What this is
TRACE is not a game in the traditional sense.
It is a deterministic system designed to explore how agents — human or artificial — behave when the space of possible actions shrinks irreversibly over time.
Every action leaves a permanent trace. Nothing is forgotten. What becomes impossible never becomes available again.
A different class of problem
Most games — from Chess to Go — rely on reversible state spaces: mistakes can be compensated, positions revisited, strategies reset.
TRACE belongs to a different class.
It implements a Deterministic Game with Irreversible Global Memory: a system where the legality of future actions depends on the entire history of previous actions, not just the current state.
No randomness. No hidden information. No chance. Only accumulation.
What is being tested
TRACE does not reward speed, efficiency, or local optimization.
Instead, it probes a different ability:
The capacity to read approaching limits
before they become explicit constraints.
Failure rarely comes from a single wrong move. It emerges from patterns — repeated decisions that silently consume future possibilities.
The system does not warn you when a choice is bad. It only makes that choice unavailable later.
Why this matters
From a computational perspective, TRACE maps onto hard classes of problems related to sequence-constrained traversal on dynamic graphs, with strong analogies to NP-hard and PSPACE-complete settings.
From a cognitive perspective, it exposes a blind spot shared by humans and many AI systems:
Competence at optimizing locally
versus competence at preserving future freedom.
Standard reinforcement learning, tree search, and heuristic planning struggle here — not because the rules are complex, but because the reward structure is inverted: what looks optimal now often destroys the long-term possibility space.
No solution, only traversal
TRACE has no final solution. No optimal policy. No winning strategy in the classical sense.
Success is measured in how long meaningful choices can be sustained before the system closes in on itself.
What remains, at the end of each run, is a structure: a trace of decisions, constraints, and limits encountered.
Some are short. Some are surprisingly resilient. All are irreversible.
Who this is for
Researchers in computation, complexity, and AI.
Designers of puzzles and formal systems.
Players interested in mastery beyond reflex or optimization.
Anyone curious about how limits emerge — not suddenly, but silently.
TRACE does not explain its rules. It lets you feel them.