Part 9: Review of Game 4: Lee Sedol's brilliant move reveals weaknesses AlphaGo (The historic match of deep learning AlphaGo vs. Lee Sedol)

[NL Versie]

Review of Game 4:  Lee Sedol's brilliant move reveals weaknesses AlphaGo

After Lee Sedol went under in the match against AlphaGo with three successive losses, and after his mental succumb under incredibly high psychological pressure yet early in the third game, the main question is what Lee Sedol still has up his sleeve when trying to defeat AlphaGo.

And how unimaginable resilient, persistent, and believing in himself Lee Sedol must be, after the new sledgehammer blow of the third game, to try to collect himself, to show that he is able to deliver, and to prove all humanity is right?

The opening of the fourth game is the same as that of the second game (in which Lee Sedol had white also) up to move 12. Lee Sedol response to AlphaGo's probe move (yosu-miru) in the bottom left corner is aggressive (triangle in Dia. 1), attacking both the black stone on the left and black's probing move. 

Dia. 1:  Game 4, after black 25 (circle, Lee Sedol is white) 

Lee Sedol forces AlphaGo to make a choice but the program simply plays somewhere else (tenuki, circle in Dia. 1). With a shoulder hit black first wants to see how things develop in this area before deciding how to answer white's double offensive (triangle in Dia. 1).

Dia. 2:  Game 4, after black 35 (circle, Lee Sedol is white)

Lee Sedol concentrates on taking territory in the edges and bottom corners while AlphaGo conquers the upper side and builds influence in the center (Dia. 2). Lee Sedol follows the amashi strategy: the amashi player takes territory in exchange for influence, and hopes to withstand the outside power and attacks later.

Dia. 3:  Game 4, after black 47 (circle, Lee Sedol is white)

When Lee Sedol tries to neutralize AlphaGo's strong center influence (triangle in Dia. 3), AlphaGo completely ignores his move and plays another shoulder hit at the right side (circle in Dia. 3). Again, black wants white to answer before deciding about attacking white's center group.

When black cuts off the four white stones in the upper center (triangle in Dia. 4) and white extends (circle in Dia. 4), a complex and ambiguous fight develops. AlphaGo heightens the tension and forces Lee Sedol to act with his center stones to prevent that black cashes in the influence built in the center. 

After white's extension (circle in Dia. 4), black's stones on the middle right comes under very severe pressure. How is black able to involve these stones in the game without weakening it's earlier investments in the center?

Dia. 4:  Game 4, after white 64 (circle, Lee Sedol is white)
AlphaGo's last move (triangle) cuts off the white group in the center.

Since multiple important and large-scale fights make it much harder for AlphaGo to judge the board position correctly, these are the ideal and complex battles to initiate for Lee Sedol.

Where to play next, is a matter of 'looking' forward at least 20-25 moves in this ambiguous fight in the center and right edge. Lee Sedol's fabulous intuition and immense skills in assessing all options and starting points in such explosive positions, are herewith invaluable (Dia. 4). Both AlphaGo and Lee Sedol have to rely on a very crude estimate of how the position may develop gradually.

Black decides to leave his stones at the middle right hand side and further strengthens his moyo to increase the pressure on white to the very maximum. For Lee Sedol it seems virtually impossible to do something  sensible with his center stones. After he cuts to make the position more complex (triangle in Dia. 5), AlphaGo answers confidently and seals off any escape routes towards the center (circle in Dia. 5).

Dia. 5:  Game 4, after black 73 (circle, Lee Sedol is white)

Although black has substantial weaknesses at both moyo edges, the white center stones appear to cooperate rather loosely and creating a living group is there out of the question. Also, to escape is not an option for white with such a majority of black stones awaiting him (center and bottom). Moreover, the aji of black's group in the middle right complicates an eventual white escape (Dia. 5).

In short, it looks like this once more will be a major victory for AlphaGo: white's position seems completely hopeless. But then Lee Sedol puts forward a completely unexpected, wonderful, and brilliant move that suddenly sets the entire board on fire.  All white center stones are activated at once and do cooperate effectively by exploiting the aji in black's center moyo in all conceivable means (Dia. 6).

