March 25, 2023
Garry Kasparov vs DeepBlue (1996-1997). First ever computer to beat the world champion.
Garry Kasparov vs Deep Blue (1996-1997)

Chess engines are computer programs that are designed to play the game of chess. These engines use advanced algorithms and techniques to analyze the game board and make intelligent decisions about the best moves to make. The use of chess engines has become increasingly popular in recent years, as they provide players with a strong and reliable opponent to test their skills against. Whether you are a beginner or a seasoned master, a chess engine can help you to improve your game, giving you valuable insights and suggestions as you play.

So, How do chess engines work?

Image of Stockfish's logo, an open source chess engine
Stockfish, an opensource chess engine

The first step in creating a chess engine is to define the rules of the game. This involves encoding the various moves that each piece can make, as well as checks and captures. These rules are the foundation upon which the engine will make its decisions, so it is important that they are accurate and up-to-date.

The engine’s programming must evaluate the strength of different positions and moves after the rules of the game have been established. This is important as this is the heart of chess. An algorithm that can always figure out the best move in any position can be said to be unbeatable. Algorithms use something called an openings database to guide it along the first couple moves of the game. These databases contain an optimal set of first couple of moves, which ensures perfect decision making for the engine.

After the opening phase of the game, engines use heuristics to determine the best moves. Heuristics simply refers to general guidelines that the engine uses to make decisions. Common examples of this include counting the number of pieces on the board, considering the mobility of these pieces, and possible future threats and opportunities. These methods provide the engine with a quick and efficient way to evaluate a position without too many calculations.

In order to make decisions quickly, chess engines often use specialized hardware, such as GPUs or custom ASICs, to perform the calculations required for their evaluations. These calculations are done in parallel, allowing the engine to consider many possible moves and positions at the same time. This parallel processing helps to speed up the evaluation process, allowing the engine to make decisions in real-time as the game is being played.

Should we care?

Algorithms such as Stockfish and AlphaZero are about 1,000 elo rating above the best human players on Earth, but should these players care about this? Engines like the ones mentioned can help increase the quality of play by humans as we can use them to prepare for tournaments and try to find faults in the opponents common plays. Even so, there is one outstanding threat. Chess at a high level is often complained about being really dry. A very limited number of openings are played as only a few of them maintain the solidity and balance that is achieved by near perfect play. As these engines get better, chess may become even more dull. They might refute common plays and attacks that are now in the trend, promoting solid play and perhaps only one opening, that would make every game seem like the same and even lose audience.

Leave a Reply

Your email address will not be published. Required fields are marked *