The assess-> plan-> execute model put into context
This technical white paper, focuses on the assess-> plan -> execute model which is emphasised in many middlegame books as a method of arriving at an effective plan. It makes a parallel with software engineering, in order to provide a fresh outlook on this approach and questions the use and value of re-usable objects and templates in chess.
Definition of assess-> plan-> execute
Chess middlegame books often emphasise the use of a structured thought process when formulating a plan from a given position.
This involves first assessing the position, and the elements in that position, such as space, the centre, pawn structure, king safety. No variations are considered at this stage. The elements in the position are simply looked at and studied. The "periodic table" of elements approach reflects a scientific approach to the game, and evidence of this scientific approach can be found in the games of players such as Karpov, Botvinnik and Kotov.
It then involves formulating a plan, given the consideration of those elements. The plan will make effective use of these elements to one's advantage. Hopefully the plan will be sufficiently decisive. Alternatively and more realistically the plan will be just be a stepping stone to gaining a decisive advantage. It may just accumulate another advantage which puts the opponent under a little more pressure.
Finally the details of implementing that plan are worked out, i.e. the execution of the plan is defined. This will add consideration of tactical variations which need to be considered, precise move orders, etc. The nuances of the position are hit upon by implication. King safety is a particular factor which often comes to light in variations, because optically a king may be safe, when simply assessing the position. Concrete variations may prove the vulnerability of the King. The triumph of computer chess, is a triumph for execution based thinking. However as humans, we need the first two stages to act as filters to the millions of possibilities inherent in a chess position. The first two stages act to put us on the right path, to be analysing the relevant variations. In chaotic tactical positions however, more emphasis has to be given to analysis over judgement. This stage would therefore rise in significance.
Marshall Vs Capablanca
New (23), 1909
"The dominant feature in this position, is the 3 to 2 pawn majority that I have on the queenside."
"To exploit the pawn majority to create a winning passed pawn which Frank cannot do anything about"
"a5 and b4 will be necessary rolling moves to create the passed pawn. Rd1 check is a nice move to put white under greater pressure because the rook can attack b2. This will make my initial plan more effective. I analyse the variation Rd1+ Kg2 a5, and then I see a few variations:-
A) 28 h4 b4 29 axb4 axb4 30 Rc2 b3 Re2 h5 with Frank still under pressure
B) 28 Kf3 b4 29 axb4 axb4 Rc2 b3 Re2 Bh3! (I bet Frank's going to miss this tactic!) Kf1 Bf1 winning material
C) 28 Bc2 Rd6 Bb1 Rd2 Rc2 Bd5+ Kh3 Rd1 Ba2 Ra1
The game continuation was:-
26...Rd1+ 27.Kg2 a5 28.Rc2 b4 29.axb4 axb4 30.Bf3 Rb1 31.Be2 b3 32.Rd2 Rc1 33.Bd1 c3
34.bxc3 b2 35.Rxb2 Rxd1 36.Rc2 Bf5 37.Rb2 Rc1 38.Rb3 Be4+ 39.Kh3 Rc2 40.f4 h5 41.g4 hxg4+
42.Kxg4 Rxh2 43.Rb4 f5+ 44.Kg3 Re2 45.Rc4 Rxe3+ 46.Kh4 Kg7 47.Rc7+ Kf6 48.Rd7 Bg2 49.Rd6+
So this approach seems extremely logical... and what is wrong with it?! Why can't we just use this assess-> plan-> execute model in all our games?!
A parallel in software engineering
Let us fly off on a tangent.......
"Top-down" design in software engineering is the approach of assessing the problem in order to understand the requirements, forming a high level design which meets those requirements, and then from the high-level design, detailed design, and then from the detailed design, coding.
Top-design however is not only the only approach in today's computing environments. There is much said for the re-use of standard components, and software re-reuse in general. This approach reflects the need to build software quickly and effectively. Object orientation is extremely popular because it offers re-use. Within this context, problems are approached differently. An object hierarchy is modelled based on the requirements. If objects are available off-the-shelf, then these may be bought, rather than written from scratch. The windows programming environment itself offers plenty of re-usable components such as dialogue boxes, combo boxes, etc.
But it is not only low level interface components which can be re-used. Software IT (Information Technology) consultancies within particular sectors, may have standard models which they can apply to meet different customer requirements within that particular market sector. These standard models represent an infrastructure that enables re-use. They do not want to build whole systems from scratch. They want to re-use as much as possible from previous efforts.
