Translate this page
We are still ironing out some wrinkles in the website redesign. In the meantime enjoy this month's Novice Nook. Please support this column with a purchase from our chess shop.
Piece Evaluation vs. Position Evaluation
Quote of the Month: Evaluation tells you how valuable something is.
I have devoted many Novice Nooks to the suggestion that the main chess skill is analysis. I have also stated the companion thought process skill to analysis, and thus "second in command," is evaluation. My definition for evaluation of a position is determining which side stands better, how much better, and why?
It is also important to differentiate between static evaluation and dynamic evaluation:
Static evaluation occurs by considering the static elements of the position, e.g., what material is on the board, positional factors like weak pawns, open files, etc, and "visual" king safety – the apparent ability of the other side to attack a king. Static evaluation does not involve any analysis; that is, you do not move the pieces in your head to see what might happen. You just look at the board as it is now.
Dynamic evaluation uses the same criteria as static evaluation, but it does not occur on the current position on the board. Instead, dynamic evaluation occurs at quiescent nodes in analysis. In other words, you need to analyze the position and evaluate when you arrive at positions of quiescence (i.e., when the smoke clears; no more forcing moves of consequence) or, rarely, when you arrive at positions that require speculative sacrificial judgment. Dynamic evaluation takes place after you see whether analysis reveals that one side might lose material, get mated, have the pawn structure ruined, etc.
The difference between static and dynamic evaluation is starkly illustrated in the following extreme example:
Either side to play – Who stand better?
Statically, the position is perfectly equal. Both sides have the same piece placement.
Dynamically, however, the situation is quite different. Any reader would agree that whoever has the move is winning with QxQ#. This also illustrates, rather trivially, that ultimately you should trust your dynamic evaluation(s) more than your static ones.
In Evaluation Criteria I listed the four major elements of evaluating a position (static or dynamic):
In my first book, Elements of Positional Evaluation: How the Pieces Get Their Power (written on a typewriter forty years ago!), I listed the seven elements that constitute the evaluation of a piece:
This piece evaluation concept can be applied to a piece's instantaneous situation in a given position to help compare to its average value. For example, if in a particular position one has a "good bishop," then its elements, taken together, should indicate that the bishop's aggregate elements – and thus current value – should be greater than those of an average bishop.
Let's address the differences, similarities, and overlaps between a few of these evaluation concepts. In that sense the column is somewhat a bridge between Evaluation Criteria and Elements of Positional Evaluation.
Before we start, an ironic aside about Elements of Positional Evaluation: even though the title contains the words "Positional" and "Evaluation," it is not about evaluating positions. The latter was to be addressed much later in Evaluation Criteria. The word "positional" refers to the static aspect of the board. When you quickly evaluate a position and begin by counting material, you do not simultaneously try to analyze ten moves ahead to see if someone can win something. You just count the material in the current position, statically. I tried to make this distinction about Elements addressing piece, and not position, evaluation clear in the subtitle, the back cover, and the Introduction to Elements. Nevertheless, one reader gave the greatly expanded fourth edition of Elements a low rating in an online review because he wanted to learn about evaluating positions, and he was disappointed when the book delivered on its promise and did not primarily address his concern!
I hope addressing both piece evaluation and position evaluation criteria together in the same column will go at least a short way toward preventing any future misunderstanding.
It should be clear that some concepts should apply more to positions than pieces and vice versa. For example, speed is simply a measure of how fast a piece can move across the board and cover squares. It makes a lot of sense that the power of a piece should include its speed, but the "speed" of an entire position is not so clear. We could say that if all your pieces remaining on the board are the fast ones (bishops, rooks, and queens, excepting the ever-present king), then you have a fast army as well.
In Elements I explain the difference between mobility and activity:
Activity and mobility are related. While it is possible that a piece can have high activity and low mobility and vice versa; generally, there is a high correlation between mobility and activity. It is interesting that I chose to make activity an exclusive element of positions and mobility an element of a piece. There is some argument that this distinction is arbitrary, but there is also some sense in it.
When you are considering positions, the mobility of all your forces (i.e., the total number of moves at your command) is an interesting trait, and higher mobility correlates with many positive factors. However, I think all strong players are much more interested in the activity of their pieces; for example, a piece that can only attack a few squares around the enemy king (or central squares, or anything tactical or vital) is almost always more valuable than one with lots of access to relatively inconsequential squares.
At the piece level, things are not quite as clear, as mobility and activity are both important. However, it turns out that a piece's activity is tied in with other piece elements like central control, flexibility, and coordination. In that sense activity is less of an independent element at the piece level; it is more like a conglomeration of several aspects of multiple elements. Thus, army activity became part of evaluating positions, but more an output of the elements in evaluating individual pieces. If you prefer thinking of activity as a piece element too, I understand.
