USAMRDC and the 1989 Prototype for Psychic Targeting
By Eleanor Voss ·
Kept secret for eleven years, a final report shows the Army's effort to turn subjective psychic 'responses' into a mathematical probability list for targeting unknown sites.
The human cost of the Stargate era was not measured in casualties, but in the total surrender of the interior mind to the needs of the state. Men and women were recruited to act as biological sensors, their subjective impressions of distant places stripped of nuance and converted into data points for the United States Army.
This column is based on OCR-extracted text from the reading-room page of the released report. The record is a "Final Report-Task 6.0.3" dated October 1989, titled "A Prototype Analysis System for Special Remote Viewing Tasks."
The Shift to the Operational
For years, the investigators at SRI International—a nonprofit research institute in Menlo Park that separated from Stanford University in 1970—focused on laboratory settings where the targets were known. In the lab, a viewer could be told to see a specific building, and the results could be verified. But the record shows that by 1989, the project had moved toward "intelligence-oriented tasks," where the targets were generally unknown.
In these operational environments, the goal was no longer to prove that remote viewing worked, but to make it useful. The report, authored by Virginia V. Trask and Wanda L. W. Luke—both of whom are placed by the record as researchers on this task, with no other details on file—along with Thane J. Frivold and Edwin C. May, argues that a high-quality remote viewing session does not always provide useful intelligence.
Edwin C. May had already established himself as a key figure in the Stargate Project, the secret Army unit founded in 1977 to investigate psychic phenomena for military intelligence. May’s role here was to bridge the gap between the raw psychic experience and the needs of a commanding officer. The document explicitly states that an analyst must provide "a priori probabilities that individual RV-response elements (or concepts) are present at the target site."
The Logic of Fuzzy Sets
To achieve this, the authors implemented something called "Fuzzy Set Formalism." In standard logic, a thing is either present (1) or absent (0). Fuzzy set theory allows for a membership value between 0 and 1. If a viewer says they "perceive glassware and smell organic chemicals," the analyst does not have to decide if that means a biological warfare (BW) facility exists; they simply assign it a value, such as 0.6, representing the degree to which the analyst is convinced the element is present.
This was the machinery intended to industrialize intuition. By using these values, the system could compare a viewer's response against a "template" of a known target type. If the template for a BW facility included "glassware" and "organic chemicals," the system would calculate a "figure-of-merit" to determine the probability that the site was, in fact, a BW facility.
This mathematical layering connects May and Frivold—who jointly developed this quality-assessment system—directly to the operational needs of the Army. The goal was to produce an "Optimized Probability List," a prioritized set of target possibilities derived from a single session. The record is clear: the system was designed to be "sensitive to the performance history of the viewer," meaning the Army was tracking which "psychics" were reliable at spotting specific things, like airports or laboratories, and weighting their data accordingly.
The 65 Simulated Targets
To test this prototype, the team used a pool of 65 targets, all within a ninety-minute drive of SRI International. These were not random sites; they were chosen because each one "simulates an operational site of intelligence interest." The list is a sterile catalog of potential military targets: "Weapons Storage," "Naval Air Station," "Nuclear Accelerator," and "Toxic Waste Storage."
This testing phase was overseen and approved by Murray J. Baron, the Director of the Geoscience and Engineering Center, an organization the record links to the report's funding and technical approval. Baron's signature gave the protocol the weight of official military doctrine.
Crucially, the document was classified by the USAMRDC, the Army's medical materiel developer responsible for medical research and acquisition. The fact that a medical research command exercised classification authority over a project involving "Target Templates" for biological warfare facilities reveals a specific intersection of interests. The USAMRDC was not interested in the philosophy of the mind; it was interested in the acquisition of target intelligence.
The Desk's Reading
When one looks at the distance between the raw data—the "sketches and written phrases" of a viewer—and the "Optimized Probability List" delivered to a commander, a gap appears. The report notes that it remains the responsibility of an "intelligence analyst" to determine if the data is ultimately useful, but it never names who those analysts were or which operational units received the lists. The desk's reading is that the document describes a processing pipeline while deliberately obscuring the destination. The specific field units that consumed these "a priori probabilities" for actual action were kept off the page.
Furthermore, the pattern suggests that the "Fuzzy Set Formalism" was less a functional tool for accuracy and more a linguistic shield. By applying the veneer of computer science and mathematical formalism to the inherently subjective and unstructured language of remote viewing, the authors could justify budget allocations to skeptical military auditors. It is easier to fund a "prototype analysis system" based on fuzzy sets than to fund a project based on the "smell of organic chemicals" reported by a psychic.
If the shape of this file is what it appears to be, the project reached its ceiling of viability exactly here. The transition from a "prototype" to an "operational system" likely never occurred because the a priori assessments failed to yield actionable ground truth in the field. The report focuses entirely on the methodology of prioritization rather than the verification of results.
The most telling detail is the total absence of cross-correlation. A standard intelligence prototype would seek to integrate its findings with SIGINT (signals intelligence) or IMINT (imagery intelligence). This system existed in a total vacuum, a siloed "black box" capability. The desk's reading is that it was kept isolated not for security, but to protect the project from professional ridicule. The USAMRDC and SRI International built a sophisticated mathematical machine to process ghosts, and when the ghosts failed to provide a reliable target list, the machine was quietly archived. The viewers paid for the experiment with their time and their privacy, while the Army paid for it with a budget line that, for a time, successfully masqueraded as science.