Strategy for Perfect Play

In my previous two blogs I discussed the First Scientific Theory of Chess and Steinitz’ Theory of Perfect Play.  To briefly restate his theory, it is simply that if both players were to play perfectly the natural result of the game would be a draw.  In my last blog I offered some initial empirical evidence that supports the correctness of the theory.

Now I continue these explorations by discussing how chess strategy also conforms to Steinitz’ theory.  It is important to note that this discussion of strategy is not evidence for Steinitz’ theory, only that the strategic ideas we see are in accordance with his theory.

In 1885 when Steinitz was at the peak of his prowess, Lasker was born.  Not his more famous friend Emanuel Lasker, who succeeded Steinitz to the throne of the world chess championship, but the less famous International Master Edward Lasker.  In 1911, Lasker the Lesser wrote an influential book on chess strategy called Schachtheorie which was translated into English four years later. 

In Chess Strategy, Lasker pays vague homage to Steinitz’ theory when he writes, "if both sides have succeeded by careful play to preserve equality of material, a draw will generally ensue."  Lasker defines strategy by writing that, "In each game the strategy of chess should set us the tasks which must be accomplished."  

Nearly a century later Grandmasters Lev Alburt and Sam Palatnik are explicit in their tribute to Steinitz’ continuing influence on modern chess strategy.  In their Chess Strategy for the Tournament Player they write that "The basis of modern positional, or strategic, play is the theory of the first World Chess Champion, Wilhelm Steinitz."  In their definition, "Strategy is the art of forming an overall plan."  They go on to say that "the words positional and strategic are frequently used interchangeably." 

This view is nearly identical to the definition provided by the Oxford Companion to Chess which says that “Strategy [is] the planning and conduct of the long-term objectives in a game.  Moves directed primarily towards this end are commonly referred to as positional play.”

The positional nature of strategy is a logical consequence of Steinitz’ Theory of Perfect Play.  This is true simply because while perfect play will result in a draw, another simple truth is that even the best players in the world play imperfectly. “The greatest human weakness is inconsistency,” writes Garry Kasparov in his essay An Evolutionary Theory of Chess (which does not qualify as a scientific theory of chess, by the way).

Given that even the best players suffer from weaknesses in their play, the positional nature of strategy follows.  This could be stated as a kind of general guideline:  Strengthen your position by systematically eliminating your own weaknesses with every move, and be prepared to capitalize on your opponent’s inevitable error. 

This guideline is certainly easier to state than it is to follow.  To strengthen your position it is necessary to understand the general as well as the specific principles of strategic or positional play, and to this end are the great variety of texts on chess strategy to be found. 

Alburt and Palatnik begin their own text on strategy by summarizing Steinitz’ nine basic elements that are key to properly understanding a position.  Those elements are:

1.  Development

2.  Mobility

3.  Center Control

4.  King Position

5.  Weak and Strong Squares

6.  Pawn Structure

7.  Queenside Pawn Majority

8.  Open Files

9.  Advantage of the 2 Bishops

 It is quite interesting to compare this list of ‘elements’ to Jeremy Silman’s list of seven ‘imbalances’ in a position, which is not dissimilar to Steinitz’ list:

1.  Space

2.  Control of Key Squares

3.  Initiative

4.  Material

5.  Minor Piece Superiority

6.  Pawn Structure

7.  Development

 Steinitz himself proposes a strategy that has been dubbed by many as the theory of the accumulation of small advantages.  The idea is to maneuver to obtain a small, but perceptible advantage versus your opponent.  This may not be enough to secure the win, just as the advantage of first move is not enough to guarantee a win for White.  But the combined strength of two or more small advantages will lead to a significant result, in the same way that Senator Everett Dirksen noticed in monetary policy that “A billion here, a billion there – pretty soon, you're talking real money.”

Steinitz’ statement of the accumulation of small advantages is not quite as memorable as Dirksen’s, but has certainly been very influential in chess, if not finance.  Steinitz advocates “steady development without any sacrifice of material, circumspective attention to the balance of forces and of position on all parts of the board, and the accumulation of small advantages if possible.”

Here is a fine example of Steinitz obtaining an accumulation of small advantages in an 1882 game against Blackburne.  After move 17 notice how Steinitz has the following advantages:

1.     More space.

2.     A center pawn.

3.     Half open files for both rooks, giving them great mobility.

4.     Two bishops vs. knight and bishop.

Notice the positional nature of the play that continues from this point, leading to yet other advantages.  I have added annotations in the moves to point out some of these notions.

