I'd like to understand chess game endings from two perspectives: 1. **By rule/category**: How many distinct types of game endings exist in chess? (e.g., checkmate, stalemate, resignation, draw by agreement, threefold repetition, fifty-move rule, insufficient material, timeout, etc.) 2. **By final positions**: Mathematically, how many unique final board positions are possible? Or if that's not calculable, what estimates or bounds exist? For the second part, I'm interested in whether this has been computed or studied, and what makes it complex to determine.
How Can a Chess Game End?
Part 1: Distinct Types of Game Endings
There are 12 distinct ways a chess game can end, which fall into three broad categories:
♚ Decisive Results (One player wins)
| # | Ending Type | Description |
|---|---|---|
| 1 | Checkmate | The king is in check with no legal escape. The fundamental way to win. |
| 2 | Resignation | A player concedes defeat voluntarily. |
| 3 | Timeout (flag fall) | A player runs out of time and the opponent has sufficient material to theoretically mate. |
| 4 | Forfeit | Loss due to rule violation (illegal move, phone ringing in tournament, failure to show up, etc.) |
🤝 Draws
| # | Ending Type | Description |
|---|---|---|
| 5 | Stalemate | The player to move has no legal moves but is not in check. |
| 6 | Draw by agreement | Both players mutually agree to a draw. |
| 7 | Threefold repetition | The same position (same pieces, same player to move, same castling rights, same en passant possibility) occurs three times. A player must claim it. |
| 8 | Fivefold repetition | Same position occurs five times — draw is automatic (no claim needed). Added in 2014 FIDE rules. |
| 9 | Fifty-move rule | 50 consecutive moves by each side with no pawn move and no capture. Must be claimed. |
| 10 | Seventy-five-move rule | 75 moves without a pawn move or capture — automatic draw. Added in 2014 FIDE rules. |
| 11 | Insufficient material | Neither side can possibly checkmate (e.g., K vs K, K+B vs K, K+N vs K, K+B vs K+B with same-colored bishops). |
| 12 | Timeout vs. insufficient material | A player's clock runs out, but the opponent lacks sufficient material to checkmate — result is a draw. |
Note: Some online platforms add additional categories like "abandonment" (disconnection), but these are platform-specific implementations, not formal chess rules. Under FIDE laws, they'd typically fall under forfeit or arbiter decisions.
Part 2: How Many Unique Final Board Positions Are Possible?
This is a fascinating and largely unsolved problem. Let's break it down in layers:
The Shannon Number (Upper Bound Context)
Claude Shannon (1950) estimated:
- Legal positions in chess: ~10⁴⁴ to 10⁴⁷
- Possible games (sequences of moves): ~10¹²⁰
These are all positions, not just final positions, but they set the scale.
Refined Estimates of Legal Positions
Victor Allis (1994) estimated ~10⁴⁶.25 legal positions. More recent work by John Tromp and others has refined the upper bound of legal positions to approximately:
4.82 × 10⁴⁴ (Tromp, 2023 — computed upper bound)
Final Positions Specifically
Final positions are a strict subset of all legal positions. Here's what we can say about each category:
| Ending Type | Estimated Unique Positions | Computability |
|---|---|---|
| Checkmate positions | Vast but finite; partially enumerated for ≤7 pieces via endgame tablebases (Lomonosov tablebases cover all 7-piece positions). Estimated in the billions for all piece combinations. | Partially computed |
| Stalemate positions | Much smaller subset than checkmates, but still enormous for full 32-piece games | Not fully computed |
| Insufficient material | Relatively tiny — only a handful of material categories, but multiple board arrangements of each | Enumerable in principle |
| Threefold/fivefold repetition | Any legal position could theoretically be the repeated one | Essentially the full set of legal positions |
| 50/75-move rule | Any legal position could be the 50th/75th move | Essentially the full set of legal positions |
| Resignation/agreement/timeout | Any legal non-terminal position | Nearly all legal positions |
Why This Is So Hard
- Legality constraints: Not every arrangement of pieces on a board is reachable from the starting position. Determining reachability is extraordinarily complex.
- Context-dependent endings: Threefold repetition and the 50-move rule depend on game history, not just the board state. Two identical positions may or may not be "final" depending on what came before.
- Combinatorial explosion: Even with just checkmates, you must consider every possible material combination × every legal arrangement on 64 squares.
- Promotion: Pawns can promote to any piece, creating positions with, say, 9 queens on one side — rare but legal, vastly expanding the space.
Bottom Line Estimate
The number of unique final board positions (ignoring game history) is likely on the order of:
10⁴³ – 10⁴⁴ (very rough estimate)
...because almost any legal position can be a final position through resignation, draw agreement, or timeout. The set of final positions is essentially co-extensive with the set of legal positions, minus some edge cases.
The number of forced terminal positions (checkmate + stalemate + insufficient material) is significantly smaller — perhaps on the order of 10³⁷ – 10⁴² — but this has not been rigorously bounded.
Key References
- Shannon, C. (1950). Programming a Computer for Playing Chess
- Allis, V. (1994). Searching for Solutions in Games and Artificial Intelligence
- Tromp, J. (ongoing). Calculations on legal chess positions — tromp.github.io/chess/chess.html
- Lomonosov / Syzygy Tablebases: Complete solutions for all positions with ≤7 pieces
Comments