Dia. 6:  Game 4, after white 78 (circle, Lee Sedol is white)

Lee Sedol's move just dropped out of the sky, both for all live commentators and the many top Go-profs that closely followed the game.  It unites all Lee Sedol's power, skills, perseverance, creativity, intuition, willingness to fight and to seek reprisal for all frustration and powerlessness in the previous three games. When seeing this move during the live commentary in China, top Go-prof Gu Li (9p) described Lee Sedol's move 78 (triangle in Dia. 6) as a "God's move". 

With black 79, AlphaGo fails to play the best response and what happens is a sequence of moves that the program calculated as the strongest answers. In exchange for it's group lost on the right, AlphaGo just gets a poor sente move that the program also could have played immediately (Q13 in Dia. 7).

Dia. 7:  Game 4, after white 94 (circle, Lee Sedol is white)

Commentator Michael Redmond (9p) showed that instead of black 79 much better responses were possible than AlphaGo actually played in the game. In the end, AlphaGo has to give up the stones in the middle right, so that Lee Sedol not only has secured his group and largely destroyed black's center moyo but in addition cashed in a pile of points on the right. 

With move 110 (circle in Dia. 8), white secures his relatively weak stones in the center by attacking the black group on the left (center). In this manner, white creates enough options to connect with the group in the bottom left in case of emergency. 

Direct comparison of Dia. 7 and --more than thirty moves later-- Dia. 8 immediately demonstrates that white has achieved the unimaginable: it is as if white's invading center stones were able to escape through an invisible tunnel (David Ormerod). Lee Sedol efficiently utilized the aji in black's center that he correctly estimated 25 moves earlier.

Dia. 8:  Game 4, after white 110 (circle, Lee Sedol is white)

With this successful center fight, Lee Sedol takes a clear lead. The South Korean live commentators cannot hide their excitement when it becomes clear that Lee Sedol --after three hours into the game-- probably will notch a win. 

There is nothing left to AlphaGo other than to secure it's group on the middle left and to allow that Lee Sedol meanwhile connects his center group --in sente--. Then white further erodes black's center by connecting the three white stones on the upper side. Black's entire moyo has been neutralized (Dia. 9).

AlphaGo tries to catch up with active play but Lee Sedol plays magnificently sharp. An example of this is Lee Sedol's counter move in Dia. 9: white ignores black's threat at the bottom right corner and plays a larger counter threat (circle in Dia. 9; that also prevents black's sente moves in this area). 

Dia. 9:  Game 4, after white 136 (circle, Lee Sedol is white).

Another example is Lee Sedol's ignorance of a big endgame move by AlphaGo (triangle in Dia. 10) which he replies with an even more binding move (circle in Dia. 10).

Dia. 10:  Game 4, after white 146 (circle, Lee Sedol is white)

About twenty moves later, AlphaGo threatens with a pointless dame point to cut off white's meanwhile large center group (triangle in Dia. 11). In response, Lee Sedol plays a big endgame move settling his group at the same time (circle in Dia. 11). If AlphaGo would have played here first, the program would have gained several points --in sente--. Thus, with this move (triangle in Dia. 11) AlphaGo in fact resigns.

Dia. 11:  Game 4, after white 168 (circle, Lee Sedol is white).

Obviously, AlphaGo's logic is not optimal anymore at all since it plays many forcing moves generated by the meager potential that the opponent might make a mistake (Michael Redmond). After playing a handful of doubtful moves and several mistakes that even lost additional points, AlphaGo resigned (Dia. 12).

Dia. 12:  Game 4, end position after white 180 (Lee Sedol is white). 

Dia. 12 shows the final position after Lee Sedol's move 180, nearly five hours into the game, at which AlphaGo resigned after it's notion that the probabilities of winning this game were falling below it's critical threshold of 10% for resignation. At this point, AlphaGo is behind at least 5 points (komi included) and therefore needs to make more than ~20 points in the bottom center area (without Lee Sedol getting any compensation for that) in order to catch up.