Kasparov writes that, “By 1870 Steinitz had begun to develop his advanced theories of defense, weaknesses, and strategic play. This is what divides the chess timeline into “pre-Steinitz” and “post-Steinitz” periods.”

Certainly the game above illustrates a beautiful win for Steinitz, who played in accordance with his own theory.  Small errors of play by Blackburne led to small advantages for Steinitz, who knew how to take advantage of the opportunities provided to him with appropriate tactical play. 

As Edward Lasker writes, “Sound strategy, when setting the task, must never lose sight of tactical practicability, and only a thorough knowledge of tactical resources makes correct strategy possible.”  I will write about this interplay of strategy and tactics as I continue to pursue the brilliance of Steinitz in a future blog.


  • 7 years ago


     nice game!

  • 7 years ago



    Ok, but it doesn't have anything to do with the nit I was picking. :)

  • 7 years ago


    @ dmvdc : "It’s just you and your opponent at the board and you're trying to prove something."  Bobby Fischer

    Think about it.

  • 7 years ago


    ...but in chess, as in war, to not lose is to win. If your opponent agrees to a cease fire and peace (Draw) it is ALMOST as nice as conquering them.

    That's about as wrong-headed as you can possibly get when you're talking about war (not least b/c war is not necessarily about "conquering"). Go re-read your Clausewitz.

  • 7 years ago


    I said it before and I will say it again. Black and white board, black and white pieces, dominate and submissive, male and female, lead roll and follower role positive and negative, dance, Olympic mat Wrestling, Yin and Yang.

    Think about it!

    Its not that complicated people?

  • 7 years ago


    White_claw. I like your thoughts, equal might be an easier word. This would allow for style differences, ability, etc. as long as they are perfectly matched- they are equal. As far as the tactics of one side having an advantage, if the 1st attacker has less pieces to attack that square, the defender wins, but if it is equal play, both sides would have an equal chain reaction, meaning the 1st to take would be the winner. This goes for any multiple exchanges, make sure you have more pieces defending the square, then take in the proper way as to continue the attack. But this leads to the fun and dubious in-between moves as well as many other positional opportunities if one does not go along with the planned trade off.

  • 7 years ago


    Actually, as said before, perfect play (or equal (for if one side plays perfect, other does too, or in other words if other side doesnt play perfect then other side doesn't too, making a game equal to perfect as same mistakes will be made, which means it would be the same as making no mistakes, only the level of the game differs)) will result in white having to give one of his pieces for capturing and if black does capture, white has to recapture ( i think that if it is not the case game can't be equal and, as explained above (if you understood me xD), it can't be perfect play on both sides) so actually it would make black one move ahead (theory, i don't know if i'm right, you can say i'm new to chess :S) and from this point either black will have to give one of his piece away and recapture (and the game would go on and on until one side was left with a king and 1 piece standing (too bored to count which one) and that side would win (logically))or the game ends in black winning (as i said i can't be sure as i'm new to chess).

  • 7 years ago


    After some thought, it seems clear to me that, contrary to Steinitz thoughts on perfect play and the corresponding strategy that seems linked to it, chess players would be better off NOT making the 'perfect' move, at least sometimes.  If so, then any chess strategy that only and always endorsed only the 'perfect' move would be necessarily incomplete and flawed.  In my blog, I try to defend this thesis as a response to Steinitz and his perfect-play followers.

  • 7 years ago


    arashi_star has an interesting thought. Although one could also argue, that white would be able to make the exchange on his terms and if his tactics are accurate, he would end with one piece or pawn ahead, meaning the winning advantage. 

    So the chicken or the egg debate continues... 

    @SpaceOddity - to answer your final thought "why does the 'perfect play' 'theory' really matter when considering the 'strategy' that Steinitz (or anybody else) offers? "

    We all want to win in chess, but in chess, as in war, to not lose is to win. If your opponent agrees to a cease fire and peace (Draw) it is ALMOST as nice as conquering them. 

  • 7 years ago


    k, I hate to burst anyones bubble, but black technically should always win with perfect play from both sides.  Your probably wondering "whats your logic?"  Well this is the logic: is since there are 64 squares on the chess board (an even number), it will end up being that white will (with perfect play) be in a position against black where after enough moves, he will have to do a move where in the end he will have to give up material and lose the game.

  • 7 years ago


    Steinitz's "incremental advantages" idea is incredibly important. It's how modern chess is played, OK. Everyone agrees. Awesome. 

    The notion of perfect play, however, is interesting but lacks application because people (and computers--as of now) aren't perfect. So while I see the theoretical aspects of perfect play, if you take a step back, it's kind of questionable. 

