How did AlphaZero destroy Stockfish in 100 games match?
Written by: Emadeldin Teama
The era of Artificial Intelligence in chess!
On December 5, 2017, the DeepMind team of developers and software engineers released a preprint introducing AlphaZero, which within 24 hours of training achieved a superhuman level of play in chess by defeating world-champion programs Stockfish, Elmo, and the three-day version of AlphaGo Zero.
In each case it used google supercomputer hardware to run Alphazero. AlphaZero was trained solely via “self-play”. After 9 hours of training itself, the algorithm defeated Stockfish 8 in a time-controlled 100-game tournament (28 wins, 0 losses, and 72 draws).
In this article we are going to discuss some of the games released between Stockfish and Alphazero and what does it mean for the chess game?
What is AlphaZero and how did it learn chess?
Alpha Zero is an AI (Artificial intelligence) algorithm made specifically for games with infinity numbers of possibilities. In the match between Stockfish 8 and Alphazero we noticed a certain pattern in the move choice by the AI. It always choose aggressive outrageous moves that made the world of chess remember the good old 18th Century chess geniuses such as Paul Morphy , Jose Capablanca and Alexander Alekhine .
The thing about this AI is that it didn’t have any access to any opening book or endgame tables. It was taught only how the pieces move and the very basic rules of a chess game.The AI using neural network technologies was given 8 hours to train itself. After 8 hours of self-training, the AI declared itself ready. Let’s check a couple of its games against Stockfish8 the strongest chess engine in 2018
Game (1)
GM Jon Ludvig Hammer characterized AlphaZero’s play as “insane attacking chess” with profound positional understanding. This is exactly the theme of the 18th Century Grandmasters as we will notice in the next games . It feels like a brilliant human being playing on board and not a computer!.
Demis Hassabis, a chess player himself, called AlphaZero’s play style “alien”: It sometimes wins by offering counterintuitive sacrifices, like offering up a queen and bishop to exploit a positional advantage. “It’s like chess from another dimension!
This following game shows an outrageous maneuvering power in a position where normal engine would evaluate as equal. This game actually broke the internet when it was released. The way Alphazero trapped the black queen is unbelievable.
Zugzwang in the middle game!
And here the full game:
Critics to AlphaZero match vs Stockfish 8:
There has been few critics to the match as it was being unfair. The truth is Alphazero was running on google super computers which gave it a serious advantage over any other chess engine running on any hardware.
Grandmaster Hikaru Nakamura stated “I don’t necessarily put a lot of credibility in the results simply because my understanding is that AlphaZero is basically using the Google supercomputer and Stockfish doesn’t run on that hardware; Stockfish was basically running on what would be my laptop. If you want to have a match that’s comparable you have to have Stockfish running on a supercomputer as well.
Another critic was made related to Alphazero by the Top US correspondence chess player Wolff Morrow.He claimed that AlphaZero would probably not make the semifinals of a fair competition such as TCEC where all engines play on equal hardware. Morrow further stated that although he might not be able to beat AlphaZero if AlphaZero played drawish openings such as the Petroff Defence, AlphaZero would not be able to beat him in a correspondence chess game either!!
Positive Feedback from the chess community :
Human grandmasters were generally impressed with AlphaZero’s games against Stockfish. Former world champion Garry Kasparov said it was a pleasure to watch AlphaZero play, especially since its style was open and dynamic like his own. Every Grandmaster loving the fact that AI is not playing passively like other chess engines and it goes for extremely aggressive and unusual positional sacrifices that may change some of our understanding to the chess basic positional rules and how we think about them!
In the computer chess community, Komodo developer Mark Lefler called it a “pretty amazing achievement”, but also pointed out that the data was old, since Stockfish had gained a lot of strength since January 2018 (when Stockfish 8 was released). Fellow developer Larry Kaufman said AlphaZero would probably lose a match against the latest version of Stockfish, Stockfish 10, under Top Chess Engine Championship (TCEC) conditions. Kaufman argued that the only advantage of neural network–based engines was that they used a GPU, so if there was no regard for power consumption (e.g. in an equal-hardware contest where both engines had access to the same CPU and GPU) then anything the GPU achieved was “free”. Based on this, he stated that the strongest engine was likely to be a hybrid with neural networks and standard alpha–beta search.
Conclusion:
AI has taken over many aspects of our lives.it is interesting to see how it can affect the chess community and revolutionize the game of chess. it was also very noticeable how the AI without any prior knowledge to chess opening was playing the normal modern opening we play today.
We believe that AI will revive many chess openings that were once considered not good or out of fashion and we are hoping that google would release more and more games of this outstanding AI!.We would like to end with a 10 game study on Lichess between Alpha zero and Stockfish 8
Here is the link on Lichess
This is also a priceless collection of all the games played by AlphaZero on Chessgames Website