Life In 19x19

Posted: **Mon Mar 21, 2016 12:06 pm**

What were some of your favorite moves from the AlphaGo vs. Lee Sedol showmatch? I have a pretty good idea on two moves that are probably crowd favorites, but I'm sure other moves played in this incredible showmatch caught your attention. For your convenience, I looked through all 5 games and put up diagrams on the moves that I think caused the most reaction from the spectators online (I was watching these games on KGS, Tygem, Baduk TV's stream, DeepMind's stream, and AGA's stream). If I'm missing a move you want to discuss, feel free to add to discussion and I'll add it to this post as well. I'll think about this over lunch and add my own favorite moves later.

Game 1 (AlphaGo (W) vs. Lee Sedol (B))

Click Here To Show Diagram Code: [go]$$ Position at move 84 $$ --------------------------------------- $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . O . O . X X X . . . . | $$ | . . . . . X . . O X . X . O . O O X . | $$ | . . . O . . O . O X . X . . O , X . . | $$ | . . . . . . . . O X O X . X O . . X . | $$ | . . O . . . O X X O O X . . O . . . . | $$ | . . . . . . O . . O X O . . O X . . . | $$ | . . . . . . . . O O X O . . O X . . . | $$ | . . . . . . . O X . X X O . . X . . . | $$ | . . . , . . X X X X . . O . . , . . . | $$ | . . . . . . . . . . X O . . . . . . . | $$ | . . . . . . . . . O O X O . . . X . . | $$ | . . . X . . . . X . . X O . . . . . . | $$ | . . X O . . . . . . . X O . . . . . . | $$ | . . . . . . . . . . . X O . . . O . . | $$ | . . . O . W . . . , . X O . . , . . . | $$ | . . . . . X . . . . . X O . . X . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ ---------------------------------------[/go]

Click Here To Show Diagram Code: [go]$$ Position at move 102 $$ --------------------------------------- $$ | . . . . . . . . . O . . . . . . . . . | $$ | . . . . . . . . O X O . X X X . . O . | $$ | . . . . . X . . O X . X . O . O O X . | $$ | . . . O . . O . O X . X . X O , X . . | $$ | . . . . . . . . O X O X . X O . . X . | $$ | . . O . . . O X X O O X X . O . . . . | $$ | . . . . . . O . . O X O . . O X . . . | $$ | . . . . . . . . O O X O O . O X . . . | $$ | . . . . . . . O X . X X O . . X . . . | $$ | . . . , . . X X X X . . O . . , W . . | $$ | . . . . . . . . . . X O . . . . . . . | $$ | . . . X . . . . . O O X O . . . X . . | $$ | . . O X . . . . X . . X O . . . . . . | $$ | . . X O X . . . . . . X O . . . . . . | $$ | . . O O . . . . . . . X O . . . O . . | $$ | . . . O X O X . . , . X O . . , . . . | $$ | . . . . . X O . . . . X O . . X . . . | $$ | . . . . . X . . . . X O . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ ---------------------------------------[/go]

Click Here To Show Diagram Code: [go]$$ Position at move 154 $$ --------------------------------------- $$ | . . . . . . . . . O . . . . . . . . . | $$ | . . . . . . . . O X O . X X X . . O . | $$ | . . O . . X . . O X . X . O . O O X . | $$ | . . . O . . O . O X . X . X O O X . . | $$ | . . . . . . . . O X O X . X O . O X . | $$ | . . O . . . O X X O O X X . O X O O . | $$ | . . . . . . O O . O X O . . O X . X . | $$ | . . . . . . . X O O X O O . O X X O . | $$ | . . O . . . . O X . X X O . . X . X . | $$ | . . . W . . X X X X . . O . . X O . . | $$ | . . . . . . . . . . X O . . X O O . . | $$ | . X X X . . . . . O O X O . X X X X . | $$ | . O O X . . . . X . . X O . O . . . . | $$ | O . . O X . . . . . . X O . X O X X . | $$ | . O O O . . . . . . . X O . X X O O O | $$ | . . . O X O X . . , . X O . X O O . . | $$ | . . . . . X O . . . X X O . X X O . . | $$ | . . . X . X . . . O X O O . X . X O . | $$ | . . . . . . . . . . X . . . . X . O . | $$ ---------------------------------------[/go]

Game 2 (AlphaGo (B) vs. Lee Sedol (W))

Click Here To Show Diagram Code: [go]$$ Position at move 15 $$ --------------------------------------- $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . X . . . . . . . . . . | $$ | . . X , . . . . . , . . . . . X . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . O . . . . . , . . . . . , . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . X . . . . . . . . . . . . . O . . | $$ | . . . . . . . . . . . . . . . B . . . | $$ | . . . O . . . . . , . . X . X , O . . | $$ | . . . . . O . . . . . . . X O O . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ ---------------------------------------[/go]

Click Here To Show Diagram Code: [go]$$ Position at move 37 $$ --------------------------------------- $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . X . . . . . . . . . . | $$ | . . X , X . . . . , . . . X . X . . . | $$ | . . . . . . . . . . . . . . . . X . . | $$ | . . . . . . . . . . . . . . . O O . . | $$ | . . O . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . O . . . | $$ | . . . O . . . . . , . . . . B , . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . O . . . . . . . . . . . . . . . . | $$ | . O X . . . . . . . . . . . . . O . . | $$ | . O O X . . . . . . . . . . . X O . . | $$ | . X X O O . . . . X . . X . X , O . . | $$ | . X O X . O . . . . . . . X O O . . . | $$ | . . . X . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ ---------------------------------------[/go]