Coming back to chess.....
So is there an equivalent in chess of re-use? And what would re-usable objects be? And what would be the equivalent of standard models in particular sectors?
The equivalent of re-use in chess
In chess, re-use would be imply using the crystallisation's of past experience. Looking at a position reached after an opening position from the perspective of assessment, plan, and execution, may be justified fully for someone that has never seen that position before or the position is sufficiently unusual to merit this approach.
However an experienced player in that particular opening, may be very familiar with the themes which arise in the middlegame from that opening. For example, a Sicilian dragon player, may already have the standard plan, of getting a knight to c4, in order to encourage white to give up the white squared bishop, and then later to sacrifice on c3, for very strong pressure on white's position if white has castled queenside.
Effective re-use would also take account of the unique features of the position, and put this crystallisation into context. An implication of this is that the assess-> plan -> execute model is not as widely used in practice as is sometimes given the impression of in middle game books. An alternative methodology is the effective application of templates and re-usable objects.
This is not talked about a great deal in middle game books, yet many chess players use templates and re-usable objects in practice.
The equivalent of re-usable objects in chess might include certain standard "methods" which could get used by a chess player and which form part of the players' strategic armoury. For example the great chess teacher, Nimzowitsch taught several strategic ideas in his classic book "My System". These include "Overprotection", "Restraint", "Rooks on the 7th", "Blockade". These are worthy of study by beginners and advanced player's alike.
These strategic ideas could be used over and over again in different games. They are in effect re-usable objects. Nimzovich one could argue, has had a profound impact on the thinking of many modern grandmasters through his teachings of these objects as they have pervaded the conscious thinking of millions of players.
An example of a re-usable object
The pawn "blockade" is an interesting example of a reusable object. It is taken in principle as being an effective method of reducing the opponent's counterplay. The pawn loses it's mobility by the act of being blockaded.
Nimzowitsch was aware of this property, and wrote about the blockade in great detail. He favoured the knight as an effective blockader. However if the knight was unavailable then other pieces could be used as blockaders. He spoke out the implications of a poor blockader such as the queen which because of its high value, might have to move more frequently and therefore lose its effect as the blockader.
We only have to look at a very recent game from the 1997 FIDE world championship for a blockade example:-
Anand vs Adams Game 2
The game continued:
The Knight is chosen as a more appropriate blockader of white's passed d pawn.
41.Be3 Qc6 42.Qb4 Nd7 43.c5 Qd5 44.Qa4 Qc6 45.Qf4 Kg8 46.Qc4 Kf7 47.Qf4 Kg8 48.Qc4
Openings with some imagination can be considered as analogous to a market sector for an IT software consultancy. An opening represents a certain scope for requirements. The plans of one opening can be fundamentally different to another. For example the Nimzo-indian defence is fundamentally different to the Kings-Indian defence. The set of problems of the Nimzo-indian include in certain variations, how to blockade the opponents doubled pawns. In the Kings Indian, the set of problems might include, how to maximise the effectiveness of the g4 breakthrough.
If a chess player gathers experience in a particular opening, the plans arising after that opening become more and more familiar, and template re-use can be utilised. The danger is obviously in playing stereotyped moves, which do not fit exactly into the unique nuances of the position. However there is a strong argument that a chess player should develop a standard model which can be applied to the set of problems arising out of a particular opening.
In the Kings Indian Defence, Black often decides to launch a pawn storm attack against White's king. Black often gets a pawn to f4, and then tries to maximise the effect of the breakthrough g4, by manoeuvring pieces behind this breakthrough. Manoeuvres include Bg7-f8 Rf7-g7 g7-g5, h5, g4, Ng6-h4 often occur in the games of Kings Indian Defence players. This template occurs deep into the game, however some might argue that this is part of opening theory.
