THE HOOK THE HORIZON EVIDENCE STANDARD
Check our work.
Fishing runs on good stories, real local knowledge, instincts people earned the hard way, and a pile of claims that got repeated until nobody could say where they started. Here’s how we try to keep those apart, and what to do when you think we got one wrong.
Before you trust anything else here
Almost nothing in fishing has one permanent answer. Regulations get rewritten, rivers move their channels, the fish move whenever they feel like it, and a rod that felt perfect turns into a different rod the moment somebody spools a different line on it. So sounding sure isn’t worth much. What’s worth something is knowing what we know, where it came from, what conditions it came out of, and where it stops holding up.
If a claim on this site can’t show its work, it hasn’t earned the right to talk you into anything.
Six kinds of evidence
Rules and anything with a clock on it
Rules, closures, license requirements, access restrictions, fire orders, consumption advisories, flow notices and emergency conditions come from the agency that issues them, not from somebody’s summary of them. When timing or jurisdiction matters we go to the original agency or the land manager. A cached search result or a five-year-old forum thread is a lead. It’s where you start looking, not where you stop.
Specs
Model numbers, dimensions, materials, line ratings, warranty terms, published operating limits and manufacturer instructions get checked against current manuals, spec sheets, or whatever the maker actually documents. That establishes what a product is designed to do and what the company is willing to claim about it. It says nothing at all about how the thing behaves in wind, with cold hands, on the third day.
Biology and the published research
For fish biology, water temperature, handling, habitat, weather and environmental relationships, we go to fisheries agencies, universities, extension programs, peer-reviewed work and recognized technical organizations. A relationship that holds up broadly and a prediction you can set a watch by are different animals, and we try not to let the first one get dressed up as the second. Nothing swimming out there agreed to match the chart.
Who actually did it matters
A firsthand account belongs to the person who lived it. Hunter is the authorial voice of Hook the Horizon. If he fished the water, used the gear, ran the test, caught or lost the fish, changed the setup, got it wrong, or noticed something worth keeping, he can say I because it is his experience. Rachel can edit, research, build, check sources and contribute behind the scenes. Her experiences do not become Hunter’s, and Hunter’s do not become hers. Any other contributor’s firsthand experience stays clearly attributed to that person too.
The account itself is the source for the event. Field notes, photographs, receipts, logs, messages, measurements and test records get pulled in when the claim is specific enough to need them. Waterbody or regional context goes in where it’s safe to share, along with the season or date, the conditions, the tackle, the effort, what it got compared against, and the result — when those actually matter to the point.
One day on the water can point at a pattern. It can’t establish one. This happened to us and this is generally true are two different sentences, and the second one needs more behind it.
People who know the water
Guides, biologists, shop staff, anglers who have fished one river for decades, land managers and local communities hold things no database has. When attribution matters, we say who is talking. And we keep track of which part is something they watched happen, which part is professional judgment, which part is opinion, and which part you could go verify yourself.
Guesses, labeled as guesses
Sometimes the evidence leans one direction without settling anything. When we’re interpreting, estimating, trying an idea out, or working with holes in the information, we say so. We don’t know yet is a more useful sentence than a confident guess wearing a clean shirt, and it’s usually the honest one.
How a claim gets on the page
- Name it. Work out what’s actually being claimed, and what somebody might go do because of it.
- Source it. Go to the most direct evidence that exists for that kind of claim.
- Try to break it. Hunt for the condition, species, water, setup, jurisdiction or sample size that makes the conclusion fall over.
- Say what it is. Official fact, manufacturer information, published guidance, firsthand observation, judgment, inference, or unknown.
Step three is the one that costs time, and it’s the one that catches things.
How hard we chase it depends on what it costs to be wrong
A knot variation and an emergency closure don’t get the same treatment, and pretending otherwise just means one of them is being checked badly. The more a claim touches legality, access, personal safety, fish health, habitat, real money, or a trip somebody planned their whole summer around, the stronger and more current the evidence has to be.
Rules and conditions move
Confirm the current rule with the authority that controls that water or that land before you act on anything you read here. We link primary agencies where it’s practical, name the jurisdiction, and try not to flatten a local rule into a universal one. For weather, flows, tides, fire conditions, ice, closures and advisories, the timestamp is part of the information. A condition we published is a photograph of one moment, and it keeps aging after we hit publish.
Water we won’t pin
Being accurate doesn’t require publishing every coordinate. We’ll generalize or hold back a location when the precision would add pressure to fragile habitat, expose a small fishery, encourage trespass, or burn somebody’s private access. What we won’t do is invent a different location to fill the hole.
Gear testing
We test things, and a test is only worth reading if you can see what was actually tried and how big a conclusion it can hold up. Where it matters, that means the exact product or configuration, the rest of the setup it lived in, the conditions, the technique, how many trips or how long, what it got compared against, what failed, and what happened.
An afternoon with a reel isn’t a long-term test. A first impression isn’t a durability review, and a demonstration run by the company that makes the thing isn’t an independent result. None of this has to read like a lab report — but if a claim is doing work, the details underneath it should be recoverable and roughly the size of the claim.
If a brand supplied the item, loaned it, discounted it or sponsored it, that gets disclosed where it matters. Whether something pays an affiliate commission has never once decided a verdict here. And best only means something attached to a job, a budget and a set of conditions. On its own it’s a marketing word.
Things that look like proof and aren’t
Worked through at length: what a photo can and cannot prove, why your best memories are your least reliable data, and what a real lure-color test would have to look like.
The short version: a generated image documents nothing, and one catch is one catch. Between an anecdote and a controlled comparison sits a lot of work that nobody did. A product page is the manufacturer talking, a search snippet is a pointer to a source instead of a source, and an undated regulation can’t tell you whether it’s current — which was the whole job the date had. A claim also doesn’t get truer by being repeated, no matter how many forums repeat it, and confidence isn’t a citation even when it’s written really well.
AI, and who did the fishing
Straight answer: yes, AI is part of the production stack. It can help organize research, compare sources, structure notes, edit drafts a person wrote, check repetitive language and put site material together. It does not author Hunter’s voice, supply his personality or judgment, invent the first-person I, or become the person who went fishing. If the wording stops sounding like Hunter, technical correctness does not rescue it. When a post comes out of something that really happened — a trip, a test, a purchase, a catch, a failure, a preference, something somebody noticed — getting help with the editing doesn’t take that authorship away or turn it into a day a model had.
What AI doesn’t get to do is make things up. Not participation, results, measurements, quotations, preferences, reactions, memories, field notes or locations. Any material claim about the outside world still goes through human review and still needs evidence that fits the claim.
When we get it wrong
This stuff ages. Gear gets replaced, access changes, agencies rewrite rules, and better evidence turns up and knocks over something we were fairly sure about. When a real error is confirmed, we fix the page. When the facts underneath it simply moved, we update the guidance and date the revision where the timing matters.
If you catch something that looks wrong, send the page, the exact sentence, and the best source you’ve got. You disagreeing doesn’t automatically mean we’re wrong — but a specific correction with something behind it gets read, every time. That includes corrections from people who are brand new at this and worried about sounding stupid sending one. You won’t.