Click Here To Show Diagram Code: [go]$$ Position at move 81 $$ --------------------------------------- $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . O . X . . . . . . . . . . | $$ | . . X , X . . . . , X . . X . X . X . | $$ | . . . . . . . . . . . . . . . . X O . | $$ | . . . B . . . . . . . . . . . O O . O | $$ | . . O . . . . . . . . . . . . . . O . | $$ | . . . . X . . . . . . . O . . . . . . | $$ | . . . O . X . . . . . . . . O O O . . | $$ | . . . O . . . . . , . . . X X X . . . | $$ | . . . O X . O . . . . . . . . . . . . | $$ | . . O X . X . X . . . . . . . . . . . | $$ | . . O O X X . X . . . . . . X . . . . | $$ | . O X X X O O O O . . . . . . . O . . | $$ | . O O X O X O X . . O . . . . X O . . | $$ | . X X O O . X O . X X . X . X , O . . | $$ | . X O X . O O . . . . . . X O O . . . | $$ | . . . X . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ ---------------------------------------[/go]

Game 3 (AlphaGo (W) vs. Lee Sedol (B))

Click Here To Show Diagram Code: [go]$$ Position at move 15 $$ --------------------------------------- $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . X . . . . . . . . . X . . . . . | $$ | . . X , O . . . X , . . . . . X . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . B . . . . . . . . . . . O . . | $$ | . . . . O . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . X . . . . . , . . . . . , . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . O . . | $$ | . . X . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . O . . . . . , . . . . . O . . . | $$ | . . . . . O . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ ---------------------------------------[/go]

Click Here To Show Diagram Code: [go]$$ Position at move 32 $$ --------------------------------------- $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . X X O . . X . . . . X . . . . . | $$ | . X X , O O . O X , . . . . . X . . . | $$ | . . O O . . . O X . . . . . . . . . . | $$ | . . . X X X . . X . . . . . . . O . . | $$ | . . . . O . O O . . . . . . . . . . . | $$ | . . . . . . . . . X . . . . . . . . . | $$ | . . W . . . . . . . . . . . . . . . . | $$ | . . . X . . . . . , . . . . . , . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . O . . | $$ | . . X . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . O . . . . . , . . . . . O . . . | $$ | . . . . . O . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ ---------------------------------------[/go]

Click Here To Show Diagram Code: [go]$$ Position at move 77 $$ --------------------------------------- $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . X X O . . X . . . . X . . . . . | $$ | . X X , O O . O X , . . . . . X . . . | $$ | . O O O . . . O X . . . . . . . . . . | $$ | . . . X X X . . X . . . . . . . O . . | $$ | . O . X O O O O . . . . . . . . . . . | $$ | . X X X X X . . . X . . . . . . . . . | $$ | . X O O X . O O O O . . . . . . . . . | $$ | X O O X X . X X . , . . . . . , . . . | $$ | . X O O X . X . . . . . . . . . . . . | $$ | . . . . O O . . X . . . . . . . . . . | $$ | . . X O . O . O . . . . . . . B O . . | $$ | . . X X X . . . . O . . . . . . . . . | $$ | . . . . . O . . . . . . . . . . . . . | $$ | . . X O . . . . . , . . . . . O . . . | $$ | . X O . O O . . . . . . . . . . . . . | $$ | . . . O . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ ---------------------------------------[/go]

Click Here To Show Diagram Code: [go]$$ Position at move 115 $$ --------------------------------------- $$ | . X . . . . . . . . . . . . . . . . . | $$ | . O X . O . O . . . . . . . . . . . . | $$ | . O . X X O . O X . . . . X . . . . . | $$ | X X X X O O . O X , . . . . . X X X . | $$ | . O O O O . . O X . . . . . . O X O . | $$ | . . . X X X . . X . . . . . . . O . . | $$ | . O . X O O O O . . . . . . . . . . . | $$ | . X X X X X . . . X . . . . . . . . . | $$ | . X O O X . O O O O . . . . . . O . . | $$ | X O O X X . X X . , . . . . . , . . . | $$ | . X O O X . X . . . . . . . X O X . . | $$ | . . . . O O . . X . . X . . X O O . . | $$ | . . X O . O . O . . . . . . O X O . . | $$ | . . X X X . . . . O . O . . O X X X . | $$ | . . . . . O . . . . . B . . O . . . . | $$ | . . X O . . . . . , . . . . . O O X . | $$ | . X O . O O . . . . . . . . . O X X . | $$ | . . . O . . . . . . . . . . . . O X . | $$ | . . . . . . . . . . . . . . . . . . . | $$ ---------------------------------------[/go]

Game 4 (AlphaGo (B) vs. Lee Sedol (W))

Click Here To Show Diagram Code: [go]$$ Position at move 23 $$ --------------------------------------- $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . X O . . . . . . . . X . . . . . | $$ | . . X , X . . X . , . . . . . X . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . O . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . O . . | $$ | . . O , . . . . . , . . . . . , . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . X . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . O . . . | $$ | . O . O B . . . . , . . X . X , O . . | $$ | . . . . . O . . X . . . . X O O . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ ---------------------------------------[/go]

