, or vice versa.
 
p.40"The power of the [Knight] declines dramatically near the edge of the board and in the corner... The [Bishop] on the other hand still retains considerable power near the edges of the board."
 
The two bishops are a powerful force, but their effectiveness depends on the placement of pieces, the location of pawns, and the pressure that they can exert on both wings of the board.
 
p.46"The two [Bishops] can also be a weapon in endgames in which both sides have symmetrical pawn structures. Although their influence is much reduced, they can still exert a constricting effect by denying outposts to the opposing knights."
 
Knights are best in the center of the board, and when supported by a pawn.
 
p.47"The knight is at its best in the centre of the board and when it is immune to attack by pawns of pieces.
  So the [knight] needs a Central Support point to operate most effectively."
 
The combination of queen and knight is often better than queen plus bishop. Our computer chess program should know this.
 
p.48"Q+N usually combine together much better than Q+B."
 
A knight on the rim is dim.
 
p.51"Again the [knight] showed its power once it reached a good central square immune from attack."
 
Perhaps we should reward our knight when it is performing blockade duty.
 
p.52"The [knight] has a particularly strong effect when blockading enemy weak pawns"
 
Sometimes we can create rules for exchanging bishops and knights. The rules should take into account the activity of the pieces.
 
p.54"A [Bishop] may be exchanged for an enemy [Knight] if by doing so a significant weakness is created in the opposing pawn structure. Note the word significant... Another case in which the exchange of B for N may be favourable is in blocked positions... Finally, the exchange can sometimes be made in order to harass the enemy pieces, but this is hard to judge, since it depends on the amount of initiative obtained."
 
In order to accurately determine the value of a bishop, we need to see what it can do a few moves into the future - it might bump into a sea of friendly pawns that restrict its mobility, or alternately it might have superior mobility  to distant targets.
 
p.54"We have already met the so-called bad bishop... in which White's pawns were all on the same light-coloured squares as his [Bishop] and it proved fatal."
 
Rooks seem to need open or half-open files to mount effective positional pressure.
 
p.60"Since the rooks need open lines we shall first turn our attention to this topic, to the creation and significance of open files."
 
A rook exerting pressure is one thing, but to be truly threatening on open files the rook must have a penetration square.
 
p.63"Throughout most of this game [currently being discussed] Black managed to control the open b-file, but it was worthless because he had no penetration squares for his rooks... Rooks need penetration squares."
 
Exchanging a rook for a minor piece might be to our advantage, but again it is an accurate assessment of the activity of the pieces that determines whether this will work for us or against us.
 
p.71-74"The value of winning the exchange is very difficult to judge, so it is worth thinking a long time before taking an enemy rook; it can involve a serious loss of initiative... It is often better to retain active minor pieces than to take the exchange [rook for minor piece] and give away the initiative."
 
Queens must be able to trace mobility through squares not controlled by other pieces. Queens often attract threats from other pieces and so must be able to trace mobility to safe squares.
 
p.83"The queen is so valuable, however, that the opposing forces will try to surround  and capture her; the enemy army is drawn to her as though to a magnet. So the queen can have a positional effect by displacing minor pieces in their attempts to capture her."
 
A king can be safe in the center - a traditional chess program might just reward the king if it is in the corner and there are pawns in the vicinity without too many 'defects' in their structure. Perhaps we should be tracing 'realistic mobility' of our opponent's pieces and use that to determine if our king is safe. If the king has the potential to move to safer squares, that also should be taken into account.
 
p.91"The king in the centre is easy meat for the opponent's forces except when the central files are blocked. One should be wary, therefore, of leaving the [king] in the centre unless it is clear that he cannot be got at [attacked]."

Enter supporting content here