Validation

Riskiest assumption

That one van can serve six or more independent groups in a day without the
service degrading into unacceptable waiting times. If the van can only really
look after two groups, the price has to double and the product becomes the guided
tour it was meant to undercut.

Nothing in the international research checks this assumption. An earlier
draft used Vine Hopper's "~75-minute headway" as an external sanity check; that has
been withdrawn — those are departure times from one depot, not a measurement of how
often a vehicle passes a point on a loop. The 45-minute circulation assumption
remains completely unvalidated.

Cheapest test that proves it

Do not build anything. Ride the loop and drive the loop on the same Saturday.
One person rides it at a realistic tasting pace with stop times logged; one
person drives the circuit continuously and logs how long a response to any point
on the loop actually takes. Simulate six groups by generating six random pickup
requests across the day. Cost: a weekend and a hire car.

• Pass: median response under 20 minutes with six simulated groups.
• Weak pass: under 35 minutes — viable but the price has to carry a second van sooner.
• Fail: over 35 minutes, or the loop can't be driven continuously.

Second test — demand

A landing page with real dates and a real price, and a $50 deposit. Ten paid
deposits before a single bike is bought. Deposits, not survey answers.