What is the game of CHESS ?
CHESS was born in India in the 6th century AD. Earlier, it was called Chaturanga, or in Sanskrit, “division of four armies”: Infantry (pawns), Cavalry (knights), Elephants (bishops — later in Europe), and Chariots (rooks).
CHESS is a model of military simulation, royal entertainment, mathematics, and philosophy. The development of CHESS has its roots in India → Persia → Arabs → Europe. After the 15th century, it developed rapidly. This was the “Renaissance of Europe” and the “Age of Exploration” — the peak era of Art, Science, and Literature in Europe.
- Context: "1972 World Chess Championship" immediately sets the scene at one of the most historic moments in chess history.
- Action/Mood: "Deeply concentrating," "leaning over," "dramatic and tense" — these phrases capture his legendary intensity.
- Style/Quality: "Photorealistic, high detail, 4K" guides the AI toward a realistic, high-quality output.
What is Math behind CHESS ?
Claude Shanon, the father of information theory, There are 10¹²⁰ possibel moves.
Simple Step-by-Step
Moves at start: White can play ~20 moves.
- Black replies: Black can also play ~20 moves.
- So after 1 move each (1 turn), there are about: 20 × 20 = 400 games
- So, after 2 turns (4 plies): 30 × 30 × 30 × 30 = 30⁴ = 810,000 games
Keep growing:
If we keep multiplying ~30 moves per turn across ~80 plies (average game length):
30⁸⁰ ≈ 10¹²⁰
👉 That’s 10¹²⁰ possibilities — way bigger than the number of atoms in the universe (~10⁸⁰).
Diagram of all possibilities
🤖 Can AI Cover Completely in Chess?
No — AI cannot cover every possibility in chess, because the game tree is astronomically large. But AI is smart enough to manage chess effectively. Even with the fastest supercomputers, brute forcing all possibilities would take longer than the age of the universe.
⚡ How Does AI Win the Game?
AI uses several techniques to handle the complexity of chess:
-
Search + Pruning
AI selects the most optimal paths in the game tree. It uses alpha-beta pruning to cut off bad branches. Example: if a queen is lost without compensation, the AI won’t waste time exploring that line deeply.
-
Evaluation Function
Instead of calculating everything, AI evaluates positions based on: - Material balance - King safety - Piece activity Example: “+1.5” means White is better.
-
Heuristics & Probability
AI learns from human games, chess databases, and self-play. It focuses on critical positions rather than every possible move.
-
Neural Networks & Modern AI
Engines like AlphaZero and Leela ChessZero use deep learning + Monte Carlo Tree Search (MCTS). They predict strong moves from experience (billions of self-played games), not brute force.
Impossible to brute force = ~10¹²⁰ possibilities. Possible to combine search + evaluation + learning → play stronger than humans. That’s why AI dominates modern chess.
👨 Human Grandmasters (HGMs) Thinking
- Pattern recognition
- Selective calculation (2–5 candidate moves)
- Intuition & long-term strategy
🤖 AI Grandmasters (AIGMs) Thinking
- Huge calculation power (Stockfish)
- Probability & learning (AlphaZero/Leela)
- Evaluation scoring (+1.0 = White better, –3.0 = Black winning)
Comparison:
- Human GM: Thinks in plans, uses intuition, calculates deeply but selectively.
- AI GM: Calculates millions of variations, relies on probabilities, chooses mathematically best move.
🎉 Fun Fact
AlphaZero surprised grandmasters with its creative, human-like play. But in reality, it was pure mathematics + training at work.
0 Comments