Sources Policy
Almost nothing on this site is a matter of opinion. Every claim is about someone else's software or someone else's game, which means it is either right or it is wrong, and it is checkable. This page sets out where the checking comes from.
Why a page like this exists
Guide sites tend to accumulate claims that nobody can trace. A number gets repeated from one article to another until it has been in circulation long enough to feel true, and by then the original source, if there ever was one, is unrecoverable. That is how a game community ends up confidently wrong about a mechanic for two years.
The defence against it is boring and works: rank your sources, prefer the primary one, link to it, and say plainly when you do not have one. What follows is the hierarchy this site actually uses, in order of what wins when two sources disagree.
The source hierarchy
Official documentation and the games themselves
Grinding Gear Games' developer documentation, patch notes, official announcements, and the observable behaviour of Path of Exile and Path of Exile 2. Anything stated here overrides every tier below it, including our own testing.
The platform being described
When a guide explains what a screen shows or where a setting lives, that interface is the source. The same applies to any statement the platform makes about its own methodology.
Our own hands-on testing
Steps performed by hand, in the current league, on the current version, with the date recorded. Testing can confirm a procedure and can flag that something has changed, but it cannot outrank an official statement.
Established community reference
The community wiki, Path of Building's own repository, and long-standing community tools. Useful for terminology and orientation, and occasionally the only place a niche interaction is written down at all.
What does not count as a source
Never used to support a claim
- Forum, Discord or Reddit posts repeated as fact, however confidently they were written
- Other guide sites, either as a source or as material to rewrite
- Anything we cannot independently check, including private data and unnamed insiders
- Machine-generated summaries of any of the above, which launder unsourced claims into confident prose
- Our own earlier articles. A claim is checked against its original source, not against something we published last year
That last one matters more than it sounds. Self-citation is how a site's mistakes become permanent. If a guide here is the only thing supporting a number, the number is not supported.
How sources appear on the page
Where a claim rests on something official, the guide links to it directly rather than paraphrasing and hoping. Links go to the primary source, not to an article about the primary source.
We do not reproduce other people's work to make a point. No copied paragraphs, no lifted tables, no rehosted screenshots of another site's interface. Where someone else has explained something better, the honest move is to link out and say so, which costs nothing and is more useful to you anyway.
When sources disagree
Official statements win. If the documentation says one thing and community consensus says another, the guide follows the documentation and notes the disagreement, because the disagreement is usually the interesting part.
If our own testing contradicts an official statement, we say both, describe exactly what we did, and treat it as an open question rather than quietly picking a winner. That situation usually means a patch changed something and the documentation has not caught up yet.
When there is no source
Some things genuinely are not documented anywhere. Internal weighting, exact filtering thresholds and undisclosed implementation details fall into this category, and no amount of confident writing turns a reasonable inference into a fact.
Where that happens, the guide says so. Language like "appears to", "in practice" and "not publicly documented" is doing real work when you see it here. It marks the boundary between something you can rely on and something that is our best reading of observed behaviour.
A worked example
Rather than describe the process abstractly, here is one real claim from this site and the full chain behind it.
How one claim gets on the page
Sources go stale
A correct source can become a wrong claim without anyone touching the page. Patches change limits, interfaces move, and documentation gets revised. Every guide that depends on a source is flagged for review when that source changes, and anything we have not been able to re-verify is marked rather than left looking current.
Challenging a source
If you think something here cites the wrong source, misreads the right one, or states as fact something that was never documented, that is worth reporting and it will be checked against the original. Disagreements about sources are the most useful mail this site receives.
