Time-Gel · live simulator · water

Does freezing water slow its own clock?

Supercooled water at −10 °C freezes at a moment nobody chooses. Room A freezes. Room B1 gets the same heat from a heater. Room B2 gets a heater tuned to produce the same entropy as the freezing. Each room has a clock. Under the rate reading, a clock slows with entropy produced; under the stock reading, it slows with the ice that stays.

A · Freezing cellSupercooled water crystallises, holds, melts

0
lag this cycle

B1 · Heater, same heatSame latent heat dumped at bath temperature

0
lag this cycle

B2 · Entropy-matched heaterSame entropy as A, no crystal

0
lag this cycle
Cycle phase (compressed, not to scale)
  1. Supercooled liquid
  2. Recalescence
  3. Growth
  4. Hold (ice)
  5. Melt

Real numbers for one cycle (default timings from the protocol)

What you see is magnified. The clock hands drift by the model's real lag multiplied by the factor shown. The real slowdown during growth is about one part in a hundred million, far below what a hand could show. The full parameter set (timings, efficiency, clock noise, number of cycles) is in the founder's ice clock simulator. Protocol: Information Looking for Its Owner v1.0.

Model output, not experimental evidence. No ice clock has been run yet.