Opening Theory
Opening theory is the accumulated body of knowledge about the first moves of a chess game.
6 frameworks, from 1600 to today · Chess · Games
Overview
Opening theory is the accumulated body of knowledge about the first moves of a chess game1. Its central question is deceptively simple: once the pieces leave their starting squares, what should each side try to do, and why? A player who knows theory can rattle off ten or twenty moves of a well-studied line; a player who does not may lose before the middlegame begins. But opening theory is not a fixed script. It is a living argument about what a good position is, and that argument has changed repeatedly over five centuries—from the swashbuckling gambits of the Romantic era, through the scientific positional school, the hypermodern reaction, the Soviet systematization of preparation, and finally the engine and neural-network revolutions that have overturned much of what humans thought they knew.
The Romantic Era: Gambits and the Cult of Attack
Chess theory in the modern sense begins in the late fifteenth century, when the game's rules took their present form—the queen and bishop gained their long-range powers, and castling and the double pawn move appeared. The first printed opening treatise, by the Spanish bishop Ruy López de Segura (1561), already analyzed what we now call the Ruy López (1.e4 e5 2.Nf3 Nc6 3.Bb5). In the seventeenth and eighteenth centuries, Italian masters such as Gioachino Greco built a tradition of brilliant attacking play, and the "Italian Game" (1.e4 e5 2.Nf3 Nc6 3.Bc4) became the standard opening.
The dominant style of the eighteenth and nineteenth centuries is now called the Romantic school. Its practitioners prized rapid development, open lines, and sacrificial attack2. The King's Gambit (1.e4 e5 2.f4), offering a pawn for the center and open lines, was the era's signature opening13. The model game was the "Immortal Game" between Adolf Anderssen and Lionel Kieseritzky in London in 1851, in which Anderssen sacrificed both rooks and his queen to mate2. The Romantic assumption was that material could be given up for initiative and that the attacker's chances justified almost any risk2. What the Romantics could not solve was the problem of defense: against a cool-headed opponent who declined the sacrifices and consolidated, the gambiteer often simply lost material.
Steinitz and the Classical Positional School
The first great challenge to Romantic assumptions came from Wilhelm Steinitz, the first official world champion (1886). Steinitz argued that a sound attack must rest on a positional foundation—accumulated small advantages such as a better pawn structure, control of key squares, and the two bishops4. His teaching, developed in the 1870s and 1880s and codified in his writings, became the Classical positional school. Its central idea was that the attacker should not strike until the position is ripe; premature attacks, Steinitz held, rebound on the attacker3.
A concrete illustration is the treatment of doubled pawns. The Romantics saw doubled pawns as a mere blemish; Steinitz and his successors taught that they could be a lasting weakness, especially if fixed on an open file, because they cannot defend one another and obstruct the pawns beside them. Steinitz's successor as champion, Emanuel Lasker, extended the approach with psychological and practical flexibility, and Siegbert Tarrasch popularized it in widely read books, insisting that certain opening principles—occupy the center with pawns, develop quickly, avoid premature pawn moves—were nearly laws of nature.
The classical school's problem was dogmatism. If the center must be occupied by pawns, what was one to make of openings that deliberately allowed the opponent a big pawn center in order to attack it later?
The Hypermodern School: Attacking the Center from Afar
The answer came from the Hypermodern school, whose leading figures were Richard Réti, Aron Nimzowitsch, and Gyula Breyer, writing in the 1910s and 1920s. Their claim was that the center could be controlled from a distance—by fianchettoed bishops on the long diagonals, by knights and pawns harassing the opponent's pawn center—rather than occupied immediately. Nimzowitsch's My System (1925) set out the doctrine in detail, including the idea of "overprotection" (defending a key point more times than seems necessary) and "prophylaxis" (preventing the opponent's plans before pursuing one's own).
The practical test came in the 1920s. Réti's opening 1.Nf3 followed by a fianchetto, and the King's Indian Defense (in which Black allows White a broad pawn center and then strikes at it with ...e5 or ...c5), embodied the new thinking156. The hypermoderns did not refute the classics; they widened the repertoire. Their contribution was to show that pawn centers can be targets as well as assets, and that flexibility and indirect control could be as valuable as immediate occupation1.
The Soviet School: Opening Preparation as a Science
The next transformation was organizational as much as intellectual. From the 1930s onward, the Soviet Union built an unprecedented system of chess training, state sponsorship, and collective study. The Soviet school of opening preparation treated the opening not as a set of principles to be understood but as a body of concrete variations to be mastered, extended, and—crucially—kept secret until a tournament game revealed it.
The method was systematic: teams of players and analysts worked through candidate moves with boards and notebooks, cross-checked lines against published games, and prepared "novelties"—new moves in known positions—for specific opponents. Mikhail Botvinnik, world champion from 1948 to 1963, was the emblematic figure, combining deep preparation with a scientific temperament. The approach produced a generation of players—Vasily Smyslov, Tigran Petrosian, Boris Spassky, Mikhail Tal, and later Anatoly Karpov and Garry Kasparov—whose opening knowledge was formidable. The Soviet system's problem was that it depended on human memory and human analysts; the volume of variations eventually outran what any team could hold.