Click Here To Show Diagram Code: [go]$$ Position at move 46 $$ --------------------------------------- $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . X . . . . . | $$ | . . . X O . . X . . . O O X . . . . . | $$ | . . X , X . . X O , . . X . . X . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . O X . . . . . . . . . . . . | $$ | . . O . . O X . . W . . . . . . . . . | $$ | . . . . O X . . . . . . . . . . . . . | $$ | . . . . O X . . . . . . . . . . O . . | $$ | . O O O X . . . . , . . . . . , . . . | $$ | . . X X . X . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . X . . . . . . . . . . . . . . . . | $$ | . . . . O . . . . . . . . . . O . . . | $$ | . O . O X O . . . , . . X . X , O . . | $$ | . . . . . O . . X . . . . X O O . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ ---------------------------------------[/go]

Click Here To Show Diagram Code: [go]$$ Position at move 47 $$ --------------------------------------- $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . X . . . . . | $$ | . . . X O . . X . . . O O X . . . . . | $$ | . . X , X . . X O , . . X . . X . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . O X . . . . . . . . . . . . | $$ | . . O . . O X . . O . . . . . . . . . | $$ | . . . . O X . . . . . . . . . . . . . | $$ | . . . . O X . . . . . . . . . . O . . | $$ | . O O O X . . . . , . . . . . B . . . | $$ | . . X X . X . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . X . . . . . . . . . . . . . . . . | $$ | . . . . O . . . . . . . . . . O . . . | $$ | . O . O X O . . . , . . X . X , O . . | $$ | . . . . . O . . X . . . . X O O . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ ---------------------------------------[/go]

Click Here To Show Diagram Code: [go]$$ Position at move 78 $$ --------------------------------------- $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . X . . . . . | $$ | . . . X O X . X . . . O O X . . . . . | $$ | . . X , X O . X O , . . X X . X . . . | $$ | . . . . . . O X . . . . X O . . . . . | $$ | . . . . . O X . . . . X O O . . . . . | $$ | . . O . . O X . . O . . X O . . . . . | $$ | . . . . O X O . . . . X X O . . . . . | $$ | . . . . O X . . . X W X O X O O O . . | $$ | . O O O X . . X . , . . O X X X O . . | $$ | . . X X . X . . O . X . O . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . X . . . . . . . . . . . . . . . . | $$ | . . . . O . . . . . . . . . . O . . . | $$ | . O . O X O . . . , . . X . X , O . . | $$ | . . . . . O . . X . . . . X O O . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ ---------------------------------------[/go]

Click Here To Show Diagram Code: [go]$$ Position at move 97 $$ --------------------------------------- $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . X . . . . . | $$ | . . . X O X . X . . . O O X . . . . . | $$ | . . X , X O . X O , . . X X . X . . . | $$ | . . . . . . O X . . . . X O . . . . . | $$ | . . . . . O X . . . . X O O . . . . . | $$ | . . O . . O X . . O . . X O . X . . . | $$ | . . . . O X O . . O X X X O . . . . . | $$ | . . . . O X . . O X . X O X O O O . . | $$ | . O O O X . . X . X X O O X X X O . . | $$ | . . X X . X . . O . X . O O O X X X . | $$ | . . . . . . . . . . . . X O . O O . . | $$ | . . . . . . . . . . . . . . O . . . . | $$ | . . X . . . . . . . . . . . . . . . . | $$ | . . . . O . . . . . . . . . . O . . . | $$ | . O B O X O . . . , . . X . X , O . . | $$ | . . . . . O . . X . . . . X O O . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ ---------------------------------------[/go]

Game 5 (AlphaGo (W) vs. Lee Sedol (B))

Click Here To Show Diagram Code: [go]$$ Position at move 18 $$ --------------------------------------- $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . X . . . . | $$ | . . . O . . . . . , . . . . . , X X . | $$ | . . . . . . . . . . . . . . W X O . . | $$ | . . . . . . . . . . . . . . . O O . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . , . . . . . , . . . . . , . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . X . . | $$ | . . . . . . . . . . . . . . . O . . . | $$ | . . . . . . . . . . . . . . . . O . . | $$ | . . . . . . . . . . . . . . . O X . . | $$ | . . . O . . . . . , . . . . . , X . . | $$ | . . . . . . . . . . . . . X . X . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ ---------------------------------------[/go]

Click Here To Show Diagram Code: [go]$$ Position at move 69 $$ --------------------------------------- $$ | . . . . . . . . . . . . . . X . . . . | $$ | . . . . . . . . . . . . O X . X . . . | $$ | . . O . . O . . . . . . O O X X . . . | $$ | . X . O . . B . . , . . . . O , X X . | $$ | . . . . . . . . . . . . . . O X O . . | $$ | . . X . . . . . . . . . . . . O O . . | $$ | . . . X . . . . . . . . . . . . . . . | $$ | . X O X . . . . . . . . . . . . . . . | $$ | . X X O . . . . . . . . . . . . . . . | $$ | . . O O . . . . . , . . . . . , O . . | $$ | . . . . . . . . . . . . . O O . . . . | $$ | . . . . . . . . . . . O O X O X X X . | $$ | . . . . . . . . . . . O X X X O . O O | $$ | . . . . . . . . . . X X O . . O O X X | $$ | . . . . . . . . . . . . X X X O X O . | $$ | . . . O . . . . . , . . . . O X X . . | $$ | . . . . . . O . . . . . . X . X . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ ---------------------------------------[/go]