Even though Lee Sedol played under ultra-high pressure for more than one hour (from move 90 on), using his last byo-yomi period up to the max each move, he was able to maintain and solidify his advantage built up in the middle game. This is the first time that AlphaGo lost a game against a top professional Go player (in the official match games without handicap). Lee Sedol's great victory is an extraordinary and historic achievement which has led to the deepest respect and admiration from all over the world.

A Korean Go-pro stated: "Lee Sedol just fought the 1000 years history of Baduk and I am proud of him". An Younggil (8p) concluded that the game was "a masterpiece for Lee Sedol and will almost certainly become a famous game in the history of Go". 

Lee Sedol's win against AlphaGo shows --and does prove for the first time-- that deep learning AlphaGo does not play perfect all the time and sometimes has severe problems in judging correctly complex middle game positions. Despite the considerable pressure, Lee Sedol could play his own game and exposed a clear glimpse of AlphaGo's weaknesses.

During the press conference afterwards, the public and press were wild with enthusiasm and chanted: “Lee Sedol!” ”Lee Sedol!” “Lee Sedol!”. Several minutes flashing recorded a radiant, visibly overjoyed and genuinely pleased Lee Sedol: "I have never been congratulated so much just because I won one game. Winning this game is still valuable, the drive behind the win today is the trust and the compassion of you all. This one win is so valuable and I will not trade this for anything in the world". 

After the game Lee Sedol answered a question about his mental condition: "after loosing the first three games and thus the match against AlphaGo, I could not say that there was no psychological shock ... but it was not to the extent that I would have to stop playing the ongoing match because at any moment of the game, I really enjoyed the game. I can tell you that I've not retained any severe damage and I'm very happy to say that I won this single game".

And asked what Lee Sedol was thinking when he was played his brilliant move (move 78, Dia. 6): "I thought I would be able to gain some profit quite easily but that wasn't the case. It became more difficult than I had expected and at that point in time, at time of 78th move, that was the only move that I could see, there were no other moves, no other placement that I could think of. That was the only place I could place my stone, it was the only option for me, so I put it there. So I am quite humble for all the praise that I'm getting for that".

Demis Hassabis said: "AlphaGo made a mistake around move 79 but it realized it's mistake only after the 87th move, when it was too late to recover from. AlphaGo thought it was doing very well at the start of the complex center fight but was pressured by Lee Sedol in some mistakes. It is because of the incredible fighting spirit Lee Sedol showed that a creative genius like him is able to find out and exploit any weakness of AlphaGo. The loss by AlphaGo is very valuable: we will look at all the statistics and are looking forward to learn from these weaknesses and fix these problems".

David Silver, Google DeepMind's team leader, added: "We have designed AlphaGo to learn by itself by repeating self-plays. But such an algorithm must have holes in it that we cannot know about. Lee was able to push AlphaGo to expose such a weakness".

Lee Sedol concluded: "I believe AlphaGo has two weaknesses: when I made an unexpected move, AlphaGo responded as if the program had a bug, indicating that the machine lacked the ability to deal with surprises. Furthermore, AlphaGo finds it more difficult to win the game when it has black as opposed to when it has white. Therefore, I would like to ask Demis Hassabis and David Silver: 'would it be possible for me to take black in the last game instead of drawing lots?'". The answer was: "yes, it's fine". 

This fourth game of the match was decided by Lee Sedol's profound and fabulous move in a complex middle game fight that forced deep learning AlphaGo to a sequence of mistakes. After this, the game was completely turned around and AlphaGo was unable to defend against all weaknesses exposed in it's center position. After three successive losses against the program, this game Lee Sedol clearly outperformed AlphaGo for the first time: a great and historic achievement by Lee Sedol.  

