A study of distance & information
The Order Echo
How does one busy shopping week
become a factory surge?
Every change starts the same chain again from week 1.
Shop, week 5
Where are the outstanding crates?
After this week’s decision. Every outstanding crate is counted in exactly one place; nothing has arrived merely because it was ordered.
All five traces use one scale. Shelf counts show the selected week. Decimal crates are quantities in a simplified model.
Three things to try
Select a stage name to see why it placed its order. The wave is calculated from every shipment and commitment; it is not a recorded animation.
What is this model assuming?
This is an original, deterministic supply-chain illustration inspired by the Beer Distribution Game. Four stages hold stock, fulfil old and new orders, and order replacements. Unfilled commitments wait; physical shipments take the chosen number of weeks per link. Messages to the next stage take one week. The factory’s production source is unlimited.
Each stage smooths the demand it sees, aims for three weeks of stock, and corrects 35% of a gap each week. The till switch changes the forecast input only. Counting the pipeline adds a correction for orders already placed, including crates delayed at a supplier. Full rules appear in each inspected decision. These are chosen teaching parameters, not fitted business data or a recommendation.
The general problem of amplification and delayed feedback is described in John Sterman’s MIT account of the Beer Distribution Game. Research by Croson, Donohue, Katok and Sterman also shows why knowing customer demand need not remove instability. The exact values and comparisons here belong only to this model.
No price changes, capacity limits, perishability, human decisions, random failures or real company data are included. All work stays in this page. Reloading restores the starting experiment.