AlphaZero's stunning victory over chess engine Stockfish raises hopes of Artificial Intelligence's… – Firstpost

AlphaZero, a chess-playing Artificial Intelligence (AI) designed by DeepMind, a Google-owned AI company, became a worldwide sensation when it defeated Stockfish, the world’s strongest chess engine, in a prolonged match. A year since its December 2017 match, the AI has played about a thousand more games, adding another 155 wins while losing only six.

Computers today, with their rating strengths easily crossing Elo 3400, are known to play near perfect chess. To put this into perspective, the highest Elo rating ever achieved by a human was Carlsen’s 2882 in May 2014. Since both opponents in a computer vs computer match play such high-quality chess, the games usually end in draws. So when AlphaZero won the 100-game match against Stockfish with a 28-0 score (with 72 draws), it was bound to catch the fancy of the world.

But more than the difference in score, what startled the bigwigs of the game of chess was the way AlphaZero played. Unlike a traditional chess engine (or a program), AlphaZero played in an active, dynamic style. Its strategies were visible to the human eye – if not at the outset, at least after a few moves into any given game. This was a welcome difference when compared to the unreasonable-looking or at times even ugly but unquestionably winning moves that traditional chess engines found.

Image credit: DeepMind Technologies Limited

Image credit: DeepMind Technologies Limited

In the opinion of the former world chess champion, Garry Kasparov, chess had been “shaken to its roots”.

Besides, there is also this whole fascinating story about how AlphaZero taught itself from scratch. AlphaZero was fed nothing but the rules of the game and within nine hours, it trained itself to convincingly defeat the strongest chess program on the planet.

As per DeepMind’s website, “To learn each game (Chess, Shogi and Go), an untrained neural network plays millions of games against itself via a process of trial and error called reinforcement learning. At first, it plays completely randomly, but over time the system learns from wins, losses and draws to adjust the parameters of the neural network, making it more likely to choose advantageous moves in the future.”

GM Matthew Sadler and WIM Natasha Regan, the authors of the book Game Changer: AlphaZero’s Groundbreaking Chess Strategies and the Promise of AI, say that this all-new silicon monster has helped us discover that there are lots of fresh potential in chess.

“What I found with AlphaZero is that a lot of things that I considered incidental – stuff that I would not really aim for but, if it came up in a game, I would use – became the centrepiece of its strategy.

“One very good example of it is restricting the movement of the enemy king. There are a lot of games where AlphaZero moves its rook’s pawn all the way up to h6 against the black king on g8. That pins the black king to the back rank, restricts its movement and then, somehow, the rest of the game is about opening lines or maybe getting a rook to the back rank. Giving up material doesn’t really matter then because there is a long term advantage to exploit. Those sorts of strategies, those sorts of things are what I found amazingly instructive and they changed the way I think about stuff,” Matthew Sadler told Firstpost.

However, strategy isn’t only where AlphaZero flourishes. When it comes to opening theory, the AI, in its several games against Stockfish, has come up with several different ideas in various openings which while now seem obvious, hadn’t ever been deployed before.

“Sometimes it comes up with ideas that make you think why they have not been played before, and yet it is the case. The Schliemann [Defence] is a very nice example. With black, it came up with this idea of Bb7 at some stage, castling queenside and throwing the ‘h’ pawn forward — something that has never been seen before but actually, it’s a very fine idea. We were also seeing the world championship games — these complicated Sveshnikov positions — with AlphaZero and it was finding all sorts of gorgeous plans,” Sadler pointed out.

Regan added that it is due to this plan-based play of AlphaZero that we could learn a lot more from it than from a traditional calculation based program. “AlphaZero’s plan is a little bit clearer. So it might put its pieces on certain squares and it will do it regardless of what exactly the opponent’s playing and come up with a plan that can be easy to follow,” she said.

When it comes to endgames, chess has already seen engines play provably perfectly. Positions containing up to seven pieces have been already solved using retrograde analysis. This gives rise to some important questions in the light of such technological advancements as AlphaZero: Will such chess programs change the way we play chess? Will chess be solved one day? How would the solving of chess impact the game as a whole?

Regan thinks that we are not quite at the stage where the game of chess can definitively be solved but there is a good chance that ideas generated by programs like AlphaZero change the way Grandmasters think. An aggressive approach might galvanize top players into trying to push for wins.

“I guess in the near future, the top players will try out these attacking techniques and try some new styles. I know that in top chess, sometimes, there are quite a lot of draws. So it might bring back some confidence in the possibility of attacking.

“I could also imagine these young players who get really, really strong very early get even more terrifying if they study all AlphaZero games and know how to conduct such attacks. I could imagine them being really fearsome.”

Sadler also shared his co-author’s view that the royal game is still quite far off from being mathematically solved and what AlphaZero has essentially done is opening up lots of fresh possibilities.

“What AlphaZero has shown is that we really thought that Stockfish was basically the strongest we can get and then suddenly, there seems to be a whole space of positions where there is still lots and lots of scope for all sorts of things. And who knows, maybe, what AlphaZero has found isn’t the last word and that Stockfish is going to come back and maybe something new will come along. I think all we have discovered is that there is lots of fresh potential in chess. I think it’s going to be a long time before it gets solved.”

He also pointed out that even if chess was mathematically solved, it would not be an easy task for the players to remember all the possibilities.

“You couldn’t remember everything, of course. But you could know the direction you need to take in every game, in every variation. With a lot of work, you could get to a sufficiently promising position. I mean, they do it already in stuff like the Berlin Defence where they go 40 moves deep or so.

“I think it will have an impact on the professional game if it was absolutely solved but I don’t think it will have any impact whatsoever on club players.”

“But then, even if it is solved, you’ve always got Chess960 and there is so much stuff you can do,” Regan said.

Aditya Pai is an editor at ChessBase India

Firstpost is now on WhatsApp. For the latest analysis, commentary and news updates, sign up for our WhatsApp services. Just go to and hit the Subscribe button.

Leave a Reply

Your email address will not be published.