Spies, drugs, and mind tricks: what actually held up

ACH is not just an analysis tool — it is a structured way of forcing any team to commit to falsifiable hypotheses before seeing the outcome. The banned move is adding hypotheses after the evidence is in. Run it before a product launch decision or a credit model revision and you get a pre-mortem with teeth.

The CRV structured session format — encrypted cue, first ideogram, sensory probing, concept elaboration, summary — is a disciplined protocol for separating first impression from reasoned inference. It works as a structured intuition-elicitation method regardless of whether remote viewing is real. The stages prevent the analyst from anchoring on their first guess.

The Society for the Investigation of Human Ecology was funded partly to cover research the CIA explicitly did not want, creating a credible academic profile that included things the sponsor genuinely did not care about. The cover works because it is not entirely fake — real research happens, just not the research that matters. The structure is fully documented here.

Fictional private foundations that fund real university researchers — who do not know the true sponsor — provide a decade-long information suppression layer. The structure works because the researchers genuinely believe in the independence of the grant. The Senate documents confirm it operated undetected for over ten years.

The CIA's FBIS systematically translated and tagged foreign open-source scientific and technical press to get early warning of rival research programmes at near-zero marginal cost. The competitive intelligence was in the open literature; the advantage was the systematic collection and tagging, not access to secrets.

The STAR GATE programme used a consistent structured format — Data Summary, Project Evaluation Feedback, Annotated Map — to aggregate multiple viewer outputs per task and score each against ground truth. The closed feedback loop between session producer and operational consumer is rarer in intelligence than it should be and is the only thing that allowed any calibration to happen.

The 1969 psychiatric technology workshop documents a tiered privilege system where the cost of privileges decreases at higher compliance levels — a token economy that demonstrably shaped behaviour in institutional populations. The key finding is that the reward needs to feel earned, not given, and that the tier structure itself motivates progression more than any individual reward.

SAIC researchers proposed using Shannon entropy to characterise how much information a target contains, as a predictor of session quality. The underlying idea — that stimulus or task complexity determines how much signal a human-judgment system can surface — is transferable to any domain where you are selecting tasks to evaluate a model or analyst.

Commissioning a second team to challenge the same evidence base with explicit licence to reach different conclusions — Team B — is not just a debate exercise. The key design feature is that Team B has full access to the same raw intelligence as Team A and produces a written product of equal formal standing. The failure mode is equally documented: Team B 1976 overestimated Soviet capability significantly, showing that adversarial challenge without additional evidence generates confident errors as readily as confident truths.

SAIC's Decision Augmentation Theory reframes anomalous effects not as physical influence but as information-biasing decision outcomes — the effect shows up in which decisions get made, not in the physical system being influenced. This framing is testable where direct physical effect is not. The statistical signature of decision biasing is different from the signature of random chance.

A CIA subproject proposed empirical behavioural rating categories for assessing personnel indirectly — inferring psychological profiles from observable behaviour without the subject's knowledge. The method does not require access to the subject; it requires a systematic taxonomy of observable behavioural proxies mapped to the traits you care about.

Every STAR GATE session was followed by a standardised intelligence evaluation sheet scored by the operational customer — not the producer. This separated quality assessment from production in a domain where the producer has enormous incentive to rate their own output highly. The customer score, not the viewer's self-rating, determined whether a session was counted as useful.