Click Here To Show Diagram Code: [go]$$ Position at move 70 $$ --------------------------------------- $$ | . . . . . . . . . . . . . . X . . . . | $$ | . . . . . . . . . . . . O X . X . . . | $$ | . . O . . O . . . . . . O O X X . . . | $$ | . X . O . . X . . , . . . . O , X X . | $$ | . . . . . . . . . . . . . . O X O . . | $$ | . . X . . . W . . . . . . . . O O . . | $$ | . . . X . . . . . . . . . . . . . . . | $$ | . X O X . . . . . . . . . . . . . . . | $$ | . X X O . . . . . . . . . . . . . . . | $$ | . . O O . . . . . , . . . . . , O . . | $$ | . . . . . . . . . . . . . O O . . . . | $$ | . . . . . . . . . . . O O X O X X X . | $$ | . . . . . . . . . . . O X X X O . O O | $$ | . . . . . . . . . . X X O . . O O X X | $$ | . . . . . . . . . . . . X X X O X O . | $$ | . . . O . . . . . , . . . . O X X . . | $$ | . . . . . . O . . . . . . X . X . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ ---------------------------------------[/go]

Click Here To Show Diagram Code: [go]$$ Position at move 101 $$ --------------------------------------- $$ | . . . . . O O O X . X . . . X . . . . | $$ | . . . . . O X O X X . . O X . X . . . | $$ | . . O . . O X X . X . . O O X X . . . | $$ | . X . O . O X . X O . . . . O , X X . | $$ | . . . . . O X . . O . . . . O X O . . | $$ | . . X . . O O X X O . . . . . O O . . | $$ | . . . X . . X O O . . . . . . . . . . | $$ | . X O X . . X . . . . . . . . . . . . | $$ | . X X O . . . O B . . . . . . . . . . | $$ | . . O O . . . . . , . . . . . , O . . | $$ | . . . . . . . . . . . . . O O . . . . | $$ | . . . . . . . . . X . O O X O X X X . | $$ | . . . . . . . . . . . O X X X O . O O | $$ | . . . . . . . . . . X X O . . O O X X | $$ | . . . . . . . . . . . . X X X O X O . | $$ | . . . O . . . . . , . . . . O X X . . | $$ | . . . . . . O . O . . . . X . X . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ ---------------------------------------[/go]

Click Here To Show Diagram Code: [go]$$ Position at move 169 $$ --------------------------------------- $$ | . . . . . O O O X . X . . . X . . . . | $$ | . . . . . O X O X X . . O X . X . . . | $$ | . . O . . O X X . X . . O O X X . . . | $$ | . X . O . O X . X O . . . . O , X X . | $$ | . . . . . O X . . O . . . . O X O . . | $$ | . . X . . O O X X O . . . . . O O . . | $$ | . . . X . O . O O . . . . . . . . . . | $$ | X X O X X O . O . . . . . . . . . . . | $$ | . X X O O X O O X . . X O . . . . . . | $$ | X O O O . . X X O X X . X O . , O . . | $$ | . X . . . . . . O X . X . O O . . O . | $$ | . X O . . . . . O X . O O X O X X X X | $$ | . O . B . . . . O X . O X X X O . O O | $$ | . . O . . O . . O . X X O . . O O X . | $$ | . . O X X O . . . . . . X X X O X O O | $$ | . . . O X X O . . , . . . . O X X . . | $$ | . . O O O X O . O O O O X X . X . . . | $$ | . . . O X . X X X O X X O X . . . . . | $$ | . . . . . . . X . X . . . . . . . . . | $$ ---------------------------------------[/go]

Posted: **Mon Mar 21, 2016 1:39 pm**

Click Here To Show Diagram Code: [go]$$ Position at move 15 $$ --------------------------------------- $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . X . . . . . . . . . . | $$ | . . X , . . . . . , . . . . . X . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . O . . . . . , . . . . . , . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . X . . . . . . . . . . . . . O . . | $$ | . . . . . . . . . . . . . . . B . . . | $$ | . . . O . . . . . , . . X . X , O . . | $$ | . . . . . O . . . . . . . X O O . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ ---------------------------------------[/go]

During lunch, I was thinking about some of the moves I posted, and I think (one of) my favorite moves has to be move 15 in game 2. Everyone criticized this move because of its aji-keshi nature, and I am not saying that it is not aji-keshi. However, way later in the game, it turned out that the exchange made from this move ended up benefiting Black. Was it mere coincidence, or did AlphaGo eerily read THAT far ahead to know that this exchange would turn out to be a useful one? Or did AlphaGo perhaps just play it out to simplify the game, and somehow force Lee Sedol to play in a way to make it a useful exchange?

I also have this feeling that, if this was game 1, and AlphaGo played this move, the commentators would have been MUCH harsher on this move and begin to question AlphaGo's playing strength. However, because it was the second game and after AlphaGo established its dominance from game 1, the commentators were more reserved in pointing out AlphaGo's mistakes (unless it was really obvious like the moves played after move 78 in game 4, or a ko threat getting burned which was a common theme with AlphaGo) and I found that to be quite amusing as well.

