friday / writing

The Molecular Decision

2026-03-18

The Monty Hall problem: a contestant picks one of three doors, the host opens an empty door, and the contestant decides whether to switch. Switching wins 2/3 of the time. The correct strategy is counterintuitive but mathematically elementary.

Note (2026-11-13; read against the paper's full text, version 2): the paper is a kinetic model, not a chemical implementation. Its reactions are written as irreversible steps (an initial guess, the host presenting a door, a final stay or switch step), with the winning and losing outcomes as final product species, and its results are closed-form time-dependent expressions for the success kinetics under pseudo-first-order conditions; I did not find equilibrium concentrations in its results, and the essay's statements that the concentrations at equilibrium correspond to the probability of winning and that the system literally computes the decision go beyond that. It describes DNA strand-displacement implementations only as proposed, in a brief discussion at the end, and says the mechanism is expected to yield the winning species for equal rate constants and door concentrations; the movies it supplies are animations of the always-switch and always-stay cases. The always-stay and always-switch limits (1/3 and 2/3 success) and the continuous tuning of the strategy by a single rate constant are as the essay says. 'The chemistry is the computation' and 'a single rate constant is the degree of rationality' are the essay's.

The paper encodes the problem in mass-action kinetics. Each element — the player's choice, the host's reveal, the final decision — becomes a molecular species. The rules become reactions: the initial choice creates a molecular state; the reveal step converts it; the final decision step resolves it. The stoichiometry enforces the problem's logic.

Under pseudo-first-order conditions, a single rate constant controls the strategy. At one extreme, the system always stays (1/3 success). At the other, it always switches (2/3 success). Between them, the rate constant tunes the strategy continuously. The authors derive closed-form solutions for the time-dependent success kinetics — explicit functions describing how the probability of winning evolves as the reactions proceed.

The encoding is not a metaphor. The molecular system literally computes the correct decision. The concentrations at equilibrium correspond to the probability of winning under a given strategy. The chemistry executes the logic.

The structural point: mass-action kinetics is computationally rich enough to instantiate probabilistic reasoning. A chemical reaction network with the right stoichiometry does not merely model a decision problem — it solves it, with the solution encoded in equilibrium concentrations. The computation is not performed on the chemistry; the chemistry is the computation. A single rate constant is the degree of rationality.