A retired police officer from Burbank lay back in a shielded room in California, closed his eyes, and started describing a crane in Soviet Kazakhstan.
That was July 1974. The viewer was Pat Price, the target was a facility the CIA knew only as URDF-3, and the gantry crane he sketched had eight wheels. So did the real one, when analysts later compared his drawing to a tracing made from satellite imagery. Sessions like that are why a programme most people file under Cold War comedy ran for two decades and got through roughly $20 million of public money.
Why anyone funded this
Paranoia, mainly. The Federation of American Scientists’ summary of STAR GATE records that between 1969 and 1971, US intelligence sources concluded the Soviets were running serious “psychotronic” research, with spending guessed at around 60 million roubles a year by 1970 and several times that by 1975. Nobody in Washington had firm evidence the Russians had actually found anything. What they had was a gap they couldn’t measure, and gaps like that have historically been the most expensive kind to fund.
Funding began flowing to a coordinate-based project called SCANATE around 1970, and laboratory work started at Stanford Research Institute in 1972 under physicists Russell Targ and Harold Puthoff. Over the following twenty years the effort bounced between the CIA, Army intelligence and the Defense Intelligence Agency (DIA), picking up a new codename almost every time it moved: GONDOLA WISH, GRILL FLAME, CENTER LANE, SUN STREAK, then STAR GATE. Around forty people cycled through, about 23 of them viewers.
Staff had a term for a result so good it rattled the room. They called it an eight-martini result.
What the crane story leaves out
How good was that famous session, really? A CIA analyst sat down with two cassette tapes, 79 pages of transcript and 30 sketches, and wrote it up day by day. His review, now mirrored in the declassified analysis of the URDF-3 remote-viewing experiment, is not the document a believer would choose.
Before Price said a word, the experimenters handed him coordinates and showed him the site on maps, sixty miles west-south-west of Semipalatinsk. They called it a scientific military research and test area and said it sat 25 to 30 miles from the Irtysh River. He then produced a great deal that was simply wrong: a 500-foot antenna, an outdoor swimming pool, an array of telephone poles, none of which existed at the site. When an experimenter floated an alternative reading of one detail, Price changed his answer on the spot, which the analyst flagged as a sign he could be led. Asked whether four structures he had drawn might be the surface elements of an underground building, he said no. They were.
The analyst’s verdict, in full: the experiment was unsuccessful.
None of that undoes the drawing. Price sketched an eight-wheeled gantry running on widely separated rails above a two- or three-storey building, and that is what was there. Two things being true at once is what makes this material so slippery to handle.
The Nature paper and the argument that followed
In October 1974, Targ and Puthoff published their remote viewing results in Nature, reporting that independent judges matched viewer transcripts to real locations at rates well above chance. Publication in that journal was as close to instant legitimacy as science gets, and the backlash came fast.
Psychologists David Marks and Richard Kammann ran duplicate experiments in New Zealand and, writing in Nature in 1978, reported that their results did not verify the SRI conclusions. Their more damaging point came next. Marks argued in a 1981 letter on sensory cues that the SRI transcripts were littered with clues about the order in which sites had been visited: dates, references to earlier targets, offhand remarks. Judges could use those to match transcripts with no psychic help at all, and Marks reported perfect accuracy on some of them without visiting a single site. Puthoff answered with a rebuttal in the same volume, and the exchange rumbled on for years.
Two reviewers, one pile of data
“Psychic functioning has been well established.” That was statistician Jessica Utts in 1995, hired alongside psychologist Ray Hyman by the American Institutes for Research, after Congress ordered the programme moved back to the CIA for review. They were handed the same material and reached almost opposite conclusions.
Utts found in her assessment of the evidence that by the standards used anywhere else in science the effect held up: results far beyond chance, and consistent with findings from other labs. The size of that effect, using the measure statisticians rely on for how strong a pattern is, fell between what they call small and medium. Hyman, reading the same material, wrote in his evaluation of the programme that the earlier SRI work was methodologically inadequate, and that the newer experiments, however clean, had never been replicated by anyone outside the small circle that produced them.
They agreed on one thing: the numbers were strange. They split on whether strange numbers from one research lineage amount to a phenomenon, and Hyman’s position won the institutional argument. AIR recommended termination, and the CIA concluded no intelligence operation had ever been guided by remote viewing data.
What survived the shutdown
Twelve million pages of records went public in 2017, as documented by History, and the internet has been mining them for wonder ever since. What circulates is the hits and the eight-martini folklore. Nobody shares the four days of wrong guesses that came with them.
That’s a selection effect, and it has nothing to do with psychics. Give any process twenty years, a generous budget and enough attempts, and it will produce a highlight reel. The question worth asking of any programme is not whether it generated stunning results, but whether anyone bothered to keep the boring ones.