Another interesting aspect is time (i.e., tempos, not clock time, although the latter is an off-the-board vital aspect of evaluating on which player you should bet!) On the surface, I could easily argue that time is more a positional feature than a piece feature. Early in the game, the minimum number of moves to reach a given configuration can be considered the "time" – when one players wastes tempos, his opponent has a time advantage, which optimally should mean more pieces developed. For more on the "tempo" aspect, see The Value of a Tempo.
However, after the opening the concept of development, as applied to positions, starts to fade. What we are really interested in is the activity of the relative forces. But for each piece, time is often a factor: how much time will it take to get a given piece to perform a given function? There is a great amount of overlap between how much time it would take to have a piece perform a given function and how much time it would take to get the entire forces to perform a function (say promote a pawn). Again, I had to cut this distinction in some way to factor it in the evaluation, and I chose to place it as a piece issue.
The important issue, as I discuss at length in Elements of Positional Evaluation, is whether the elements cover the intended object (evaluation) adequately. If together the elements lack the ability to evaluate certain situations, then the coverage is inadequate and additional or more powerful elements are needed. If the elements "over-cover" the object and are redundant, then weaker or fewer elements are needed. In both cases (evaluating positions and individual pieces) the elements I chose have the desired coverage, and perfection may be an impossible target.
For example, in Activity is the Real Goal, I discuss the role of space in conjunction with activity when evaluating positions. In that column I note that the goal of space is to confer activity; if it succeeds, as it usually does, then space is a big plus; if it fails, as it can, then having more space can backfire. So the real goal for your pieces is activity, and, as a big bonus, activity covers many more aspects of evaluating positions than space. As a result, activity became an element of evaluation, and space an addressed concept.
Of course, there are other ways to slice the pie; many other authors have addressed evaluating positions; fewer address piece evaluation. Each author chose their own "coverage" of evaluation, and the list of components often vary. That is acceptable, so long as the author defines his terms clearly, creates adequate coverage, and you can use his list in a practical manner to help you evaluate.
When I ask my students to evaluate a position, I often list the four elements listed earlier: material, king-safety, activity, and pawn structure. In doing so, I can always (at least so far!) cover all the necessary aspects. If the student chooses to bring in another aspect, such as tempos or space, in order to clarify his meaning, that is quite acceptable.
Flexibility and vulnerability can both be used in discussing piece and position evaluations, but let's first address coordination, which was the subject of last month's column. At first glance it seems clear: since coordination, almost by definition, requires multiple pieces, it should be a position evaluation element, and not a piece element – and it could be, as stated above.
However, at the position level, coordination, like space, confers activity. That is an example of how activity is so subjective and wide ranging – it covers a multitude of concepts. But the value of each piece, taken individually, often varies quite a bit by how well it coordinates with the other pieces: is it blocking your other pieces? helping attack a square or enemy piece? helping coordinate to block the retreat of an enemy king's retreat? In The Evaluation of Material Imbalances, GM Kaufman discusses the value of the bishop-pair, and he notes that one reason they coordinate well is by never getting in the way of each other, while rooks and knights can. His designation for a pair of knights that guard each other is redundant knights, implying a negative factor.
Bottom line: if you would rather assign the concept of coordination exclusively to position evaluation instead of for pieces, that is fine; in my case it just fit better the other way around.
Good coordination means the whole is greater than the sum of its parts. Ironically, this would mean that if you simply took the average value of each piece on a position, that a well-coordinated position should indicate an overall advantage greater than the material balance between the two forces. Alternately, we could assign higher values to each piece that contributes to the coordination, so that the sum would remain greater than their "base" material value.
Flexibility is another interesting element. I have addressed piece flexibility and plan flexibility, but less so position flexibility. Yet we could easily define position flexibility as plan flexibility, so there is really only a semantic difference (as there is with several aspects of these concepts). With piece flexibility, a piece can do multiple things, or possibly just one thing in multiple ways. For example, a piece that controls squares in the center, but also has an eye on the possibility of attacking an enemy king is likely more valuable than one that can only perform one of those functions.
Similarly, at the plan level, a plan that allows for many future possible positive paths is, in general, superior to one the burns the bridges and either works or goes down in flames if the main idea crashes. As I wrote in Elements, I often found that when my plans lacked flexibility – usually out of necessity, not choice – then I was probably on my way to defeat. On the other hand, when I was faced with a "confusing" array of tempting choices, that was usually an embarrassment of riches, and not the result of an inferior position.
I am not going to address all the evaluation elements in this short column; that is the reason Elements was written. However, I hope I have clarified some of the differences between evaluating entire positions and determining how good a particular piece is in any given position. A final word of warning: when comparing individual piece placement with the position that contains them, sometimes the whole can be greater – or less – than the sum of the parts, just as it was with coordination.
