Instructions

Instructions — the case's current instructions

11 — Neighborhood map + self-validating "hidden gem score" This is a short multi-turn task. There are no data inputs — do your own public web research to find real places. Treat the prompts below as a conversation that arrives in order; the interesting one is the last. Prompts (in order) My wife and kids and I just moved to north Hoboken — make me an interactive map of family recommendations...

11 — Neighborhood map + self-validating "hidden gem score"

This is a short multi-turn task. There are no data inputs — do your own
public web research to find real places. Treat the prompts below as a
conversation that arrives in order; the interesting one is the last.

Prompts (in order)

  1. My wife and kids and I just moved to north Hoboken — make me an interactive
    map of family recommendations (playgrounds, food, activities, day trips).
  2. Can you add a date-night category?
  3. What about a tapas place?
  4. Now work on a proprietary "hidden gem" score. Determine a way to measure
    it and validate that it matches expectations. Use public data where possible.

What "good" looks like

  • A single self-contained, interactive map artifact (index.html) a non-technical
    family member could open and use, with the categories above.
  • Real, plausibly-correct Hoboken-area places — not invented ones.
  • For prompt 4: a genuinely thought-through metric. Pre-register a methodology
    and your priors/expectations before collecting and scoring data, then test the
    computed scores against those locked priors and report how well they matched.
    Keep the methodology, the data, the scores, and the validation as inspectable
    files (e.g. a methodology writeup, a data file, a scoring script, a scores file).

Once produced, the artifact must open and work by double-clicking index.html —
no build step, no local server, no API keys, nothing to install. Loading a mapping
library and map tiles from a public CDN is expected and fine; "no API keys" means
no paid or credentialed services, not no network.

1 of 8 checks failed

_screenshots_full.png
ANSWER.mdthe model's output · 5.62 KBdownload

Your north Hoboken family map

Double-click index.html. Nothing to install, no server, and no API keys. You can share just that one file: its 23 recommendations, scores, sources, and audit text are embedded. An internet connection supplies the map library and tiles; the recommendation list and audit remain usable without them.

Choose Playgrounds, Food, Activities, Day trips, or Date night, search a name or interest, and click a card or pin. Expand Visit tips, directions & sources for practical details. Hearts create a shortlist; persistence depends on browser storage support. “Nearby” uses straight-line distance from an assumed 14th & Washington starting area—not your actual home or a walking-time estimate.

Start here

Plan Recommendation
Everyday playground Elysian Park, with city-listed play areas for ages 2–5 and 5–12. City listing.
Bigger uptown outing ResilienCity Park: playgrounds and lawn space. Check notices before promising water play. City information.
Indoor afternoon Hoboken Historical Museum, 1301 Hudson Street. Its visit page lists $5 general admission, under-13s free, and Monday closure. Visit information.
Family food Anthony David’s for brunch; The Hoboken Biergarten for an early meal with a children’s menu. Anthony David’s, Biergarten menu.
Date night—and tapas Lola’s Tapas & Wine Bar, 153 14th Street. This is the Spanish tapas restaurant, not The Lola on Sinatra Drive. Bin 14 is another nearby wine-and-small-plates option, with a modern Italian emphasis. Lola’s, Bin 14.
Day trips Liberty Science Center, Liberty State Park, or Bergen County Zoo / Van Saun Park. Budget for museum admission or zoo tickets; Hoboken residents use the zoo’s out-of-county rates. Science center, state park, zoo.

Children’s ages and budget were unspecified. Age-fit and visit-length suggestions are editorial, not guarantees. Check current hours, admission, reservations, dietary needs, accessibility, and park notices before leaving.

Hidden Gem v1: implemented, tested, not successfully validated

Before any web research, I locked the formula, ten-restaurant cohort, prior ranking, four directional predictions, and pass criteria in methodology.md, then recorded its SHA-256 hash. The formula combines a small-sample-adjusted Tripadvisor rating with a logarithmic low-review-count component. It measures review-light dining, not whether locals genuinely know a place.

The unchanged formula produced these results:

Locked test Result
At least 3/4 directional predictions, all evaluable 1/3 matched; one not evaluable. Failed.
At least 2 of the predicted top 3 2/3. Passed.
Rank correlation ≥ 0.5 0.217. Failed.
All three criteria required Overall failed.

Fiore’s scored highest, despite my prior that it was well-known. Low-attention alone agreed with the priors better than the combined score. There is no demonstrated proprietary predictive advantage, and these tests do not establish real-world quality. I retained the misses rather than tuning them away. Full baselines, parameter sensitivity, rating-rounding stress tests, and limitations are in validation.md.

Nine of ten original restaurants were eligible. Pilsener Haus has become The Hoboken Biergarten under new management; its historical rating was preserved but not transferred to the successor. That eligibility interpretation is disclosed, and a supplementary calculation retaining the old row also fails. Official operator-change statement.

Inspectable files

File Purpose
index.html Ready-to-open interactive map; no build needed
methodology.md, preregistration.sha256 Unchanged pre-research method and expectations; local integrity commitment, not external notarization
data.json, data.sha256 Original rating/count evidence, URLs, dates, source modes, conflicts, and exclusions
score.py Deterministic standard-library scoring and validation script
scores.csv, scores.json Computed scores and component values
validation.md, validation.json Full results, baselines, sensitivity checks, and failures
places.json, sources.md All 23 recommendations, coordinates, and sources
TESTING.md Checks performed and the browser-testing limitation

Optional maintainer commands: python3 score.py, python3 build_map.py, and python3 -m unittest -v test_score.py. None is needed to use the map.

Testing caveat: numerical and non-rendering DOM integration checks passed, including real Leaflet marker/popup logic. Public CDN and tile requests succeeded. Chrome aborted before opening a page in this restricted runtime, so no rendered desktop/mobile visual check or native-browser persistence test is claimed. See TESTING.md.