What kind of page do AI answers actually cite?
The evidence
The narrow site beat the big publisher.
Of one engine's seven cited domains, exactly one was mainstream press. The other six were fan-built wikis and single-franchise sites. A registered expectation naming the obvious industry publishers and the manufacturer returned zero matches across three calls.
Both of the other engine's top two positions were the same kind of site.
Positions one and two went to single-subject sites, ahead of the BBC, IGN, CNET and USA Today, all of which it also cited further down.
The owner of the fact was not the source for the fact.
The company that sets the date, announces the date and sells the product was cited zero times out of three on a question about that date. Its own domain was in the registered match list before the pull.
This is a direction, not a mechanism.
These sites were cited; nothing here measures why, and a correlation between being single-subject and being cited is not a causal rule. One query, one day. It is enough to change what you test next and not enough to rebuild a site around.
What would have to be true for this to be wrong
One query string to ChatGPT and Perplexity, three calls each, plus Google organic, Houston, Texas, October 24, 2026, via DataForSEO. Every cited domain published in full on the study page. Platform-default budget-tier models.
Every row behind this answer is published in the standing study at cultmath.com/research/games/gta, which is re-run on a schedule and corrected in place, dated.
Published 2026-10-24 by Ipalibo Da-Wariboko, CULT+MATH LLC. Where a number could not be sourced, this page says so and gives none.
Keep your own call on this question on record. DEOS is the decision app built on this research: talk it through, and it writes the record, checks it against your past calls and brings it back when it is time to review.
Free for 30 days, no card. Then $99 a month.
“What kind of page do AI answers actually cite?”
This answer, its date and the sources under it are available to ChatGPT and Claude through the CULT+MATH app. Connect it in either one, ask the question above, and the findings come back cited rather than summarized.