[Deep Learning AlphaGo vs Lee Sedol Overview]

[Part 1: Details of the Match of the 21st century]

[Part 2: AlphaGo under a Magnifying Glass]

[Part 3: Lee Sedol about the Go board in his head]

[Part 4: AI versus Human: Expectations and Predictions] 

[Part 5: Google's biggest press conference ever: opening of the match AI vs. Mankind]

[Part 6: Review of Game 1: Lee Sedol underestimates AlphaGo's incredible fighting power]

[Part 7: Review of Game 2: AlphaGo's new move and devastating aggression]

[Part 8: Review of Game 3: Lee Sedol's opening mistakes due to enormous mental pressure]

[Part 10: Review of Game 5: AlphaGo unfamiliar with common tesuji in ultimate moyo game]

Part 7: Review of Game 2: AlphaGo's new move and devastating aggression (The historic match of deep learning AlphaGo vs. Lee Sedol)

[NL Versie]

Review of Game 2: AlphaGo's new move and devastating aggression

At the press conference after the first game, Lee Sedol estimated himself to have a 50% chance of winning the second game. The main question is whether AlphaGo will play (now with black) as aggressive and offensive as in the first game.

You can imagine that Lee Sedol has adapted severely his playing style in multiple ways after losing the first game. In contrast, in the basis AlphaGo is unable to adapt in any way as training a new version of the program takes about 4 - 6 weeks. So it is not possible (and also not desirable) to apply important changes in AlphaGo's algorithms during this seven-day match (except in case serious bugs would occur).

Lee Sedol concentrates before the start of the second game.
Referee Fan Hui (left) watches. 

Lee Sedol plays the first moves in any case more conservative and calm. According to commentator Michael Redmond (9p): "Lee Sedol perhaps has considered a slightly more careful way of playing considering how well AlphaGo has performed in the first game. However, in this game AlphaGo not only plays even more aggressive but it seems that the program also makes less mistakes".

Dia. 1:  Game 2, after black 37 (circle, AlphaGo is black) 

AlphaGo's 19th move (circle in Dia. 1) is an equally astonishing as powerful move. Lee Sedol appears to be as surprised as everyone: he leaves the match table immediately, takes a (allowed) break while his clock remains ticking. Michael Redmond: "This is a creative move. It is a move which I believe I have never seen before in a top professional game". Virtually all AlphaGo's stones are cooperating together effectively with this creative 'shoulder hit'. 

When Lee Sedol returns at the match table, he takes an unusually long time to answer and after only just forty moves in the game, he is already 20 minutes behind in thinking time compared to AlphaGo. Redmond: "He has clearly difficulties to handle such a strong move that he has never seen before. It is a new and unique move and Lee Sedol has to think about it. This is the reason why people become professionals". 

Dia. 2:  Game 2, after black 41 (circle, AlphaGo is black) 

After this well balanced move AlphaGo evidently is the attacker. Younggil An (8p) mentioned AlphaGo's move (see Dia. 1) "a rare and intriguing shoulder hit but Lee Sedol's answer (see Dia. 2) was excellent and beautiful".  

Lee Sedol prevents AlphaGo to build up a moyo at the upper center and leaves the option open to make some territory himself at the right side. However, AlphaGo keeps the initiative and with black 41 (circle in Dia. 2) the program creates possibilities to play on one of the triangles (Dia. 2). Thereby Alphago can try to attack the white stones while making territory at the lower side. In addition, this move of AlphaGo helps to build the center moyo and prevents a white extension at the bottom side. 

Dia. 3:  Game 2, after black 61 (circle, AlphaGo is black) 

With AlphaGo's move 61 (circle in Dia. 3) it builds influence in the upper left center, provides some extension air to it's group in the center left (preventing white's attack), and threatens to invase white's group at the lefthand side (and to cut off a white stone). It is crystal clear that AlphaGo keeps putting pressure and controls the initiative.

Dia. 4:  Game 2, after black 81 (triangle, AlphaGo is black) 

AlphaGo seems not to worry about Lee Sedol's previous strengthening of his groups to prepare a fight at the upper edge. AlphaGo just continues building up it's own framework and, while doing so, exhibits an incredible flexibility.

In response to white's invasion (square in Dia. 4) black doesn't respond with obvious moves either on top or to the left of the invading stone, but instead plays a very solid move at the outside (triangle in Dia. 4), removing any weaknesses there so that black will have all his hands free to attack the invading stone much more heavily and aggressively later on.

