Transparent method

How the Tokyo family base model works

Public facts stay public. Japan Answers’ original value is the schema, normalization, weights, load penalties, confidence labels and editorial synthesis that turn scattered network information into a reversible decision.

Version: Tokyo family base model 1.0 · 2026-08-24

The model does not predict travel time or crowding. It ranks six broad bases on a consistent 1–5 editorial scale and then applies the priorities entered in the selector.

Layer 1 · sourced facts

What comes from official public inputs

Airport access pages establish the operator corridors serving Tokyo. Rail network maps establish which line families meet each area. Official destination pages provide neighborhood orientation. These facts are credited; Japan Answers does not own them.

Two rail map sites returned bot-protection responses to our automated check on the review date; their public URLs remain the cited operator references. Airport and listed GO TOKYO pages returned successfully. A blocked automated fetch is not evidence that a route or service is unavailable.

Layer 2 · normalized decisions

The six 1–5 dimensions

DimensionWhat 5 meansWhat it does not mean
Haneda / Narita arrival fitThe area is structurally well placed for that airport in the published networkA direct train at your arrival time
Station easeLower navigation burden relative to this six-area setStep-free access from every platform to every hotel
East / west reachGood fit for repeated days on that side of TokyoA measured minute-by-minute journey
Intercity launchUseful access to long-distance departure patternsEvery intercity train leaves there
Evening easeUseful food/services or a manageable local return patternA safety, nightlife or restaurant-quality rating

Layer 3 · formula

Weights and penalties stay readable

The selector computes a weighted average of the relevant 1–5 scores, converts it to 100 points, then subtracts a physical-load penalty. Airport fit receives 1.4× weight. A chosen trip direction receives 2.5×; balanced trips weight east, west and intercity equally at 1.2×. Station ease receives 2.4× for low tolerance, 1.4× for medium and 0.7× for high.

Each large bag—and one additional stroller/mobility-load unit—creates a larger penalty at areas with lower station-ease scores. Groups above four receive a small additional complexity penalty. Scores are rounded only after the complete formula. The source function lives in src/lib/tokyo-base-score.mjs and is covered by mutation and invalid-input tests.

Judgment and uncertainty

Confidence is about the record, not your hotel

High confidence means the broad network role and decision tradeoff are well established across multiple official inputs. Medium confidence means service pattern or station geometry can change the practical result more easily. No confidence label covers a hotel’s room policy, elevator route or current price.

AreaConfidenceLast checkedKnown unknown that can reverse the result
Uenohigh2026-08-24The station still has multiple operators and exits. A hotel described as “Ueno” may sit on the opposite side from the route you expect to use.
Asakusamedium2026-08-24“Asakusa Station” refers to separate operator facilities. A through-service shown on a network does not mean every departure is direct.
Tokyo Station / Marunouchihigh2026-08-24“Near Tokyo Station” is not a walking-time guarantee. Yaesu, Marunouchi and underground approaches create very different luggage routes.
Shinagawahigh2026-08-24The main station is broad and busy. Nearby properties can still involve crossings, elevated walkways or a long platform-to-lobby path.
Shinjukuhigh2026-08-24Station side is decisive. A property on the wrong side can add confusing internal walks or force a street-level detour.
Ikebukuromedium2026-08-24East and west exits lead to different daily lives. Airport convenience is not the main reason to choose this base.

Version policy

When the model changes

A new version is required when an area, dimension, formula weight or penalty changes. Source-date refreshes without scoring changes can keep the same major version. We preserve unknowns rather than filling gaps with guessed fares, schedules or walking times.