Are chess engines good or bad?
The revolution in technology has touched many areas and improved our lives in many ways. Chess is no exception. Technology has allowed us to develop chess engines that have changed the way we understand the game, the way we study it and many other things. But they have also made cheating easier, made us lazy when analysing games and many other things, so are chess engines good or bad?
Table of content
A chess engine is a computer program whose function is to observe a chess position, analyse the different ways the game can continue and choose the most beneficial move for a given player. Currently chess engines are much stronger than humans with even 3400 Elo (any chess player has ever crossed 2900 Elo) and they keep improving every year with the technological improvements.
When we show a position to a chess engine, it uses an algorithm to evaluate the position and find the best move.
First it analyses all the possible moves in the position. This is done using a tree diagram known as minimax. Minimax is a method of decision making used in game theory. It looks for the move that loses the least advantage if we assume that the opponent will play perfectly.
Many chess engines start by making a list of all the legal moves in the position. Then they check them one by one. It makes one of the possible moves and analyses the new position again, then undoes the move and starts again with another move (depending on how deeply we want to analyse the position it will make more or less moves after the first one). This process is repeated for all legal moves.
The most difficult part is how the engine evaluates a given position. This is done using a neural network. This is a very complex structure that is useful in many computer programs. The way it works can be summarised as follows: there are several layers of neurons connected to each other. We send some information to the first layer, it processes the information and sends it to the next layer. This is how we get the "right answer" in the last layer.
Some of the world's best engines use neural networks to become stronger by playing against themselves. For example, AlphaZero was taught the rules and then plays millions of games against itself, learning from its mistakes in each game and getting stronger until it reaches a very high rating.
A lot of companies have decided to use a chess engine. Here I show you some of the most important and used engines in the world:
- Stockfish: is a free and open source chess engine. It is probably the most famous engine and is used by thousands of players all over the world. It has won all main events of the Top Chess Engine Championship (TCEC) and the Chess.com Computer Chess Championship (CCC) since 2020. It started by using only a classical handicraft function to evaluate board positions, but adopted a hybrid evaluation with the introduction of the efficiently updatable neural network in August 2020.
- AlphaZero: it is considered to be the strongest chess engine in the world. It was developed by the DeepMind team at the end of 2017. It can also play Go. As I have already explained, it learnt to play from scratch by playing against itself. In just a few hours, it was able to defeat very highly rated chess players. In a match against Stockfish, AlphaZero won 28 games and draz 75 without losing a single game.
- Leela Chess Zero: is a free, open source, deep neural network based chess engine and volunteer computing project. One of the purposes of Leela Chess Zero was to verify the methods in the AlphaZero paper as applied to the game of chess. It also learned to play with a neural network. Although some people thought it was not as strong as other engines, it is capable of playing at a level comparable to Stockfish.
Now that we know how chess engines work and which ones are among the most important, we will try to understand whether they have influenced us in a good or bad way.
Since the appearance of the first chess engines we have been able to understand some positions in a very different way, which has helped us to play more interesting games. They have also helped in opening theory, suggesting new moves that are very difficult for a human to make if you don't have some help to see the continuation. They are very useful when players analyse their games, because they show some ideas that people can't go into in depth with little time.
Despite all the advantages that the engines give us, they have also brought some bad things. For example, it is now much easier to cheat because all you need is a mobile phone that tells you the right move in seconds. Also, some players just do what the engines do instead of thinking for themselves and trying to make a creative move. Some amateur players don't understand why engines make certain moves and repeat what they say instead of debating with other chess players with different points of view.
Chess engines have helped chess a lot, especially at the highest level, as we have improved our understanding of position and can now play more complex games. However, if we use this new technology badly, the game of chess can be damaged, because people's viewpoints are different and this makes the games more interesting with their creativity, mistakes and brilliant moves.
As Kasparov said:
Artificial intelligence is not here to compete with us, but to challenge us to be better.
Garry Kasparov
