What Can a Board Game Teach About Winning?
From the Beyond the Board: Winning Strategies from Go for Business, Competition, and Life collection
Most strategy advice begins with a goal and works backward to a plan. Go, the oldest continuously played board game in the world, begins with something different: a theory of what winning actually means. On a 19-by-19 grid with only two kinds of stones and one rule for capture, two players surround territory. There is no king to trap, no checkmate, no knockout. The game ends when both sides agree the board is settled and the areas are counted. A half-point lead is a complete victory.
That design difference produces a strategic philosophy unlike anything in chess, poker, or conventional business thinking. Henry Kissinger drew the contrast in On China (2011): chess seeks decisive confrontation; Go seeks relative advantage distributed across the whole board. The strongest Go players, human and machine, build many small, secure positions rather than one decisive strike. They treat each stone as an investment that must do the most work for the least cost. And they measure success by reliability, because a narrow win achieved with certainty is worth more than a spectacular win that carried a real chance of losing.
Start where structure does the work
The Chinese proverb jin jiao yin bian cao du pi — gold corners, silver sides, grass belly — captures a geometric insight that applies far beyond the board. Two edges help enclose territory in a corner; one edge helps on a side; the open center has none. Efficient players begin where existing structure does part of the work for them, then expand outward.
The same logic governs where to open a business, where to build a career, and where to invest. Starting in a corner of the market where regulation, customer habits, or partnerships already provide two walls is cheaper than starting in open space and building every wall from scratch. The instinct to go where the action is, to compete in the center, is often the instinct to fight where each unit of effort produces the least return.
Go formalizes this as the distinction between territory and influence. Territory is secured ground — revenue in hand, contracts signed, positions locked. Influence is outward-facing strength — reputation, capabilities, relationships — that pays only when it is later converted into something concrete. Neither is correct in general. The balance between the two is what separates strong from mediocre strategy, and the failure to convert influence into results is one of the most common ways promising positions collapse.
The right to choose the next battle
If whole-board efficiency is the first pillar of Go's strategy, initiative is the second. A move is sente when the opponent must respond, leaving the player free to choose where the next action happens. A move is gote when it ends the player's turn and hands the agenda to the other side.
Strong Go players choose sequences that finish in sente. They play forcing moves, sacrifice material to gain tempo, and sometimes walk away from a local fight entirely — a move called tenuki — because the initiative gained elsewhere is worth more than the local concession.
The concept transfers directly. In negotiation, the side that keeps proposing new terms and opening new fronts controls the shape of the deal. In business, the company that sets the pace of product releases and market entries forces competitors into reactive mode. Microsoft's 2015 write-down of its Nokia phone business, followed by a full pivot to cloud computing, was a tenuki: it gave up a local position to seize initiative on a board where the long-term points were larger.
A related concept, miai, sharpens the logic of optionality. Miai describes two points of roughly equal value: if the opponent takes one, the player takes the other. A position with two miai paths to the goal is safe, because every opponent action still leaves a good answer. The planning principle is clear: build strategies with two independent routes to the objective, so that a rival blocking one simply confirms the other.
Attack to gain, defend lightly
Go is full of fighting, but the strongest fighters rarely aim to kill. The proverb gong bi gu wo — "attack the opponent while watching yourself" — warns that an all-out assault often leaves the attacker's own weaknesses exposed. Professionals attack a weak group mainly to build territory and strength during the chase, collecting profit along the way rather than gambling everything on a capture that may fail.
The complementary discipline is lightness. "Heavy" stones are too committed to abandon but too weak to live comfortably. "Light" stones can be sacrificed without loss. The technique of sabaki — making shape quickly in hostile territory and giving up pieces when needed — keeps a player flexible in areas where the opponent is strong. The proverb "play away from strength" extends the idea: compete where rivals are thin and resources count, instead of challenging an incumbent in its best-defended position.
These fighting principles produce a stance toward competition that is disciplined rather than aggressive. The first of the Ten Golden Rules of Go, traditionally attributed to the Tang-dynasty court player Wang Jixin, is bu de tan sheng — do not be greedy for victory. When ahead, reduce variance and avoid unnecessary fights. When behind, take calculated risks. An overplay — trying to get more than the position justifies — is the single most common cause of reversals at every level.
What the machines confirmed and overturned
In March 2016, AlphaGo defeated the world champion Lee Sedol four games to one. In Game 2, Move 37 — a fifth-line shoulder hit that commentators first read as a mistake — proved strong and became the symbol of AI finding value where human convention saw none. In Game 4, Lee Sedol's "divine move," a wedge that AlphaGo had judged very unlikely, won the only game a human has taken from the program in an official match.
By October 2017, AlphaGo Zero had learned Go entirely from self-play, with no human games at all. Within days it surpassed the version that beat Lee Sedol. Along the way, it rediscovered many common human patterns from scratch — then discarded some of them. Much of the classical tradition was confirmed; some was revealed as habit rather than principle.
The deepest AI lesson may be how it plays when ahead. AlphaGo and its successors choose moves that maximize the probability of winning, not the score. When leading, they make quiet, "slack-looking" moves that give up points but remove paths to losing. Many AI victories are narrow. This formalizes the half-point philosophy of Lee Chang-ho, the human champion who dominated world Go in the 1990s by steering every game toward the smallest reliable margin.
A 2023 PNAS study of millions of professional moves found that human decision quality improved after superhuman AI appeared, and that players made more novel, creative moves. AI raised the level of human play instead of making it irrelevant. The sharpest conclusion is that AI sharpened the tradition rather than replacing it: it showed that much of it was right in spirit, some of it was wrong in detail, and that a willingness to question inherited "correct play" is itself a strategic advantage.
Five questions for any board
The thinking that Go teaches converges on five questions that apply wherever the stakes are competitive and the outcomes cumulative. Before acting, ask: Where is the board — what is the full territory in play? What is the count — who is currently ahead, and by how much? What is urgent — which vulnerability must be addressed before the biggest opportunity is pursued? Who has the initiative — am I choosing the next battle, or reacting to someone else's? And what does winning require — is this a game won by margin, or by probability?
A player who asks these questions consistently will start where structure helps, keep the initiative, attack for position rather than destruction, stay light in hostile territory, and aim for the win that is most likely rather than the one that is most impressive. The same stance serves a founder allocating resources, a manager setting priorities, a negotiator choosing which issues to press, and anyone facing a competitive decision with imperfect information and real consequences. The board changes. The questions hold.