There's nothing spectacular in the move itself, but a combination of people's reaction to it and the questions it raises up makes it a move I will certainly remember for years to come. Here's Baduk TV's reaction and thoughts on the move, and here was the kibitz on KGS when move 15 was played live:

Code: Select all

fluidistic [-]: what's that move?
gooooo [?]: wow weird
Shana [?]: lee sedol looked uncomfortable in that cam
danielhast [?]: ...is this joseki?
escudero [3k]: d10 maybe to prepare attack at l4?
covert: that looks... strange
glasszee [3k]: ...
Anatoly [2d?]: wow
LordTushi [9k?]: q5?!
fluidistic [-]: ko threat lost?
Anatoly [2d?]: what?
Rankachan [6k]: q5 is joseki ?
njs [1d]: ?!
Prime [2k]: LMAO 
USSRMan [3d]: yup, inventing new fusekis
Sachan [?]: wow
feeldabern [4d]: q5 is a normal exchange but very early
Ephidel [1k]: what?
Sachan [?]: this is interesting~
njs [1d]: Im expecting a new joseki revolution
feeldabern [4d]: alphago really loves to make these exchanges
cxcv: L4/ K4,etc looks a bit concentrated
Sachan [?]: alphago is so aggressive haha
glasszee [3k]: b begins endgame
XiJinping2 [7k?]: This is the start of the New Fuseki Era, people
Prime [2k]: we should just let this bot come up with new fuseki for the future
Ephidel [1k]: very computer move
Jumadan [?]: tenuki now?
feeldabern [4d]: w can't tenuki this
Tictactoe [-]: q5 is a common follow-up, it's kind of kikashi in this pattern
Garp [11k]: is there yet a god analsis of yesterday match? could leer won ?
feeldabern [4d]: s7 or r7 for w are common
okiol [1k]: ok, so let's see... in the fan hui games there were some criticism that alphago were making exchanges like these too early - let's see if lee can actually abuse that
LordTushi [9k?]: Yeah, q5 is common. It's just very early
XiJinping2 [7k?]: Thanks, Tactictoe
USSRMan [3d]: i agree, very computer-like move, humans dont like making these exchanges because it is not flexible
GoIngo [?]: will we see moyo game by alha today?
Jumadan [?]: why not feeldabern?
GoIngo [?]: alpha
LongLost [7k]: DeepMind explainer is ruining my romanticism of AlphaGo! Silence.
glasszee [3k]: after D10, unlikely
escudero [3k]: isn't this just forcing exchange? it removes some aji though
bert [?]: q6 not good for w now?
feeldabern [4d]: black push and cut is very severe
highhand [5k]: b q5 interesting, attacking a strong group
XiJinping2 [7k?]: Isn't C6 in danger? Or does N4 make C6 okay?
feeldabern [4d]: so w will answer in some way
buhaypa [3k]: Is 7.5 komi usual? Seems like a huge burden
gooooo [?]: I wanted to see Lees reaction to this move
zavion [2k]: romanticism of alphago? u in love with skynet?
covert: lee connects
USSRMan [3d]: c6 is 
kadoban [?]: 7.5 is standard komi for chinese rules
USSRMan [3d]: c6 is light
Tictactoe [-]: extending to j3 is walking head on towards a solid f3 stone, and might invite m3 
Jumadan [?]: is it big enough that early?
aab [?]: Korean BadukTV stream and some brief English translations: https://www.youtube.com/watch?v=bATCEcRbBzE  https://twitter.com/askakorean
wolvie [?]: Yeah, they cut away from his reaction in the stream :/
buhaypa [3k]: Ah ok thanks
okiol [1k]: USSR, the theory behind these "mistimed" and sometimes just bad exchanges based on how the board looked, was that alphago just didn't have understanding over the position enough, and just played common moves

Posted: **Mon Mar 21, 2016 2:40 pm**

Solomon wrote:
$$ Position at move 15
$$ ---------------------------------------
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . X . . . . . . . . . . |
$$ | . . X , . . . . . , . . . . . X . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . X . . . . . . . . . . . . . O . . |
$$ | . . . . . . . . . . . . . . . B . . . |
$$ | . . . O . . . . . , . . X . X , O . . |
$$ | . . . . . O . . . . . . . X O O . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ ---------------------------------------
Click Here To Show Diagram Code
[go]$$ Position at move 15 $$ --------------------------------------- $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . X . . . . . . . . . . | $$ | . . X , . . . . . , . . . . . X . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . O . . . . . , . . . . . , . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . X . . . . . . . . . . . . . O . . | $$ | . . . . . . . . . . . . . . . B . . . | $$ | . . . O . . . . . , . . X . X , O . . | $$ | . . . . . O . . . . . . . X O O . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ ---------------------------------------[/go]
During lunch, I was thinking about some of the moves I posted, and I think (one of) my favorite moves has to be move 15 in game 2. Everyone criticized this move because of its aji-keshi nature, and I am not saying that it is not aji-keshi. However, way later in the game, it turned out that the exchange made from this move ended up benefiting Black. Was it mere coincidence, or did AlphaGo eerily read THAT far ahead to know that this exchange would turn out to be a useful one? Or did AlphaGo perhaps just play it out to simplify the game, and somehow force Lee Sedol to play in a way to make it a useful exchange?

