Skip to content
The Chat Archive
llama-4-maverick-17b-128e-instruct· Meta LMArena Human Preference 140K

A common mentioned value for possible chess games is 10^120 . But what about reasonable…

  1. User

    A common mentioned value for possible chess games is 10^120 . But what about reasonable (competitive) games? Consider the average branching factor and the average length, or any other important factor to make estimates. Could you come up with a reasonable estimate?

  2. Assistant

    Estimating the number of reasonable or competitive chess games involves understanding the complexity of the game tree and making some simplifying assumptions. The often-cited figure of 10^46 for the number of "sensible" or "reasonable" chess positions gives a hint, but we're aiming to estimate the number of games, not positions.

    Average Branching Factor

    The branching factor in chess varies throughout the game. In the opening, it's relatively low because there are fewer possible moves. As the game progresses and more pieces are developed, the branching factor increases. In the endgame, it typically decreases again as pieces are exchanged and there are fewer possible moves.

    • Opening: The average branching factor is around 20.
    • Middlegame: It can be significantly higher, often cited around 30-40.
    • Endgame: It decreases, let's say to around 10-20 on average.

    For simplicity, a commonly used average branching factor is around 30-35 for the entire game.

    Average Length of a Game

    The length of a competitive chess game can vary significantly. However, most games tend to last between 40 to 60 moves, with some going much longer.

    • Average game length: Let's consider an average of 50 moves per game as a rough estimate.

    Estimating the Number of Competitive Games

    To estimate the number of competitive games, we can use the average branching factor and the average game length.

    1. Simple Model: Assuming an average branching factor of 35 and an average game length of 50 moves (or plies, which is actually 100 if we count each move as a ply, i.e., white's move followed by black's move), the number of possible games can be estimated as (35^{100}).

    2. Calculation: (35^{100} \approx 2.7 \times 10^{154}).

    This is a very large number, significantly larger than the often-cited 10^46 "reasonable" positions, because it counts sequences, not positions.

    Adjusting for Reasonableness

    Not all these games are "reasonable" or competitive. Many involve blunders or suboptimal play. To adjust for this, we could consider a lower effective branching factor that reflects "reasonable" choices at each move.

    • If we reduce the average branching factor to, say, 10-15 to reflect more "reasonable" or competitive play (acknowledging that a strong player has a more limited set of considered moves at each turn), the number significantly decreases.

    Let's use 12 as a hypothetical "reasonable" branching factor.

    • Adjusted Calculation: (12^{100} \approx 8.6 \times 10^{107}).

    Conclusion

    Using a "reasonable" branching factor of 12 and an average game length of 50 moves (100 plies), we can estimate the number of competitive or reasonable chess games to be on the order of (10^{108}), specifically around (8.6 \times 10^{107}).

    This is still an enormous number, far beyond what can be exhaustively analyzed with current or foreseeable computational power. It gives a sense of the vast complexity and richness of chess, even when narrowing it down to "reasonable" or competitive play.

Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.