Aug 15, 2016
“Alpha Flow” has quickly gained popularity in the weiqi community since the five-game match between AlphaGo and Lee Sedol in March. Weiqi, also known as “Go” in English, is a two-person board game originating in East Asia.
Lee Sedol is one of the world’s best Go players. In March, he lost to the computer program AlphaGo in a widely publicized series of games. This was the first loss of a human of Lee’s ranking to a computer program.
“Flow” is a common term in Go, similar to a line in chess. It refers to a strategy of setting up the pieces on the board to try and offer a player the advantage. Some of the more famous layouts include the China Flow, created by Chinese professional Chen Zude, the Korea Flow, created in the 1990s by Cho Hun-hyun and Lee Chang-ho, and the Cosmic Flow, created by Japanese grandmaster Masaki Takemiya.
Alpha Flow is term used to describe several new layouts that occurred during the matches between AlphaGo and Lee Sedol, referring especially to the variation that surfaced in the fifth game.
Many professional Go players have since embraced Alpha Flow, acknowledging the profound impact that machines have brought to the game. Many people in the public assumed that professional Go players would laud the day when computers could finally beat humans, but on the contrary, it was very exciting. The machine brought the game new variations and new possibilities.
Let’s first debunk a common misconception. Many people outside of the Go community think that Go players are exceptionally intelligent. They consider AlphaGo’s victory over Lee Sedol proof of artificial intelligence’s dominance over the human brain.
But the truth is, although there are several geniuses among us, most Go players are normal people. We’re just very diligent. Japanese Go master Fujisawa Hideyuki once said: “There are hundreds of secrets in Go, and I only know seven.”
AlphaGo is an extremely powerful machine. The memory of the AlphaGo and Lee Sedol match is still vivid in my mind. In the second game, AlphaGo shocked everyone by making moves unfathomable to the professional Go community. It seemed like the computer had made amateur mistakes.
But the program somehow managed to turn those compromising moves into a powerful flow over the course of the game. By the end of the fifth match, AlphaGo had managed to revolutionize the game. It exposed numerous flaws in existing flows and changed the way players approach Go.
Interestingly, AlphaGo never once played a perfect game. It calculated the highest probability of winning, and played lines that guaranteed it a victory, without any risks or life in its moves. AlphaGo is very good at winning games, but a simple victory has never been the real essence of Go.
Go was first created over 5,500 years ago in ancient China. The board consists of 19 lines, with 361 crossing points, and it is closely related to cosmology. The center of the board is known as taiji in Chinese, which translates to “supreme ultimate” and is a philosophical principal that refers to the creation of the universe. The black and white playing pieces represent yin and yang, a Chinese philosophy about duality that promotes natural balance in the world.
The grandmaster Wu Qingyuan once said that although Go is a competitive game, winning should never be viewed as the end goal. The ultimate purpose of the game should be to achieve neutrality, like the natural harmony that occurs when yin and yang are balanced.
The perfect game is therefore when this neutrality is reflected on the Go board — a match in which the full potential of every single piece is activated. In this way, AlphaGo hasn’t yet achieved perfection.
Nonetheless, AlphaGo has without a doubt helped Go professionals further develop the game. The most immediate proof would be Lee Sedol’s nine consecutive victories immediately after his match with AlphaGo.
Go players are discovering untapped potential. With only two variations, Alpha Flow is already a hit. If the computer continues developing more game patterns, our understanding of the game will be exponentially increased.
I’ve been studying Go since I was 6 years old. I’ve poured my heart and soul into it, but I know that I might never be in the same league as some of the world’s top players.
At 9 I began to study under Ruan Yunsheng, the best player in central China’s Hubei province. I was in the national youth team by 13. I had a solid beginning, but I have settled in the 40s in world rankings. Even though I was born in 1990, I am no longer considered a young player — the new generation born after 1995 dominates the game. One of them, Ke Jie, is currently the world’s highest ranked player.
At training, I can feel the strength and potential of the young players born after 2000. I know I am in decline, and sometimes feel pressured. This is why watching AlphaGo transform the game is particularly comforting to me. It’s helping give me faith that nothing is set in stone — as long as a player has the drive and a little creativity, age shouldn’t be an issue.
Go has never just been about competition for me. It has taught me to remain collected, to think through things, to keep my head above water, and remain confident in the face of defeat. It’s more than just a game of winning or losing, and understanding that is an important part of becoming a professional. I don’t believe AlphaGo can comprehend that.
(Header image: Bambustone/VCG)