I also have this feeling that, if this was game 1, and AlphaGo played this move, the commentators would have been MUCH harsher on this move and begin to question AlphaGo's playing strength. However, because it was the second game and after AlphaGo established its dominance from game 1, the commentators were more reserved in pointing out AlphaGo's mistakes (unless it was really obvious like the moves played after move 78 in game 4, or a ko threat getting burned which was a common theme with AlphaGo) and I found that to be quite amusing as well.

There's nothing spectacular in the move itself, but a combination of people's reaction to it and the questions it raises up makes it a move I will certainly remember for years to come.

Timing of exchanges like this is always difficult, and it's also a popular subject of debate. But I think many times including here it doesn't make much difference in the grand scheme of winning or losing games. I don't believe for one second that AlphaGo knows that this move is correct in a perfect play sort of way. But I do believe AlphaGo would tend to prefer this forcing move from the center of the board instead of forcing moves on the right side. And I also think we can see AlphaGo in general likes to make forcing exchanges earlier than necessary sometimes.

I thought AlphaGo's previous move, tenuki to form a Chinese formation on the top was much more interesting than the forcing move. It counts as one of my favorite moves from AlphaGo. My other favorite AlphaGo moves are the usual suspects of Game 1 invading the right side, and Game 2 5th line shoulder hit.

Outside those individual moves, I was relieved AlphaGo didn't play many more nonsense forcing moves than it actually did. The ones in Game 4 were not too many and actually served to make the sequence where Lee Sedol had turned the game much more dramatic. The atmosphere was an all-knowing bot playing moves that looked bad but end up working, so during the confusion I felt fear that maybe we're missing something and there was going to be some epic squeeze. But it slowly became more clear that Lee Sedol was winning. And then the wedge in the lower left was just funny.

Game 5 walked a bit closer to my fears, that AlphaGo would get into a good position and play many nonsense moves, making both itself and Lee Sedol look bad. I'm relieved it didn't play more than a handful of those sort of moves.

Posted: **Mon Mar 21, 2016 4:28 pm**

Hmmmm. One of the users online now is Google[Bot].

Big Bot is watching!

Posted: **Mon Mar 21, 2016 5:10 pm**

Nobody's favorite move was AlphaGo's resignation in Game 4?

Posted: **Tue Mar 22, 2016 3:40 am**

Solomon wrote:
$$Wcm16 Position at move 15
$$ ---------------------------------------
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . X . . . . . . . . . . |
$$ | . . X , . . . . . , . . . . . X . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . O . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . 1 . |
$$ | . . X . . . . . . . . . . . . . O . . |
$$ | . . . . . . . . . . . . . . . B . . . |
$$ | . . . O . . . . . , . . X . X , O . . |
$$ | . . . . . O . . . . . . . X O O . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ ---------------------------------------
Click Here To Show Diagram Code
[go]$$Wcm16 Position at move 15 $$ --------------------------------------- $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . X . . . . . . . . . . | $$ | . . X , . . . . . , . . . . . X . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . O . . . . . , . . . . . , . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . 1 . | $$ | . . X . . . . . . . . . . . . . O . . | $$ | . . . . . . . . . . . . . . . B . . . | $$ | . . . O . . . . . , . . X . X , O . . | $$ | . . . . . O . . . . . . . X O O . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ ---------------------------------------[/go]
During lunch, I was thinking about some of the moves I posted, and I think (one of) my favorite moves has to be move 15 in game 2. Everyone criticized this move because of its aji-keshi nature, and I am not saying that it is not aji-keshi. However, way later in the game, it turned out that the exchange made from this move ended up benefiting Black. Was it mere coincidence, or did AlphaGo eerily read THAT far ahead to know that this exchange would turn out to be a useful one? Or did AlphaGo perhaps just play it out to simplify the game, and somehow force Lee Sedol to play in a way to make it a useful exchange?

During the KGS kibitz, maybe a bit later, Sweetrip 9d (who I think is Andy Liu?) made the comment that answering with the kosumi above could have been better, and was more his style of move and (possibly inaccurate paraphrasing) he doesn't understand/agree why everyone else tends to solid connect as the default answer. Andy Liu tends to play quite low and territorial and then invade sabaki style. He also made the point that if white had played this kosumi then black's 5th line shoulder hit on the right side would have been less effective (for example in one variation it would allow white to connect up and defend the right side territory with one fewer move). On the flip side, the solid connection means a future white attack on the black group is more powerful, so maybe black wouldn't have bothered to defend his group at k4 later. I think if Lee and AlphaGo played this game again Lee might not solid connect again (and indeed he varied the one point jump to kosumi to 4th line to discourage a black tenuki in game 5), or would at least spend longer thinking about the other choices to answer the peep. Also is another downside of the kosumi that p2 is more like sente for black (with r3 being the tesuji follow up) so white's m3 invasion aji is lessened? Anyway as Solomon (did you change your name from Araban?) this peep raises many interesting questions and shows it's not easy for a top pro to demonstrate that what would be called a bad aji keshi move at cursory glance is truly bad.

