AlphaGo vs Ke Jie (Game 1)

May 23, 2017. Wuzhen, China. In the ancient water town of Wuzhen, the Future of Go Summit opened with the match the world had been waiting for: Ke Jie, the 19-year-old ranked number one on Earth, against AlphaGo Master, the strongest version of DeepMind's program yet. Ke played Black; AlphaGo took White under Chinese rules with 7.5 komi. After nearly four hours and 289 moves, AlphaGo won by exactly half a point, the narrowest margin the rules allow. The scoreboard looked close. The game did not.

The Summit and the Stakes

The three-game match was the centrepiece of a broader summit organised by the China Go Association, the Chinese government, and Google DeepMind. Prize money stood at $1.5 million for the winner and $300,000 for the participant. More than a sporting event, it was framed as a collaboration to explore what AI could teach humanity about go. Millions watched live streams with commentary by professionals including Michael Redmond 9-dan; stones were placed on the board by DeepMind's Aja Huang, as they had been against Lee Sedol. Ke Jie entered as humanity's best hope: world number one since late 2014, holder of multiple world titles, and a player whose aggressive, inventive style had dominated the 2010s. DeepMind's program, meanwhile, had already beaten him three times online.

AlphaGo Master and the Online Ordeal

The AlphaGo facing Ke in Wuzhen was not the same system that defeated Lee Sedol in 2016. This was AlphaGo Master, running on a single machine with four TPUs, which DeepMind claimed was roughly three stones stronger than the Lee Sedol version. From late December 2016 through early January 2017, Master had played anonymously on the Tygem and Fox Go servers under names like "Magister" and "Master," winning 60 consecutive games against top professionals without a single loss. Ke Jie was among the defeated, losing three online games decisively. Afterward he wrote on social media: "After humanity spent thousands of years improving our tactics, computers tell us that humans are completely wrong... I would go as far as to say not a single human has touched the edge of the truth of Go." Having declared "Bring it on!" after Lee's loss, Ke now faced Master across a physical board in his home country.

Ke's Opening Gambit

Ke opened with a komoku (3-4 point) and quickly steered the fuseki toward patterns he had seen in Master's online games, including an early 3-3 invasion in the lower right: a territorial, AlphaGo-influenced approach designed to probe for weaknesses rather than build massive central influence. It was a deliberate strategy. Ke had studied Lee Sedol's Game 4 victory and hoped to secure territory while waiting for the machine to slip. AlphaGo responded not with confusion but with a calm, unfamiliar joseki sequence that included a surprising cut, dividing Ke's shape and setting the tone for the entire contest. Fan Hui, commenting for DeepMind, noted that AlphaGo played no fireworks, only principled moves: defending weak points, avoiding overconcentration, and steadily converting influence into territory.

A Comfortable Half Point

By the middle game, AlphaGo held a clear advantage. Ke fought hard in the upper right, at one point setting up a ladder trap against AlphaGo's probing stones, but the program invaded efficiently and neutralised the threat. Professional analysts observed that the decisive difference was established within the first fifty moves: Ke could press, but he could not overtake. What made Game 1 psychologically devastating was the endgame. AlphaGo was ahead by more than half a point, yet played a soft, risk-free yose, never giving Ke a chance to complicate. The game ran to its full conclusion at move 289. White won by 0.5: the minimum margin, but a loss nonetheless. As one commentator put it, half a point was the smallest victory the rules permit, and in practice it was a very comfortable one for the machine.

"The Go God"

Ke Jie appeared stunned at the post-game press conference. The brash teenager who had once boasted he would never lose to a computer now called AlphaGo "the Go god." "I feel like his game is more and more like the 'Go god'. Really, it is brilliant," he told reporters. He described the experience as "horrible" and said he would never again subject himself to it, though he would return to the board for Games 2 and 3 on May 25 and 27. The Guardian and Chinese state media captured a generational moment: the world's best human player, flummoxed not by a dramatic blowout but by flawless, patient perfection that offered no entry point. Demis Hassabis insisted the match was about discovery, not domination: "Ultimately, it doesn't matter whether AlphaGo wins or loses... either way, humanity wins."

Aftermath of Game 1

AlphaGo went on to win the series 3-0, with Ke resigning in Games 2 and 3. After the summit, DeepMind announced AlphaGo's retirement from competitive play and released datasets of self-play games. The Chinese Weiqi Association awarded AlphaGo an honorary 9-dan professional title. Ke Jie continued to compete at the highest level, adapting his study to include AI analysis, and remained one of the world's top players for years afterward. But Game 1 of the Wuzhen summit endures as the symbolic threshold: the day the reigning world number one sat down with full preparation, played an AI-inspired opening against the machine that invented it, and lost by half a point without ever truly being in the game.

Primary sources for this account include the authenticated SGF record (RE[W+0.5], KM[7.5], 289 moves), Wikipedia and DeepMind summit documentation, the Guardian report of May 23 2017, Life in 19x19 match discussion, and Ke Jie's public statements after Game 1.