The Engine Revolution: Silicon Analysis
The computer changed the economics of opening preparation. From the 1970s, chess engines grew steadily stronger, and by the late 1990s and 2000s programs such as Fritz, Junior, and Rybka were beating strong grandmasters in analysis. The decisive public moment was Deep Blue's match victory over Kasparov in 1997, though that machine was a specialized search engine rather than a general analysis tool. What mattered for opening theory was that engines could evaluate millions of positions per second and never forgot a line.
Engine-driven opening analysis had two effects. First, it made concrete calculation cheap: a player could test a novelty against a machine and learn within minutes whether it held. Second, it exposed human assumptions as errors. Lines that grandmasters had considered sound for decades were shown to be dubious, and obscure variations were rehabilitated. The "computer opening" became a genre of its own, with players preparing lines chosen for their engine evaluation rather than their human comprehensibility. The limitation was that early engines evaluated positions by handcrafted criteria—material, mobility, king safety—and could be systematically wrong about long-term positional factors.
Neural Networks: Learning from Self-Play
The most recent turning point came from machine learning. In 2017, the AlphaZero program, developed by DeepMind, learned chess from scratch by playing against itself, using a neural network guided by Monte Carlo tree search rather than handcrafted evaluation. It defeated the strongest conventional engines of the time. Its successor, Leela Chess Zero, an open-source project, continued the approach and became a standard analysis tool.
Neural-network opening assessment differs from earlier engine analysis in kind. Instead of evaluating positions by explicit rules, the network learns patterns from millions of self-play games and assigns values that often surprise humans—favoring, for instance, certain quiet moves and long-term structural features that classical evaluation undervalued. The practical consequence is that opening theory is now partly written by machines: lines are tested and refined against neural engines, and human players study the resulting evaluations. The open question is how much of this machine knowledge humans can absorb and whether it will produce a new synthesis or simply a deeper, less comprehensible book of variations.
The Present: A Layered Discipline
Opening theory today is not one method but a stack of them. Players still learn classical principles—develop pieces, fight for the center, safeguard the king—because those principles remain the best guide for a human trying to understand an unfamiliar position. Hypermodern ideas are standard repertoire: the King's Indian, the Grünfeld, and the Réti are played at every level. Soviet-style preparation persists in the form of professional teams and databases, now digitized. Engines and neural networks are indispensable tools for checking and extending lines.
What remains open is the question that has driven the field since Ruy López: how much of the opening can be reduced to memorized variation, and how much must remain a matter of judgment. The engines have pushed the boundary of memorization further than anyone in the Soviet school imagined, yet the strongest players still win by understanding why a move works—and by finding, in the space between the machine's evaluation and the human's intuition, the novelty that no one has played before.
References
- Chess opening (introduction, Open Games: 1.e4 e5, Common aims in opening play), Wikipedia
- Romantic chess (introduction, History), Wikipedia
- King's Gambit (introduction, History), Wikipedia
- Hypermodernism (chess) (History), Wikipedia
- Réti Opening, Wikipedia
- King's Indian Defence, Wikipedia
The frameworks of Opening Theory
Each framework is a way of working the field has used: when it held, and the question it answers.
- 1600-1880
- 1870-1920
Classical positional school
What positional foundations make an attack sound?
- 1910-1950
- 1930-1990
Soviet school of opening preparation
How should opening variations be prepared and kept secret?
- 1990-Present
Engine-driven opening analysis
What does a machine calculation say about this line?
- 2017-Present
Neural-network opening assessment
What patterns do self-play networks learn about positions?
How the ideas connect
Each arrow runs from an earlier idea to the later one it shaped. The list below says what passed between them.
Classical positional school reacted against Romantic chess
Steinitz built the classical positional school by attacking Romantic assumptions: he argued that sacrificial attacks must rest on accumulated positional advantages and that premature attacks rebound on the attacker. His writings of the 1870s and 1880s defined the new school against the gambit play of the Romantics.
Carried by Wilhelm Steinitz.
Hypermodernism reacted against Classical positional school
Réti, Nimzowitsch and Breyer defined hypermodernism by rejecting the classical dogma that the center must be occupied by pawns immediately, arguing instead that it could be controlled from a distance and attacked as a target. Nimzowitsch's My System (1925) set out the doctrine as a direct challenge to Tarrasch's 'laws of nature'.
Carried by Richard Réti; Aron Nimzowitsch; My System (1925).
Engine-driven opening analysis replaced Soviet school of opening preparation
Engine-driven analysis replaced the Soviet system's human teams and notebooks as the working method of opening preparation: engines could test a novelty in minutes and never forgot a line, making the volume of variations outrun what any team could hold. The decisive public moment was Deep Blue's 1997 match victory over Kasparov.
