No photo clue here - deduce today's answer from the grid. Each guess you make in your 6 tries is scored attribute by attribute:
- Country
- IUCN category
- Established
- Area band
✓ exact≈ close🔼🔽 higher / lower✕ no match- no data
| Guess | Country | IUCN category | Established | Area band |
|---|---|---|---|---|
How to play Parkdle
Parkdle gives you six guesses to name one of 808 well-known protected areas that Wikidata records as national parks. Four columns do the work, and one of them is a classification most visitors have never had explained to them.
You get 6 guesses. Each one is scored against 4 columns - Country, IUCN category, Established and Area band - and the scoring is the whole game: you are narrowing a field rather than naming it cold.
IUCN category is exact or nothing: the cell turns green when your guess holds the same value as the answer, and stays blank when it does not. Country can hold several values at once, so it can also come back as a close match: yellow means your guess and the answer share at least one value without being the same set. Area band is a band, and it is scored on how far apart the bands sit: the same band is green, one band either way is yellow, and further away is blank. Area band also carries an arrow towards the answer's band. Established is a number, and an arrow points from your guess towards the answer, so a wrong number still tells you which way to go.
A column the source leaves blank is marked as having no data rather than scored. It cannot come back green for anyone, on any guess, which means an unrecorded value is never a lucky hit - and the gap itself is a fact about the record, so the grid shows it instead of hiding it.
How a round goes
Guess a park and the row returns Country, IUCN category, Established and Area band. Established and Area band carry arrows: the grid tells you whether the answer's decade or size band lies above or below your guess, which turns the round into two bracket searches running at once.
Country is scored as a set, because a handful of parks straddle a border. Share one country with the answer and the cell comes back as a partial rather than a hit. Established is a decade rather than a year, and Area band is a number from 1 to 5: 1 is under 100 km2, then 100 to 499, 500 to 1,999, 2,000 to 9,999, and 5 is 10,000 or more. Where the frozen record carries no area or no founding date there is no arrow to give, because there is no number on the record.
What the IUCN categories are actually for
IUCN category is the column worth learning. It is not a ranking of how wild or how well run a park is; it records the primary management objective the protected area was set up to serve. Category Ia is a strict nature reserve and Ib a wilderness area. Category II covers ecosystem protection together with recreation, which is roughly what most people picture when they hear national park.
The rest of the scale is less familiar. Category III protects one specific natural feature, IV conserves habitat or species through active management, V protects a landscape or seascape, and VI pairs conservation with the sustainable use of natural resources. The descriptions the game uses follow the IUCN protected-area management category guidelines.
In this pool the column rarely splits the field, because category II accounts for most of the entries. That is why the categories are taught in a bonus round instead of being left to the grid: the round names a park, gives its category, and asks what that category is mainly for.
Where the data comes from
Park names, countries, IUCN categories, establishment dates, areas and the supporting facts come from Wikidata under CC0 1.0. Each reveal reproduces the park's verbatim English Wikipedia extract under CC BY-SA 4.0 and links to the article.
The reveal also prints the exact figures the grid only ever bands: the founding year, the area in square kilometres, and any heritage designation on record. The other bonus rounds compare two parks by area, ask which country a park is in, and ask which decade a park was established.
Where the data comes from
Parkdle is built on 2 openly licensed sources, frozen into the game before it goes out. The licence is named next to each one above, and the same list, with the per-record detail, is on the Parkdle attribution page.
The writing you get on the reveal is english Wikipedia verbatim extract, CC BY-SA 4.0. It is quoted or generated from the source, never written by a language model while you play, and it carries its own credit and licence line wherever it appears.
The pool holds 808 entries and the calendar already runs 808 days, one answer a day, no repeats. That sequence was built offline before release and ships inside the game, so it is the same for everyone and nobody can read tomorrow out of the page. The day rolls over at midnight UTC.
Nothing is lost when a day ends. Every puzzle that has already run stays playable in the archive, and its answer, with the same reading you would have got by finishing it, is published the day after it ran.
Every past answer · Sources and licences in full · Parkdle on Gamecove
Questions
- Is it the same puzzle for everyone?
- Yes. One park a day, the same one for every player, rolling over at midnight UTC. The frozen calendar runs for 808 days, one for each park in the pool.
- Do I need an account?
- No. There is no sign-up and nothing about you is stored on a server. Your streak and your stats live in your browser's local storage, so clearing site data for the domain clears them.
- Why do some parks show an unknown IUCN category?
- Because Wikidata records none for them, and the game does not invent one. The cell reads unknown and never scores as a match, even when your own guess is also unknown.