Moves bolded in red indicate template implications
Yuferov,S - Kasparov Garry [E99/11]
1.d4 Nf6 2.c4 g6 3.Nc3 Bg7 4.e4 d6 5.Be2 0-0 6.Nf3 e5 7.0-0 Nc6 8.d5 Ne7 9.Ne1 Nd7 10.Nd3 f5 11.Bd2 Nf6 12.f3 f4 13.c5 g5 14.cxd6 cxd6 15.Nf2 Ng6 16.a4 Rf7 17.Nb5 h5 18.h3 Bf8 19.Qc2 a6 20.Na3 Rg7 21.Rfc1 Nh4 22.Qd1 Bd7 23.Nc4 g4 24.hxg4 hxg4 25.fxg4 Nxg2 26.Kxg2 Nxg4 27.Bxg4 Bxg4 28.Qxg4 Rxg4+ 29.Nxg4 Rc8 30.Nh2 Qh4 31.Rc3 Rc7 32.Rg1 Rg7+ 33.Kh1 Rxg1+ 34.Kxg1 Qh7 35.Bxf4 exf4 36.Nd2 Qd7 37.Rc4 Bg7 38.b3 Bd4+ 39.Kh1 Bc5 40.Ndf3 b5 41.Rc2 Qe8 42.Rg2+ Kf8 43.Ng5 Qh5 44.Ne6+ Ke7 45.Rg7+ Kf6 46.Rg4 bxa4 47.bxa4 Be3 48.Nxf4 Bxf4 49.Rxf4+ Ke7 50.Kg2 Qd1 51.Ng4 Qxa4 52.Ne3 a5 53.Nf5+ Kd7 54.Rh4 Qc2+ 55.Kf3 a4 56.Rh7+ Kd8 57.Ne3 Qb3 58.Kf4 a3 59.Nf5 Qb2 0-1
Kamsky,G - Kasparov Garry [E97/04]
Paris (2), 1992
1.d4 Nf6 2.c4 g6 3.Nc3 Bg7 4.e4 d6 5.Nf3 0-0 6.Be2 e5 7.0-0 Nc6 8.d5 Ne7 9.Nd2 Ne8 10.Rb1 f5 11.b4 Nf6 12.f3 f4 13.c5 g5 14.Nc4 Ng6 15.a4 Rf7 16.Ba3 Bf8 17.a5 Rg7 18.b5 b6 19.cxd6 cxd6 20.Na2 g4 21.fxg4 Nh4 22.Rb3 Nxg4 23.Bf3 bxa5 24.Qd2 a4 25.Rc3 Bd7 26.Rb1 Rc8 27.Kh1 Nxh2 28.Kxh2 Rxc4 29.Rbc1 Nxf3+ 30.Rxf3 Qh4+ 31.Kg1 Rxe4 32.Qf2 Qh5 33.Rc7 Re1+ 34.Qxe1 Qxf3 35.Qd2 Qg4 36.Nc3 f3 37.Rxa7 Qh3 0-1
For another 15 example games, where the "template" moves Rf7-g7 Nh6-g4 and g5-g4 were played have a look using a java browser.
Ineffective template re-use
Playing stereotypical moves without considering the unique aspects of the position is simply asking to lose quickly!
For example, an over-confident Sicilian dragon player may well start thinking after the opening, that he can just double up rooks on the c file, sacrifice on c3, and then destroy his opponents king on the a1-h8 diagonal. This may often be the case, but it is not always!!
An over-confident kings indian player, may think that they can just avalanche the pawns on the kingside, and sacrifice the queenside for time-gain, in order to create an unstoppable mating attack. This again, may often be the case, but there is a great scope for going wrong. If white can exchange light squared bishops for example, black's attack may sometimes be stopped dead in its track, and white may infiltrate via the c7 square into black's position.
Effective template re-use
Effective template re-use is essentially to be able to leverage the hindsight of experience, with the analytical keenness of someone who has not seen the position before in his or her life. With this analytical keenness, the uniqueness of the position is given respect. A great plan may be conceived in the following way:-
0% 60% 73% 100%
Template Unique assessment Unique analysis
The template to be reused may take the player to 60% of the way to conceiving an effective plan for the given position. Then with unique assessment, and analysis of the particular factors of the position, the plan/ game template can be tailored to the position, creating a much more effective plan which meets the unique requirements of the position. And yet, re-use has enabled the rapid formulation of the initial idea to be worked on.
This paper examines the significance of experience in chess, and argues that re-usable objects and templates have great value in a chess player's thinking. These re-use mechanisms enable a chess player to arrive at ideas quickly which can then be tailored to the specific needs of the position. It is argued that ineffective template re-use results in stereotyped moves and plans which will often fail because they do not meet the unique requirements of the position. Effective re-use makes demands of the player to work out the nuances of the position, and to tailor the templates and objects appropriately. The resulting plan will still be appropriate to the position, but the infrastructure of the player's experience has been leveraged.