    People aren't perfect. So all the data we get doesn't lead us to a strong conclusion. Maybe we can say that in high level play, draws are more likely to come. That makes sense. But saying perfect play leads to draws when people aren't playing even close to perfect is a leap in logic. What if perfect play leads to white winning? Who would ever know? There's no data to support it. And that's where Steinitz get's into trouble for me. 

  • 7 years ago


    Great blog. The issue of whites advantage has been argued longer than any of us have been alive. I believe @SpaceOddity was onto a good comparison with that of tic-tac-toe. Being a shorter game with less variables we can see that, if it is not played perfectly by both sides, the 1st player generally wins because of his beginning advantage. But, as long as the second player pays attention to his opponents moves-a draw will always ensue. This plays to black having an advantage in chess because he is able to see his opponents plan and counter it each move. 

    But in chess, with its exponential variables, the momentum can change with one wrong move, and we do not always see what our opponent is planning, therefore we make slight mistakes by trying to push our own agenda. If a persons entire goal is to negate the plans of his opponent, it will be a draw.

    I point to Karpov for this example, as he is one of the best strategic/positional players of our time, and consequently is one of the most drawn players in the game. Obviously if you are a great player you will have more wins than losses, but IMO draws are seen as weak to us because of our pride, which is also why we make mistakes, because we think we are better than our opponent. 

    One of my favorite quotes was by the 15yr old Fischer before his 1st candidates match. He told journalist Miro Radoicic, "I can draw with the grandmasters, and there are half-a-dozen patzers in the tournament I reckon to beat."[81] 

    When we think our opponent is better, we generally just try to stop him from beating us... So in theory, if we were to swallow our pride and humble ourselves from the "YOU PLAY TO WIN THE GAME!" philosophy and play less aggressive, making smaller safer moves, hoping for an opening to attack, we would be utilizing Steinitz theory. 

    Too often do we think perfect play means a win. But if we look at any sport-when both sides are extremely good and even, the game is always extremely close, coming down to clutch moments, great moves, amazing catches etc. taking advantage of one misstep, one mismatch, one weakness. 99% of these examples include an extended play (overtime, blitz or Armageddon game, etc.) some squeak it out in regulation by an amazing play, or fall just short by pushing too hard and opening a weakness themselves.

    I will suggest this again, create a collection of as many computer v computer games at their highest setting. Using the same program will most likely yield the same result, but even different programs, when set to the highest settings, should draw. Although some are obviously better programs-you will probably have several wins as well because of one minor mistake.

    Keep the discussion going, you're combining my 3 favorite things, history, chess, and philosophy. 

  • 7 years ago


    I don't know that I am right but because white is always one move ahead it seems to me that white should always win with perfect play. Just a thought.

    And the one who makes the first mistake should lose if the other player plays perfectly.

  • 7 years ago


    thanks for honoring one of the greatest underestimated players of all time.

  • 7 years ago


    I think Steinitz theory of perfect play reminds us that early aggression should only be successful as the result of a significant error of the opponent as opposed to a significantly good move. Or at least that is what I'm hearing.

  • 7 years ago


    I find the history of Steinitz thoughts and his influence quite interesting.  I appreciate your research in showing especially how even the likes of Kasparov make reference to Steinitz in this light and the lasting influence of his writings on strategy. 

    What I don't understand, however, is the relevance of Steinitz thesis (or hypo-thesis, in the strict sense) that perfect play leads to a draw.  Steinitz offers advice on strategy to help a player win, since winning is the goal of chess (or as that ex-NFL coach famously ranted, "YOU PLAY TO WIN THE GAME!").  That said, we play to win whether chess is a draw, if played perfectly, like tic-tac-toe, or if chess is a win for white if both sides play perfectly.  So the hypothesis of perfect play leads to a draw falls out as irrelevant to the strategic advice regarding how we, imperfect as we are, should play the game.  The strategic elements of chess that Steinitz outlines are indeed helpful and insightful, but they are helpful and insightful regardless of the underlying hypothesis on perfect play.  Of course, we want to capitalize when our opponents make mistakes, and that's true of just about any competitive game, sport, war or situation. 

    So my question is--why does the 'perfect play' 'theory' really matter when considering the 'strategy' that Steinitz (or anybody else) offers? 

    A strategy is good or bad to the extent that it helps players become better, ie, win more, win faster, win more easily and, as the case may be, draw less in favorable positions and draw more in unfavorable positions. The perfect play theory itself seems more of a side-note to the real strategy that Steinitz offered and does not seem to be essential or necessary to that strategy.

Back to Top

Post your reply: