DEEPFORM-STYLE
FIELD INSTRUMENT
LIVE SIMULATION
47.61° N / 122.33° W
FIELD / 01  —  SEATTLE ADVANCED LIFE SUPPORT

Moving ambulances
is a rounding error.

Two identical fleets. One year of real dispatches. The left fleet uses Seattle's actual medic posts. The right one is yours to configure. Both simulations run in this page, call by call, from the same data — nothing here is a recording.

SCROLL TO OBSERVE
01PREMISE

A relocation system shows a gain.
We want to know what it is buying.

Dynamic ambulance relocation moves idle units between posts to chase demand. It is one of the most studied ideas in emergency-services operations research and a live product category. The question this instrument answers is not whether it produces a number, but whether that number survives being charged the cost of the moves — and whether simply parking the same units somewhere better would have produced more.

02THE INSTRUMENT  ·  LIVE DISCRETE-EVENT SIMULATION
SEATTLE MEDIC ONE  ·  2025 HOLDOUT  ·  DISPATCHES
A · ORIGINAL — SEATTLE'S ACTUAL POSTS8 units · never moves
B · CHALLENGER
CLOCK PROGRESS0.0% SELF-TEST PENDING
under 4 min 4–8 8–12 over 12 ◆ staffed post → relocatingtract fill = live mean response there

Both panes consume the identical call stream at the identical clock. Service durations are the same hash-seeded draws the Python study used, exported per incident, so the two engines agree exactly. On load the browser re-runs the full year for every policy and checks its answers against the Python reference; the badge above reports the largest disagreement found.

03RESULT

Charge the moves their real travel time
and standing still wins.

FULL-YEAR RESULT, EVERY POLICY, AT YOUR CURRENT SETTINGS
MEAN RESPONSE VS COST OF A MOVE

Recomputed live. Drag the relocation-cost slider and this redraws; the flat line is the policy that never moves.

04METHOD & LIMITS

What is real

What is modelled