Game theory studies strategic interaction — how your move depends on what your opponent can do. This page introduces the Game Theory project at Norway Chess, explains its background, and presents our newest innovation, the Norway Chess Bar (Beta).
On this page:
1. Introduction
Game theory studies strategic interaction: how your move depends on what your opponent can do. This perspective fits Norway Chess perfectly. At the top level, players don’t just search for the “objectively best” move; they choose practical strategies that put real pressure on a human opponent under time stress and tournament conditions. The Game Theory Corner adds this human-centred angle to traditional engine tools, helping fans see why certain decisions were difficult, where pressure built up, and how risk eventually paid off.
A brief origin story of game theory:
It wouldn’t be too controversial to say that game theory was born from chess more than a century ago. World Champion Emanuel Lasker was one of the first to think about a general theory of games in his writings. In 1913, mathematician Ernst Zermelo published a paper on “solving chess,” widely regarded as the first academic work in game theory. From those early ideas, the field grew into what we now know as modern game theory.
Chess has also played a major role in the rise of artificial intelligence (AI). From early computer experiments to modern learning engines, chess has always been a testing ground for human and machine reasoning. Today, through the Game Theory Corner, we bring those worlds together again, combining AI with game-theoretical thinking to better understand how the best players in the world make decisions.
↑ Back to top
2. Game Theory Corner
Innovation at Norway Chess:
Norway Chess has always pushed to make elite chess more exciting and accessible. Founder Kjell Madland’s introduction of Armageddon games after classical draws changed how the tournament stays engaging for fans. The Game Theory Corner, led by Dr Seven, continues that same spirit of innovation, bringing a more human perspective into chess analysis.
During the tournament, the Game Theory Corner provides round-by-round insights into players’ practical winning chances and easy-to-understand metrics that reveal how they handle critical moments. Fans can expect short, clear explainers during live rounds and daily statistics that unpack the games.
In the most recent edition, we’ve also been testing the Norway Chess Bar — a new evaluation bar that adds human context to traditional engine analysis. It reflects practical winning chances rather than only “objective” engine evaluations, showing how likely a position is to be won or saved by real human players.
The Game Theory Corner continues to evolve each year, building on the same idea that has always defined Norway Chess: making top-level chess more understandable and more human for everyone watching.
What We Do at The Game Theory Corner
Modern chess has entered what many call the engine era. Computers have changed the game completely — both for professionals and for how fans follow it. With years of preparation and access to powerful engines, top players know the main lines so deeply that many classical games end in draws. The margins for creativity in the opening are smaller, and victories often come from small, well-timed risks rather than outright opponent mistakes.
But something gets lost when we talk only about “accuracy.” Engines judge every move in terms of objective evaluation — how close a player’s move is to the computer’s top choice. For human players, though, that’s not the full story. Chess is played under tension, fatigue, and time pressure. A move that looks perfectly equal to the engine may be extremely difficult to find over the board.
That’s why, at Norway Chess, we look beyond accuracy. We focus on risk and reward — on how players make practical decisions that target their opponents as human beings, not machines. Sometimes the most accurate player isn’t the one who wins the tournament. In fact, recent Norway Chess editions have shown this clearly:
Wesley So had the highest accuracy in both 2022 and 2023, meaning he played closest to engine perfection. Yet Magnus Carlsen (2022) and Hikaru Nakamura (2023) won those events.
So what made the difference? The ability to judge when to take risks, when to simplify, and when to push. That balance — between precision and pressure — is what we aim to measure and share through the Game Theory Corner.
By pairing engine analysis with human-centered metrics, we show how results emerge not only from accuracy but also from the psychology of competition itself. In modern chess, perfection isn’t everything. Understanding when to be imperfect can be the real winning move.
Game Theory Corner During the Tournament:
Before the round begins each day, we provide updated predictions on the players’ tournament-winning chances. These probabilities consider the unique scoring system at Norway Chess, where a classical win earns three points and draws go to Armageddon.
During the rounds, our analysis blends engine precision with human understanding. The engines run in the background, but the Game Theory Corner focuses on what really matters for people sitting at the board: time pressure, practical complexity, and human error margins. A position that the computer calls equal may still be a nightmare to defend when the clock is ticking.