For those interested in testing their skills evaluating positions, take the Evaluation Quiz.
Question How do you spot five-move combinations? I am doing the rapid chess improvement plan and its working except I cannot spot five-move combinations.
Answer Five moves as in 5-ply or 9-ply? A five-move combination is 9-ply, which is not short. I assume when you write "Rapid Chess Improvement Plan" you are referring to Michael de la Maza's book. I like his approach with one notable exception: I do not agree with Michael with regards to repetition of more difficult combinations. I describe my difference in A Different Approach to Studying Tactics.
The longer the combination, the more good visualization is required and – almost always – the more possible permutations that are involved. If you do not have the requisite visualization, you cannot "spot" anything. You also need a good feel for The Seeds of Tactical Destruction, so you know where on the board a combination might be possible.
Also, there is an entire chapter in Andrew Soltis' book Studying Chess Made Easy about looking two-and-a-half moves ahead (5-ply). I suggest you pick up a copy.
Question What is the difference between calculation and candidate moves, and how do I learn to calculate?
Answer A candidate move is one that potentially does something for you; i.e., makes some sense. It is one you can or are considering for your move. It does not have to be safe. See Initial and Final Candidate Moves.
Calculation is usually considered the part of analysis where you look at forcing lines: checks, captures, and threats. Other authors use calculation and analysis synonymously.
They write entire books (series of books!) about learning how to analyze and calculate. Part of it is pattern recognition from studying many (simple) tactical problems and reading many annotated game books. Other skills are less knowledge and more abilities. You get some of these skills by playing stronger players and then analyzing the game with them carefully afterward. See my book The Improving Chess Thinker, Soltis' How to Choose a Chess Move, or all my columns on analysis.
Question My child seems to have plateaued around 2000. She studies regularly, does chess tactics, plays in tournaments – the same routine that got her where she is, but this routine does not help her improve anymore. I am still waiting for a breakthrough, hoping that developing mature thinking will affect her chess as well. I am wondering if this is something you have seen in your practice? What should we expect, what can we do? I would greatly appreciate your thoughts.
Answer Thanks. Without seeing someone play (time management, enthusiasm, etc) it is tough to say why someone plateaus at a particular level. The higher you go, the more you need specific information. Beginners do not need specific information, such as openings and endgames, but players rated 2000+ do.
Also, the ability to find good competition, play steadily, and review games afterward is paramount as you get stronger. You need feedback from opponents (or a strong friend or coach) as to what you are doing and what they know that you do not to pull yourself up. In that sense what someone does to get from 1600 to 1700 is quite different than what they might do to go from 2100 to 2200.
Finally, enthusiasm and fun is key. To work that hard when you get strong becomes more and more difficult. Many youngsters understandably are not monomaniacal about chess (as happened with almost all my strong students like Dan Benjamin and Matt Traldi) and that is good, but not so good for their chess.
Question I am starting a blog to teach chess; can you give me some tips?
Answer Know your intended audience: are you writing for young children, their parents, or everyone?
Here are some tips:
You can find many more chess-writing tips in the Chess Journalists of America Newsletter. For example, I wrote an article about creating instructional chess material.
All of the newsletters are available via the website chessjournalism.org. You can scan their Table of Contents on the second pages of the PDFs.
If you are instructing young beginners, I have an article on that, too: Teaching Young Beginners.
Question Chess problems and puzzles are ubiquitous, but I have never seen a published methodology for solving. Can you recommend a flow sheet or algorithm to use when viewing a new puzzle?
Answer Good question. It depends on what kind of puzzle it is. You do not use the same algorithm for mate-in-two that you would for a general play-and-mate, and you would not use the same for play-and-win and play-and-draw, etc. See Understanding Chess Puzzles for more.
Similarly, you do not use the same algorithm for finding the best move possible in a reasonable amount of time vs. doing a puzzle, which is why Dr. Adriaan de Groot wanted to use positions from his games, and not puzzles, to see how grandmasters really think. How you think in games is more relevant than puzzles, because in puzzles you have a specified goal, while in games you have to make your own goal(s). This required an entire PhD thesis (!) devoted to this question, published in book form as Thought and Choice in Chess.
Novice Nook #158 (Ebook)
In PDF format. Viewable in Ipad, Itouch, Kindle, and other PDF viewers.
© 2014 Dan Heisman & BrainGamz, Inc. All Rights Reserved.
Dan welcomes readers' questions; he is a full-time instructor on the ICC as Phillytutor.
Comment on this month's column via our official Chess Blog!
Purchases from our
Home Page] [ChessCafe
© 2014 BrainGamz, Inc. All Rights Reserved.