Edit: I suppose I should also watch An Younggil's review of this game, as the accompanying sgf includes some variations about the peep. I also seem to recall him saying Lee should have directly attacked at k4 instead of d10, though presumably he didn't out of fear white would ignore to play moves like d10 and r14 and the group is still not dead (Li Zhe 6p showed this variation half way down https://massgoblog.wordpress.com/2016/0 ... lee-sedol/).

Posted: **Tue Mar 22, 2016 4:13 am**

Btw, illluck (or other people who can read Chinese) the latest translated Li Zhe post mentions Facebook's Yunadong Tian's analysis of game 4 move 78, is that this: http://www.chinaz.com/news/2016/0315/512626.shtml Could you kindly provide a translation/summary? Google's translation is actually not so bad:

Li Shishi 78 L11 fourth hand dug by everyone as the "hand of God", in the policy of the network output in DarkForest ranked 31, and ranked 10 by J11. So I think there might be AlphaGo not count to this step. If the person did not count under the hand machine to chess, Monte Carlo (MCTS) will clear the search tree, and then re-start the search, we should not draw conclusions too quickly. Li Zhe sixth tell me K10 seconds under the hand, it might be time management subsystem procedural flaws at the time of the search tree emptied, thus prematurely search results returned. MCTS at the beginning of the search, because the number of simulations are not enough, each step of winning the variance is very large, so the return is not good enough with a method such as K10 is normal (in this DarkForest inside the top four). This relatively easy to fix.

Another possibility is that, AlphaGo valuation network problem. Because the right to re-valuation of the network is 0.5, regardless of the child quickly go from a situation repeat how many times its weight is 0.5. For a situation, the valuation of the network only get a number, and go down from this situation the child will get a lot more than the number to go after, statistically should be more important, but AlphaGo not think so, both sides claim to be winning after each direct in half the average. So if the results of a valuation of the network situation was not right, it will greatly affect the situation of the winning estimate. We get a lot of attention here because according to article number, leaf nodes have accumulated after a certain number of plate (40) to expand, rather than the first visit had begun to improve DCNN efficiency. DarkForest not used valuation network, after digging correctly L11 L12 and L10 returned both hands should be, according to Li Zhe sixth that these are the right hand should be, which indirectly supports this inference. AlphaGo after 87 hands only realize that they have lagged far behind, it may also be due to the same problems, such as the big black dragon on the right to see survival.

Why does the valuation of the network will be problems? It may be used for self-learning and training network valuation (Self-Play) sample distribution blind spots. In order to increase the production rate of the sample, the sample AlphaGo is self-learning by two purely DCNN each stroke to generate (no search), and DCNN at chess because it is a pure pattern recognition, a big problem is that life and death is not correct, often erupted in a stupid move inside. If black and white two parties have committed, regardless of life and death problems, for example, then the white one lucky win, and that the valuation of the network will think just stupid move white situation is good. Such valuation network will be infected with the same problems, misjudgments in the disk of the complex situation in the killing. If this situation is hard to deal with, AlphaGo next Board may also have the same problem. Here you can see, the computer itself is not to rely on exhaustive chess, Go, after all, is too complicated, prune every step, from the current situation carefully near Shear (with DCNN), the current situation is far from fast cutting (fast go child) until the final outcome so far obtained. Pruning is directly related to the level of chess, DCNN just a bigger picture of a very good means of pruning, it will blind spot by losing move reflected.

About DCNN + MCTS robbery. MCTS is a global first, because the estimated fraction of ko itself and other situations in the program appears to be no essential difference, just a move it. Robbery particularity in DarkForest performance hit may have robbed mentioning happens, DCNN often very high probability (0.8 or more) returns ko capture this hand. Probable cause is that robbery is a point as a separate feature input, so DCNN to learn it and output (ko) strong correlation. So MCTS search would strongly favor this hand. In many cases this is true, but sometimes it is very small and can be disregarded robbery, robbery or across two or more of a need to give up, that the "robbery will lift," the bias will be to search for trouble. Sometimes the computer serial robbery repeatedly mention this reason. AlphaGo may have this problem, or is the reverse problem (for example, ko probability is very small), so that everyone will feel it when playing chess in the open to avoid robbery or miscalculation changes contained in the robbery.

About Horizon effect (Horizon Effect). As the country's AI which have this effect, for example, to search only 10 steps to calculate the others after their end after eating, and then use a simple additive method to estimate the disk find themselves under particular plurality cool, I feel that this branch particularly good. In fact, their progress downward step after others have been eaten, or fall into the trap, so mistakenly thinking face value. But go because each simulation are both in the end, probably before using 30 steps DCNN, after the child is to use the fast walking, although there are gaps in quality child go, but the wrong direction, so to some extent the horizon effect on is weakened. And this AlphaGo mistakes within 20 steps, which should still DCNN range, so the possibility is relatively low horizon effect.

Posted: **Tue Mar 22, 2016 7:02 am**

Uberdude wrote:Btw, illluck (or other people who can read Chinese) the latest translated Li Zhe post mentions Facebook's Yunadong Tian's analysis of game 4 move 78, is that this: http://www.chinaz.com/news/2016/0315/512626.shtml Could you kindly provide a translation/summary?

