The non-profit operator of the official RuneScape and Old School RuneScape wikis presents Search Console data showing steady/rising impressions but collapsing clicks — the classic 'buried in SERP' pattern. They argue that even for queries where they are the demonstrably canonical source (with Jagex officially pointing players to them), Google now surfaces Fandom pages, scraped mirrors, and AI-generated clones above them, indicating a category-level algorithmic bias against community-run referenc
By submitting the Weird Gloop post to HN with the framing 'a new Google Jail for independent wikis,' the submitter amplifies the position that this is a broader, categorical pattern rather than a one-off site issue. The 370-point score and 140 comments suggest the framing resonated with operators who have seen the same behavior.
The editorial emphasizes that pages remain crawled and indexed, so there is no technical violation to fix and no notification from Google. Standard remediation — schema audits, Core Web Vitals, freshness, backlink cleanup — has produced no measurable recovery, leaving affected operators with no actionable feedback loop and no appeal path.
Synthesizing reports from the HN thread — including Arch Wiki mirror operators and multiple wikis that migrated off Fandom for editorial reasons — the editorial identifies a specific pattern: Fandom pages outrank the originals even when the Fandom version is a stale fork. This suggests Google's ranking signals reward aggregator scale over content provenance or canonical authority.
For a decade, following Google's guidelines and producing quality content was understood as sufficient to earn organic traffic. The editorial argues that deal is now visibly one-sided — Google unilaterally changed the terms without notifying affected sites, breaking the reciprocal expectation that had underwritten the entire specialist reference-site economy.
Weird Gloop, the non-profit that runs the official RuneScape and Old School RuneScape wikis, published a post titled "Google Jail" describing what they believe is a systematic Google Search demotion targeting independent wikis. According to their data, organic traffic from Google dropped sharply and did not recover, even as the pages continued to be crawled and indexed. For queries where their content is demonstrably the canonical source — including content the RuneScape developer Jagex officially points players to — Google now surfaces Fandom pages, scraped mirrors, and in some cases AI-generated clones above them.
This isn't a deindexing story; it's a ranking story, which makes it harder to diagnose and almost impossible to appeal. The site owners describe the classic Search Console pattern of the last two years: impressions holding steady or rising while clicks crater, meaning Google still shows the page in results but far enough down the SERP that nobody clicks. Weird Gloop notes they've tried the usual remedies — schema audits, Core Web Vitals, content freshness, backlink cleanup — with no measurable effect. The post frames this as a category-level signal: independent, community-run reference sites are being algorithmically pushed behind aggregators.
The HN thread (370 points at time of writing) surfaced similar reports from operators of the Arch Wiki mirrors, several fandom-adjacent wikis that migrated off Fandom for editorial reasons, and a handful of hobbyist references. The through-line: sites that left Fandom, or never joined it, are seeing Fandom outrank them for their own content — sometimes with the Fandom page being a stale fork of the original.
For a decade, the implicit deal for anyone running a specialist reference site was: write good content, follow Google's guidelines, and organic search will find your audience. That deal is now visibly one-sided, and the party that changed the terms isn't telling you what changed. Google's public guidance still says "focus on people-first content," but the operational reality for small publishers is that ranking is now dominated by domain-level trust signals that are effectively unreachable if you're not already a large brand.
The Fandom angle is worth naming directly. Fandom (formerly Wikia) has become the aggregator of last resort for game and pop-culture wikis: ad-heavy, autoplay video, aggressive interstitials, and — critically for this story — often hosting stale forks of communities that have since moved elsewhere. When Google ranks the fork above the actively maintained wiki, it's not just annoying for readers; it materially damages the site that's actually doing the editorial work, because they lose the traffic that funds hosting, moderation, and — for Weird Gloop specifically — a paid staff.
There's a second dynamic that developers should recognize: this is the same discoverability collapse that hit independent tech blogs, Stack Overflow answers republished on scraper sites, and small package documentation. The winners of the last two years of Google ranking changes are large aggregators and AI-generated content farms; the losers are the primary sources those aggregators scrape from. The RuneScape wiki case is unusually clean because the developer of the game itself endorses the independent wiki, so "authority" arguments collapse. If a first-party-endorsed reference can't outrank a Fandom fork, the ranking signal isn't measuring authority in any sense a practitioner would recognize.
Community reaction on HN was less "this is new" and more "welcome to what everyone else has been living through." Several commenters pointed to the September 2023 Helpful Content update and the March 2024 Core Update as the inflection points where small-site traffic patterns broke. Others noted that the rise of AI Overviews compounds the problem: even when you do rank, Google increasingly answers the query in-place, so the click never happens.
If you operate or depend on an independent reference site — internal docs portal, OSS project wiki, niche knowledge base, developer blog — treat Google organic as a degrading channel, not a growth channel. Concretely:
Instrument for the demotion pattern before you get hit. Watch impressions-vs-clicks in Search Console at the query level, not just the site level. The signature of "Google jail" is stable impressions with collapsing CTR on queries where you're the canonical answer. If you see that, no amount of on-page SEO work will fix it.
Diversify discoverability channels now, while you still have the traffic to bootstrap them. RSS, an email list, a Discord or Matrix presence, cross-posting summaries to platforms where your audience already is (Reddit, HN, Lobsters, dev.to), and — increasingly — showing up in the training data and retrieval sets that LLMs use. The realistic 2026 answer for niche reference sites isn't "win Google"; it's "become the source LLMs cite," which requires clean, well-structured, licensable content and a stable canonical URL.
For OSS projects specifically: stop assuming your GitHub Pages docs will get found via search. Link them aggressively from the README, package registry description, and any first-party channel you control. If a community fork of your docs ends up on a Fandom-style aggregator, it will likely outrank you and there is no reliable remediation path.
The Weird Gloop post will resonate because it's a rare case with unambiguous ground truth — the game's own developer endorses the independent wiki, and Google still won't rank it first. Expect this to become a reference point in the broader conversation about whether search is still a viable distribution channel for independent publishers, and whether antitrust and DMA-style interventions start looking at ranking behavior, not just default-search deals. In the meantime, the practical move for anyone shipping a reference site in 2026 is to plan for a world where Google is one channel among several, and where the traffic you can't replace is the traffic you already had.
Top 10 dev stories every morning at 8am UTC. AI-curated. Retro terminal HTML email.