Hoeflin Power Test: Independent Exposure Audit

The Power Test is an unsupervised, untimed high-range intelligence test created by Ronald K. Hoeflin, the philosopher and psychometrician best known for founding the Mega Society and designing the Mega and Titan tests. It consists of 36 problems and draws on selected problems from Hoeflin's earlier Mega, Titan and Ultra tests. The problems emphasize spatial, numerical and logical reasoning: counting the parts into which a solid can be divided, visualizing intersections of surfaces, extending number sequences, and similar tasks that are intended to be solved with extended, unaided effort rather than speed.

The Mega Society, which accepts the Power Test for admission, currently lists a raw score of 34 out of 36 as qualifying. That figure is an external fact set by the society, not by this site, and it may change; consult the official Power Test page for current requirements.

Because the Power Test is taken without supervision, its meaningfulness depends on candidates working the problems themselves. Every public copy of a test prompt, every forum thread dissecting an item, every posted hint, answer request or worked solution changes what an unsupervised score can be taken to mean. A candidate who has seen a decisive method hint — even accidentally, through a search result — is no longer solving the same problem as one who has not. Exposure does not have to be deliberate cheating to affect interpretation or validity: it accumulates quietly, in public, over decades.

The Power Test inherits an additional risk from its history: because it reuses problems from the Mega, Titan and Ultra tests, material published about those predecessor tests — including decades-old journal discussions and archived mailing lists — can expose Power Test items indirectly. This predecessor and cross-test carryover is one of the largest exposure categories this audit records.

This site is a public observatory: a dated, sourced archive of where and how Power Test material has become publicly discoverable. Each finding records the affected item numbers, the type of exposure (prompt publication, discussion, answer request, method hint, candidate answer, verified solution, offers to share answers, evidence of private sharing, or predecessor carryover), a confidence level, a discovery date and an assessment. Findings are located through systematic search runs documented in the research updates, verified against the cited source where possible, and classified according to the published methodology.

The archive distinguishes carefully between verified evidence (the source was retrieved and its content confirmed), unresolved leads (material reported or suspected but not yet confirmed), and predecessor or cross-test exposure (material about a Mega, Titan or Ultra problem that carries over to a Power Test item). The full baseline analysis is published as the public-exposure audit document, and the editorial rules the project follows are described on the about and editorial policy page.

One rule overrides everything else: this site does not republish final answers, answer keys, or decisive solution steps — not even partially, and not to attract search traffic. Findings describe that and where exposure exists, never the exposed answers themselves.

The archive currently documents 77 findings, of which 4 changed an item's exposure classification after the baseline audit. Monitoring began from the baseline audit snapshot dated 4 August 2026; the figures below are live values calculated from the archive as it stands today, not the baseline numbers.

• Prompt publication / reproduction29
• Method hints20
• Candidate answers14
• Public discussion14