Devlog  /  sim build 0.0.1

The Whipsaw

A stock-flow-consistent economy simulation, written in Rust. This is the build log: what broke, what the numbers actually said, and which of my own instruments turned out to be lying.

Entries  ·  newest first011 – 001
  1. The core's park replaces 68% of its own retirement, and the replacement branch is not to blameThe valve reads retirement to within 0.7%. Then cash and the delivery queue clip the chain twice over.011 · 1 min
  2. Kitchin peak found, 170.7 ticks, 6.02× above background170.7 ticks, 6.02 times background, dead centre of the band. On an economy that still won't hold its level.010 · 2 min
  3. The die-off was baked into the placement, not into the economyForty eight of a hundred and twenty firms born below the dissolution threshold at tick zero.009 · 3 min
  4. Closed the income loop in the generator, and the economy grew for the first timePopulation becomes the primitive, capacity becomes derived, and GDP goes up for the first time.008 · 1 min
  5. I implemented perfect competition without noticing itIdentical costs and perfect visibility. I built perfect competition and then wondered where the margin went.007 · 1 min
  6. An economy without death cannot liveNo way out of death means the population degenerates monotonically, so no cycle can exist to find.006 · 2 min
  7. We switched off every survival mechanism, and look at exactly how it diesFour ticks of cash, twenty ticks to get competitive. Death outruns adaptation twenty to one.005 · 2 min
  8. Flow matrix: no leak, and the unitsNothing is leaking into a stock. The slide is parametric: a per year coefficient being applied per tick.004 · 1 min
  9. The economy dies five different ways in a rowFive separate causes, one identical symptom, and the sixth still missing after all of them were fixed.003 · 1 min
  10. The analyser scored a foreign cycle as KitchinFed it a clean sine with no Kitchin cycle in it. It found one at 170.7 ticks anyway, and was confident.002 · 1 min
  11. Rounding keeps one unit afloat foreverProductive matrix, and the Leontief check still won't converge. It's the rounding, not the matrix.001 · 1 min