Editorial policy
This page sets out commitments you can hold us to, not aspirations. Where a commitment is enforced by a build script rather than by someone remembering, we say which script. Where it depends on human judgement, we say that too.
What do we publish?
Pages that answer one question for one kind of reader. A calculator page exists because someone needs a number; a pool page exists because someone is deciding where to launch. If we cannot name the question a page answers, we do not build the page.
Every page carrying factual claims must contain at least one thing a reader cannot easily get elsewhere: a figure read out of a primary document rather than copied from another site, a worked example run through our own code, or a discrepancy between sources that nobody else has flagged. A page that only restates what is already on the first page of results does not go up.
What will we not publish?
- Unsourced figures. Not as estimates, not with a caveat. Ourfact-checking policy describes the build gate that enforces this.
- Templated pages at scale. No generating a page per species or per pool by swapping a keyword into a fixed structure. Two pool guides written properly beats fourteen generated.
- False precision. No single-number answer where the source data supports a range, and no percentage scores for things nobody has validated a model for.
- Regulatory advice as fact. We publish what a specific chapter of code says on a specific date, with a link. We do not tell you what is legal, because the agency does that.
- Invented authority. No fabricated author biographies, no citations to institutions that did not publish the thing we are citing, no "studies show".
How do we source a claim?
Primary documents first. For river geometry that means the Corps of Engineers water control manual for the dam in question, not a tourism page quoting it. For regulations it means the administrative code, not a regulations summary. For fish weight equations it means the published paper, with the sample size stated.
Each source gets a calibration grade — measured, regional or generic — describing how well the source matches the thing we are applying it to. That grade is shown on the page, and ourresearch methodology defines each grade with examples.
How do corrections work?
Report an error through the contact page and we check it against the source rather than against our memory of the source. If we were wrong, we fix the page and note what changed. If we were right, we write back and say why, with the citation.
Corrections that could affect a legal obligation — a bag limit, a size limit, a license requirement — are treated as urgent and fixed the day we confirm them. Everything else goes into the next scheduled review.
Separately, no source on this site can go more than 180 days without a recheck, because the build fails if one does. That means a page can become stale in our queue but it cannot become stale on the site.
How is advertising kept away from editorial?
Structurally, not by policy alone. Advertising is served programmatically through Google AdSense; no human at an advertiser sees a page before it publishes and no advertiser can request a change. We take no sponsored posts, no paid placement in any comparison table and no affiliate commission on tackle, so there is no gear recommendation on this site that pays us if you follow it.
Placement is constrained in the page layout rather than left to judgement: no ad unit is rendered within a screen of a calculator result panel, and every ad box reserves its height before it loads so it cannot push content under your thumb as you read.
Where does AI fit?
We use it for drafting and for checking our own arithmetic, and a person decides what publishes. The line we hold is on substance: a page must contain something derived from a primary document, our own code or our own time on the water. A language model summarising the existing first page of results has no information to add, and publishing it would be exactly the pattern search engines have spent two years demoting.
Questions about how we work
Do advertisers influence what you publish?
No. Advertising on this site is served programmatically through Google AdSense. No advertiser sees a page before it publishes, no advertiser can request changes, and we accept no sponsored posts, paid placements or affiliate commissions on tackle.
Do you use AI to write these pages?
We use AI tooling in drafting and in checking our own arithmetic, and a human decides what publishes. What we will not do is publish a page whose only content is a language model restating what other sites already say. Every page has to carry something a reader cannot get elsewhere.
What happens if I find a mistake?
Write to us. We check the claim against its source, correct it if it is wrong, and note what changed and when at the point of the correction. Errors that could affect a legal obligation, such as a bag limit, get fixed the same day we confirm them.
Why are pages signed by a team rather than a person?
Because the review is done by a team, and inventing a named expert with credentials nobody can verify would be a worse signal than an honest team byline. Every page names what qualifies us: we fish this water and we check every number against the body that published it.