How We Research Toys, Games & Collectibles

We research; we do not test

CurioRank is a research desk, not a testing lab. We do not buy every board game, model kit and collectible on the market and put them through hands-on trials, and we never write as though we had. You will not read “we played 40 games at our table” here. You will read researched, compared and evaluated, because that is what actually happened.

That limit is the whole reason this page exists. A site that is honest about what it did not do is easier to check than one that isn't.

Where the words come from

The manufacturer's own material first. Player counts, play time, piece counts, scale, recommended age, materials and the manual — published by the people who made the thing. Publishers and makers we cite directly include LEGO, Stonemaier Games, Czech Games Edition, Bandai, Robotime and Metal Earth.
Hobby databases and specialist coverage.BoardGameGeek for weight, ratings and expansion relationships; Brickset for set data; publisher and tournament bodies where a category has one. These are the reference works the hobby itself uses, and we link the specific page, not an organisation's front door.
Child-development and safety sources, where age is part of the question. HealthyChildren.org (American Academy of Pediatrics), the NAEYC, the CDC's developmental-milestone material, the CPSC for recalls and toy-safety labelling, and peer-reviewed work indexed on PubMed. We use these to check an age claim, not to dress up an opinion in a citation.
Aggregated retailer ratings, described for what they are. We read the public star average and rating count, and look for complaints that recur rather than for a quotable line. We do not read every review and we do not run a panel. A retailer aggregate is not restricted to verified purchasers, so we treat it as a public aggregate rather than as owner testimony.
Community consensus, paraphrased and attributed to the room. 45 of our articles carry a community-consensus block. It summarises what a subreddit or hobby community broadly agrees on and names which communities it came from. We never reproduce an individual comment as though it were an interview, and we never invent a username.
Confirm the listing is real and live. A recommendation pointing at a dead or mismatched product is worse than no recommendation, so a pick that cannot be bought loses its buy button rather than quietly keeping it.

What we won't do

We won't claim hands-on testing we haven't done, or imply it with staged photography or invented play notes.
We won't invent a statistic, a source or a product. If we cannot open the primary source and read the sentence that supports a figure, the figure does not go in. A citation whose title promises a specific page while the link goes to a homepage is worse than no citation, so we deep-link or we retitle.
We won't fabricate community quotes. Consensus is described; individuals are not ventriloquised.
We won't hide the negatives to protect a commission. Every pick carries what is wrong with it, and a category where nothing is worth buying gets said out loud rather than padded with a fifth-best option.
We won't take sponsored placements, paid inclusion or free product in exchange for coverage. Nobody can buy a rank, a score or a rewrite. The site is funded by affiliate commission on links readers choose to follow; that pays for the research and it does not decide the order.
We won't give a safety verdict we cannot source.Age grades and safety labelling are the manufacturer's statement or a regulator's, and we attribute them that way rather than asserting them ourselves.

What an “Updated” date means here

Most sites stamp a fresh “Updated” date on a page whenever anything in the file moves — a re-synced price, a rewritten meta description, a schema change. That is a freshness claim the article itself does not support, so we stopped making it. An article here shows an “Updated” date only when the body of the article actually changed, and shows nothing at all when it did not. A page displaying only its publish date has not been quietly re-dated; it simply has not been rewritten.

What would change our mind

A recommendation gets rewritten when the maker changes the product, when a listing dies or is replaced by a newer generation, when recurring owner complaints point at a fault we did not weight heavily enough, or when a source we relied on turns out not to say what we thought it said. We would rather publish the correction than defend the pick.

Corrections

If something here is wrong, we want to fix it rather than defend it. Prices move, listings change and products are discontinued, so parts of this site go stale between updates by default. Tell us the page, the sentence or figure you think is wrong, and the source that says otherwise if you have one. A factual error, a dead or mismatched link or a price that no longer matches the listing gets fixed; where a correction changes the substance of a recommendation rather than a number, the page says what changed instead of being edited silently.