Carried by Deep Blue; Fritz; Rybka.
Neural-network opening assessment replaced Engine-driven opening analysis
Neural-network assessment replaced handcrafted engine evaluation as the strongest analysis tool: AlphaZero in 2017 learned from self-play and defeated the strongest conventional engines, and Leela Chess Zero continued the approach. Instead of explicit rules like material and mobility, the network learned patterns that often surprised humans.
Carried by AlphaZero; Leela Chess Zero.
Soviet school of opening preparation influenced by Classical positional school
The Soviet school took up the classical positional school's emphasis on pawn structure and long-term advantages as the foundation for concrete opening preparation, combining it with systematic variation analysis. Botvinnik, its emblematic figure, was trained in and built on classical principles while treating the opening as a body of concrete variations.
Carried by Mikhail Botvinnik.
Soviet school of opening preparation influenced by Hypermodernism
Soviet opening preparation absorbed hypermodern openings such as the King's Indian and Grünfeld into its repertoire and subjected their variations to systematic analysis. The hypermodern widening of the repertoire gave the Soviet school concrete systems to prepare and refine.
Beyond the field
Where Opening Theory's ideas came from in other fields, and what they led to there.
Romantic chess provoked Steinitz's Positional Theory in Middlegame Theory
Steinitz's theory arose as a rejection of the Romantic gambit and attack style dominant in opening play, replacing it with positional preparation and sound pawn play. The new doctrine defined itself against that earlier approach.
Carried by Wilhelm Steinitz.
Romantic chess provoked Classical positional school in Chess
The classical positional school defined itself against the romantic attacking style, with Steinitz arguing that sound accumulation and defence replaced the sacrificial gambit play of the earlier era.
Carried by Wilhelm Steinitz.
Classical positional school reacted against Romantic chess in Middlegame Theory
Steinitz built the positional school by attacking Romantic gambit play, arguing that sound development, pawn structure and accumulation of small advantages beat speculative sacrifice; the school defined itself against the Romantic style.
Carried by Wilhelm Steinitz.
Hypermodernism reacted against Classical Positional School in Middlegame Theory
Nimzowitsch and Réti framed Hypermodernism as a rejection of Tarrasch's classical centre-occupation dogma, arguing for controlling the centre from the flanks with fianchettoed bishops and allowing the opponent a broad pawn centre to attack.
Carried by Aron Nimzowitsch; Richard Réti.
Hypermodernism reacted against Classical positional school in Chess
Hypermodern opening play arose as a direct attack on the Classical positional school's rules about occupying the centre with pawns, replacing them with flank control and piece pressure.
Carried by Aron Nimzowitsch; Richard Réti.
Engine-driven opening analysis influenced by Search-and-evaluation engine in Computer Chess
The alpha-beta search-and-evaluation engines developed from the 1950s onward became strong enough in the 1990s that players began preparing openings by consulting engine output, which is the basis of engine-driven analysis.
Engine-driven opening analysis influenced AI-informed opening theory in Fuseki Opening
Engine-driven opening analysis in chess established the practice of mining strong engines' evaluations for opening novelties. AlphaGo's creators and later Go analysts applied the same engine-mining approach to Go, using superhuman programs to generate and test opening moves.
Carried by AlphaGo.
Neural-network opening assessment replaced Engine-assisted preparation in Chess
Neural-network assessment superseded conventional engine-assisted preparation: where players once mined alpha-beta engine lines and databases, AlphaZero-style networks produced positional judgements that overturned established opening evaluations.
Carried by AlphaZero.
Neural-network opening assessment influenced by Neural-network policy and value learning in Computer Go
Neural-network policy and value learning developed in computer Go, culminating in AlphaGo and AlphaZero, was transferred to chess and produced the neural-network opening assessments that replaced handcrafted evaluation.
Carried by AlphaGo; AlphaZero.
Romantic chess provoked Classical Positional School in Middlegame Theory
The classical school consolidated itself by rejecting Romantic gambit play in favour of positional soundness and preparation. Its manuals and teaching defined correct play against the Romantic model.
Carried by Siegbert Tarrasch.
Soviet school of opening preparation influenced Engine-assisted preparation in Chess
The Soviet school of opening preparation established the habit of deep, systematic opening study that engine-assisted preparation continued and transformed with computational tools.
Soviet school of opening preparation influenced by Classical positional school in Chess
Soviet opening preparation inherited the Classical positional school's principles of centre control, piece coordination and pawn structure, which Soviet trainers taught and then extended into concrete opening systems.

Opening Theory is a planet in Nousfarer
Its city has a tower for each of these 6 frameworks. Land, walk in, and climb one concept at a time.
Land on Opening TheoryRelated fields
Fields its ideas are linked to
- Chess 5 links
- Middlegame Theory 4 links
- Computer Chess 1 link
- Computer Go 1 link
- Fuseki Opening 1 link