After the round starts, at critical moments, the Norway Chess Bar comes in. Instead of showing only “objective” evaluations, it shows practical winning chances: how likely each player is to win or hold, given the position and the clock. When time gets short, the bar adjusts sharply, reflecting how real pressure changes the odds.
During the breaks, we also provide various chess statistics, including historical and Armageddon winning percentages.
After the round: once the games finish, we provide the performance metrics: Missed Points and Game Intelligence (GI).
Missed Points measure how many points a player effectively “left on the table” compared to the engine’s top line, using more human-understandable language with winning chances. They’re expressed in simple terms — points rather than pawns — so that fans can understand what really changed the result. Fewer Missed Points are better: 0 missed points in a game mean perfect play according to the engine; 1 missed point means making a game-losing blunder in a winning position, whereas 0.5 missed points mean making a game-losing blunder in a drawn position.
Game Intelligence (GI) is a human-centric score that measures a player’s ability to weigh risk against potential reward. High-GI players know when to take chances and when to hold firm; they press in complex positions and stay precise when it matters most. For comparison across events, the average human GI is standardized at 100, with about two-thirds of players falling between 85 and 115. Winners of major tournaments often reach 160 or higher.
Building this model was a major research effort. Dr Seven analyzed over one billion chess moves, including more than a million from the world’s top grandmasters. He also studied every World Championship game from 1886 to 2023, tracking how player precision has evolved through time. The results show a clear trend: overall accuracy has increased, and Missed Points have steadily declined.
A few standout findings:
- Magnus Carlsen achieved a record-high GI score of 161, reflecting his ability to provoke mistakes from opponents through complex and risky play.
- Viswanathan Anand was identified as the most accurate World Champion by average Missed Points per game.
These insights help explain why Norway Chess results often defy simple accuracy charts. Players like Carlsen and Nakamura excel not just by minimizing errors but by creating positions that are hard for humans to handle. GI captures that invisible skill — the art of being practically brilliant.
Practical reading guide
- High GI & Low missed points: clinical and effective (precision and conversion).
- High GI & Medium missed points: well-timed risk that paid off under human conditions.
- Low GI & Low missed points: solid moves, but not enough practical pressure or conversion.
- Low GI & High missed points: risky or imprecise play that didn’t yield results.
↑ Back to top
3. Norway Chess Bar (Beta)
The Norway Chess Bar is one of our newest and most exciting innovations — a fresh way to read chess positions that focuses on human reality rather than engine perfection. Instead of showing an engine’s “objective” evaluation, it estimates practical winning chances: what the odds of victory really look like for a person sitting at the board, under pressure, with the clock ticking.
Norway Chess founder Kjell Madland told Dr Seven about his vision for such a bar when they met, and it took him a long time to bring it to life. After a long period of behind-the-scenes testing, the Norway Chess Bar made its public debut at the 2025 Norway Chess official broadcast, alongside five-time World Champion Viswanathan Anand, Grandmaster Christian Chrilla, and International Master Anna Rudolf.
The Bar quickly became one of the main talking points in the broadcast. It delivered more intuitive and realistic predictions than the traditional evaluation bar, adapting dynamically to the situation on the board, including time pressure. The broadcast hosts and commentators embraced it quickly, even nicknaming it the “Magical Bar.”
Traditional engine bars show how a position stands with perfect play from both sides — something that rarely happens in real games. The Norway Chess Bar, on the other hand, translates that evaluation into practical winning odds, showing how likely a player is to convert or survive given the realities of human play. When time runs low, the bar adjusts because the likelihood of finding the best moves changes.
This approach makes the game easier and more exciting to follow. Instead of saying “0.0” — essentially a drawn endgame with perfect play — the bar might show “White has a 35% chance to win,” giving viewers a more natural understanding of the tension and who is pressing for a victory in the position.
For years, chess fans have watched engine bars rise and fall without really knowing how those numbers relate to human play. The Norway Chess Bar changes that. It’s built to measure practical reality, turning abstract evaluations into relatable probabilities.
The Bar remains a beta project. We have constantly looked at the feedback from grandmasters, commentators, and viewers. Like all things at Norway Chess, it’s driven by curiosity and a wish to make chess more engaging. The early reactions have been overwhelmingly positive, and its introduction marks a new chapter in how the game can be experienced.
↑ Back to top