AlphaGo still maintains the initiave and forces too choose between: a) extending to the center to make there a living group and neutralize black's strong influence in the center, and b) to make a small living basis in the upper left corner with the great risk that black will become rock-solid at the outside and with that strength can try to cash in the built center moyo (Dia. 4).

Dia. 5:  Game 2, after black 97 (circle, AlphaGo is black) 

Lee Sedol makes the choice to attempt to neutralize the center but remains with a small heavy group without a base while AlphaGo takes points at both sides, strengthens it territory and tightens the pressure on the white group (Dia. 5).

Dia. 6:  Game 2, after black 101 (circle, AlphaGo is black)

White has no choice other than to persevere consequently but black keeps heavily attacking. And during this offense AlphaGo every time profits. With a next beautifully balanced attacking move (circle in Dia. 6) black prevents white from making eyes easily or can connect to the white group at the right hand side. And meanwhile black substantiates the center (which will be of benefit later on during the fight).

Dia. 7:  Game 2, after black 127 (triangle, AlphaGo is black)

With determination Lee Sedol tries to solidify his center group and to prevent AlphaGo from making a mountain of points in the center but the white group remains shaky and stays a heavy burden to Lee Sedol (Dia. 7). While white takes points at the lower edge (square in Dia. 7), black calmly makes territory in the center (triangle in Dia. 7). 

Redmond: "the control of the flow of the game has been alternated between the players several times. However, while fighting in the center AlphaGo played brilliant moves which forced Lee Sedol into immediate actions and choices". And while Lee Sedol seriously threatened the territory at the upper side, AlphaGo doesn't seem to worry and just cashes in points in the center: "points are points and it is completely irrelevant where on the board these points are taken" Redmond continues.

Dia. 8:  Game 2, after black 139 (triangle, AlphaGo is black)

While attacking AlphaGo takes already a tank of points in the center by cutting off a part of Lee Sedol's group. White just has no choice: otherwise his whole group will die. 

But even after white's big move in the upper right (square in Dia. 8) which threatens black's territory at the upper right edge, black plays calmly in the center. Thereby threatening to cut off another little tasty trail from white (triangle in Dia. 8).

Dia. 9:  Game 2, after black 151 (circle, AlphaGo is black)

Unfortunately, also in this situation Lee Sedol has to take eggs for his money and after another substantial bleeding is the white center group, meanwhile strongly shrunk, still not yet connected. AlphaGo keeps up the pressure and with black 11 the program plays a nice move that achieves many results simultaneously (circle in Dia. 9).

Dia. 10:  Game 2, after black 163 (circle, AlphaGo is black)

Again Lee Sedol has no other choice than to try connecting his stones without AlphaGo playfully kicking in even more points in the mean time.  Yet white looses once more some points. With black's forcing move (circle in Dia. 10) to cut again (triangle in Dia. 10), or even to catch the left over group in it's entirety (with a cut on O11 in Dia. 10) it is trying to empty the ocean with a thimble for white.  

At this point in the game Lee Sedol is behind about five points (komi included) so if black is able to cut yet more stones from white's meanwhile disintegrated center group it is definitively end of story. And this is exactly what ultimately happens: white is cut off once more. 

Dia. 11:  Game 2, end position after black 211 (circle, AlphaGo is black)

With this cut the game is over and Lee Sedol resigns about 50 moves later (see final position in Dia. 11), at that time he is about 10 points behind. AlphaGo did not relinquish control of the initiative at any moment in this game: Lee Sedol was defeated by the program in a rather straightforward and very impressive manner.

While Lee Sedol played some obvious overplays (a move that is too greedy to make more points than the opponent, more than could be considered as reasonable on the basis of your opponent's strength) in the first match, this game it is a completely other story: Lee Sedol didn't play any overplays and showed the very best of his Go play while seriously considering AlphaGo as one of his strongest opponents until now.  

Several new and interesting moves have been conjured effortlessly on the board by a very aggressively playing AlphaGo. Some of which initially were regarded as unreasonable, incorrect, or simply not working. 