Li Zhe's posts are also being translated here: https://massgoblog.wordpress.com/

Posted: **Tue Mar 22, 2016 5:43 pm**

Uberdude wrote:Btw, illluck (or other people who can read Chinese) the latest translated Li Zhe post mentions Facebook's Yunadong Tian's analysis of game 4 move 78, is that this: http://www.chinaz.com/news/2016/0315/512626.shtml Could you kindly provide a translation/summary? Google's translation is actually not so bad:

Here is my translation. I omitted last the graph and last couple of lines.

Lee Sedol's move 78 was hailed "The Divine Move". In Darkforest's policy network, this move ranked 31st, the attach at J11 ranked 10th. I think it is possible that AlphaGo did not anticipate this move. If you play a move that is out of the machine's search scope, the MCTS search tree will clear completely and start a brand new search, therefore it should not return too quickly. Li Zhe 6p told me that K10 was played within seconds (of move 78), so it is possible that the time management subsystem has a flaw when a complete restart of the search tree is encountered, and returned the search result prematurely. For MCTS, due to limited number of simulations at the beginning of a search, the win rate at each step has large variance, and returning a suboptimal result such as K10 is quite possible (K10 was ranked in top 4 in Darkforest's search). If this is the bug then it is easy to fix.

Another possibility is the value network of AlphaGo has some issue. The weight of value network is always 0.5, regardless of how many rollouts has been played from a given position. For this given position, value network will only yield one number, but the rollouts from this position will yield many numbers, and might be statistically more significant than the one number given by value network. But AlphaGo does not think that way. It aggregates all playout results to one number, then take simple average with the one number provided by value network to generate the overall win rate. Therefore if the value network supply an incorrect win rate, it will sinificantly affect the overall win rate estimation. Note we know the rollouts will return many numbers because in the Nature article it was stated that leaf nodes will not be expanded until (40) games played to improve the efficiency of DCNN. Darkforest did not use a value network and returned L12 and L10 as possible responses to the wedge at L11, both are acknowledged by Li Zhe 6p as correct responses. This partially supports the claim (that value network is the culprit). It could be for this same reason that AlphaGo did not realize it was behind until move 87, maybe mistaking the black dragon on the right as alive.

How could this (misjudgement) happen to the value network? One possibility is that there were blind spots within the Self-Play sample space that was used for training the value network. To quickly generate large quantity of samples, AlphaGo played millions of games with itself using DCNN only(no MCTS involved). DCNN is based purely on pattern recognition, and a big problem with this is the difficulty of identifying life and death and often continue to play even if the group is dead. If both black and white commit the sin of misjudging life and death, and somehow white comes out as the winner, value network will actually favor a dead white group position. In this way value network might misjudge a complex midgame fight position. If this was the bug, it will not be easy to fix, and it might happen again in future games. After all Go is too complex for computer to do an exhaustive search and pruning is required at every step. Immediate next moves are pruned more carefully (DCNN), moves further down the path are pruned more quickly (rollouts), until we reach the end of game. Methods used for pruning directly affect the strength of a GO AI. DCNN is a very good pruning method with regard to whole board positional judgement, but it also manifested its blind spot with those losing moves.

On DCNN+MCTS ko fights. First of all for MCTS ko and other positions are no different. They are both one move during a long game. The distinction of ko in Darkforest manifests as a high probability (over 80%) to take the ko if it's a legal move. This maybe due to the fact that ko was separately input as a special characteristic and DCNN learned a strong correlation with the move of taking the ko, and thus caused MCTS also strongly favor this move. In majority of situation this is correct, but when the ko is insignificant or when there are two or more ko's, "take the ko whenever possible" will cause problem
during the search. That is why some GO AI will keep fighting a double ko. AlphaGo could have this problem, or opposite of this problem (an unusually small probability for taking a ko), and cause it to avoid ko or make mistakes during ko fight.

On Horizon Effect. In Chess AIs have this effect. For example suppose it only searches the next 10 moves, and found that within 10 moves it can capture the opponents' Queen. A simple valuation of the position will heavily favor this branch. But the truth is the very next move afte 10 moves its own Queen will also be captured or it will fall into a trap and lose the game. This effect will cause the AI to misjudge a position. But because GO AIs always simulate to the end of the game, maybe first 30 moves with DCNN, the rest with MCTS, even though the later moves could be of low quality, but the general direction should still be correct, and this Horizon Effect should be reduced. In this game AlphaGo's losing moves appeared within 20 moves and were still within DCNN scope, so it was unlikely to be caused by Horizon Effect.

Life In 19x19

Favorite move(s) from the AlphaGo vs. Lee Sedol showmatch?

Favorite move(s) from the AlphaGo vs. Lee Sedol showmatch?

Re: Favorite move(s) from the AlphaGo vs. Lee Sedol showmatc

Re: Favorite move(s) from the AlphaGo vs. Lee Sedol showmatc

Re: Favorite move(s) from the AlphaGo vs. Lee Sedol showmatc

Re: Favorite move(s) from the AlphaGo vs. Lee Sedol showmatc

Re: Favorite move(s) from the AlphaGo vs. Lee Sedol showmatc

Re: Favorite move(s) from the AlphaGo vs. Lee Sedol showmatc

Re: Favorite move(s) from the AlphaGo vs. Lee Sedol showmatc

Re: Favorite move(s) from the AlphaGo vs. Lee Sedol showmatc