In an exceptionally big fight during the second half of the middle game, Lee Sedol was put under enormous pressure (Dia. 6 en 7).With clearly much too little time to calculate accurately all favorable variations and outcomes in this extremely complex, never before occurring Go position. 

Regrettably Lee Sedol didn't succeed in finding a better alternative in this ultra complicated, misleading, unclear and confusing position. On that account he made the decisive mistake that gave AlphaGo a lead of about 5 points. At the deciding moment in the game (Dia. 6) Redmond suggested that Lee Sedol possibly was already somewhat behind. 

In the end, Lee Sedol quits while being defeated more or less with his own weapon: starting ultra complex, all-or-nothing fights and then to try getting decisive advantage during that battle. For many top professionals it remained unclear until the last part of the middle game what would be the outcome of this second game. 

Demis Hassabis mentioned that from the second half of the game on, AlphaGo became more and more confident in victory from the midway point of the game (as is continuously monitored by it's probability theory of winning). Even though the professional commentators could not tell who was ahead at that stage in the game. 

Lee Sedol explained afterwards: "Today, AlphaGo played a nearly perfect game: from the very beginning of the game I did not feel like there was a point that I was leading. While during the first game I was overwhelmed and surprised by AlphaGo's way of play, today I was startled and speechless about the extremely strong moves AlphaGo generated, some of which were unexpectedly well working out. Furthermore, AlphaGo, unlike yesterday, did not make any unreasonable or incorrect moves from the second part of the middle game until the end of the game". 

Asked about any possible weaknesses of AlphaGo, Lee Sedol passed over the question to Demis Hassabis after first stating that "today I lost because I was unable to find and detect any weaknesses in the play of AlphaGo". Hassabis answered the question by stating that one of the goals of the match of AlphaGo vs. Lee Sedol is to find out if AlphaGo has any such weaknesses and to try to learn from them. 

Go-prof Hong Minpyo (9p) confirmed that Lee Sedol and he afterwards have analysed every move of the second game for five hours. After the game, Lee Sedol was very confused about why he had lost this game. This post-game analysis and research  probably helped Lee Sedol to recover somewhat. 

Today, AlphaGo apparently has beaten Lee Sedol with unprecedented accurate force power and incredibly profound computations during the most complex phase of the game. Reaching the point at which many human top professional Go players, including Lee Sedol, were unable to confidently calculate and reliably assess the outcome of the complex variations possible. 

Lee Sedol remarked that in the upcoming third game of the match he will have to focus and concentrate on the opening and first part of the game, to seek and exploit any weaknesses of AlphaGo in order to build a lead early in the game.

So now, after two incredibly fascinating and thrilling games of the match --despite their totally different type in nature both have been won by AlphaGo-- the real strength of the AI program becomes more and more apparent. 

The chances of winning this match by Lee Sedol against such a magnificent and superbly playing deep learning AlphaGo seem to decrease rapidly (down to very improbable) after considering this astonishing, fascinating and breathtaking second game of the match. Today, again history has been written, both by Lee Sedol and the team of AlphaGo in many, irreversible, indelible, and unimaginable ways.

This must feel for Lee Sedol like he was hit by a relentless gravitational wave. Struck, overwhelmed, greatly saddened, almost destroyed. What would be in his mind after this second consecutive loss? At this stage in the match Lee Sedol deserves more than the very, very best and the deepest respect, support and admiration to overcome his great disappointment. 

[Deep Learning AlphaGo vs Lee Sedol Overview]

[Part 1: Details of the Match of the 21st century]

[Part 2: AlphaGo under a Magnifying Glass]

[Part 3: Lee Sedol about the Go board in his head]

[Part 4: AI versus Human: Expectations and Predictions] 

[Part 5: Google's biggest press conference ever: opening of the match AI vs. Mankind]

[Part 6: Review of Game 1: Lee Sedol underestimates AlphaGo's incredible fighting power]

[Part 8: Review of Game 3: Lee Sedol's opening mistakes due to enormous mental pressure]