In May 1968, a U.S. submarine, the Scorpion, disappeared with 99 men aboard. The entire Navy was at a loss to find her. Among those looking for her, one man held the key to recovering her. Yet, his ideas were nearly brushed aside. They were deemed too strange.

With that piece of the puzzle, Craven was sure he had figured out what happened to the Scorpion.

It seemed to Craven that the Scorpion had been battling an active, or "hot-running," torpedo, but hadn’t been able to turn to the east fast enough. The logic, the evidence—it all fit. Craven was convinced.

There was only one problem: Almost nobody agreed with him. The sonic experts, the torpedo experts, the submarine commanders all listened as Craven held forth with his theories, his evidence, and his logic, his voice rising and falling as if offering a Shakespearean soliloquy, albeit one punctuated with his own trademark maxims of the deep sea. But nobody of any rank, from the chief of naval operations on down, thought Craven could be right.

The acoustics expert at the Naval Research Lab—who had earlier provided Craven with the information on the explosions—was convinced that Craven was reading way too much into the data and was chasing ghosts. The only thing that turned east toward the Mediterranean, he believed, was Craven’s phantom trail. His arguments instilled some doubts even within Craven. Besides, it was the expert’s lab that was guiding the search vessel, and Craven needed his support if the ship was going to turn around and start searching to the east.

The officers in charge of torpedo safety at the Ordnance Systems Command soon joined the group of naysayers. They insisted that it was impossible for a hot-running torpedo to detonate inside a submarine. For detonation to occur, the command insisted, a warhead would have to run into something while moving at top speed. Then and only then would it go off. The Ordnance Systems commanders were backed up by the Bureau of Ships. A top submarine official was also firmly in doubt.

As the debate raged on, none of the men forgot that they were looking for their own dead. Still, at one point, in an effort to lighten things up a bit, the submarine official wagered Craven a bottle of Chivas Regal Scotch whisky that he would turn out to be wrong. Operational commanders took the official’s side in the bet.

The search vessel, the Mizar, had already dug up some tantalizing clues on the homeward-bound, or western, side of Point Oscar. There were three items found on the ocean floor that could have fallen from the Scorpion: a piece of elbow pipe, what seemed to be a woman’s umbrella, and a rope tied in a "monkey’s fist," the ball-shaped knot that sailors tie at the end of a mooring line to make it easier to catch when it is tossed onto a pier.

There was some argument within the Navy about whether the monkey’s fist the Mizar found was tied in the U.S. style or in the style favored by the Italian navy, but the umbrella, the operations officers believed, had to have come from the Scorpion’s crew. The submarine had made port calls, and the umbrella could have been someone’s souvenir or gift for a woman back home. Months would pass before Navy biologists declared that what looked like an umbrella was actually alive, one of the many odd creatures that live on the ocean floor.

Given the Mizar evidence and the strong opinions around him, even Craven began to wonder whether he was wrong—just "smoking opium," as he liked to say. But then again, maybe he was the only one who was right. Craven had no trouble believing either possibility, so he kept right on digging.

He arranged to have a ship drop small explosive charges at Point Oscar. By comparing the acoustic signatures collected at the site with the signals that later reached Norfolk, he would be able to figure out once and for all whether an explosion in the area would create echoes—the sonic ghosts that others said he was hearing.

When the explosives finally signaled through to Norfolk, they came through with no echoes. And when the new data were calibrated with the earlier signals from the Scorpion, it became clear that, not only had the Scorpion been traveling east, she was traveling east even faster than Craven had thought.

Craven was back to his torpedo theory. But he still wanted more evidence.

With typical dramatic flair, Craven arranged a re-enactment of his version of the tragedy. He needed a submarine simulator, and he needed Lt. Cmdr. Robert R. Fountain Jr., a former Scorpion commanding officer who had been detached from the submarine just before she embarked on her final mission.

Fountain was put at the simulator’s helm, and a computer was programmed to factor in the orders he gave as the simulator re-enacted various possible causes of the Scorpion’s loss. Ten different scenarios were tested this way, and 10 failed to create a match with the acoustic evidence. Then, Craven’s team asked Fountain to try one last time. They said nothing about a possible torpedo explosion; they simply told Fountain that he was heading home at 18 knots, leaving it to him to choose a depth. Craven then asked him to test his torpedoes. The team waited 10 or 15 minutes, giving Fountain a chance to stand calm. Then they rang an alert. "Hot-running torpedo in the torpedo room."

Without missing a beat, without waiting, without asking questions, Fountain ordered, "Right full rudder."

There it was. The sharp, 180-degree turn that Craven believed had been executed on the Scorpion. The simulation, as it continued, generated results that seemed to confirm the chain of events Craven believed had befallen the missing submarine. Chills shot through him when he saw the results. By now, he and several others attending this test were nearly certain they had replicated the Scorpion’s loss. No one told that to Fountain. No one told him he had just possibly re-enacted the circumstances under which the men he had once helped to command had died. Maybe nobody had to tell him. He left the simulator without asking any questions, without saying a word.

Craven’s compassion for Fountain and for the crew of the Scorpion couldn’t squelch the exuberance he felt. As a detective, he had come up with important new evidence, and he now raced to Vice Adm. Arnold Schade, commander of submarines in the Atlantic, and Adm. Bernard Clarey, the vice chief of naval operations. By now, even they were becoming intrigued by Craven’s detective work, but they remained unconvinced. As did the Ordnance Systems Command, which continued to insist that there was no way a torpedo could explode on board a submarine.

Nobody was ready to face the specter that, because of a faulty torpedo, the Navy itself was responsible for the deaths of those 99 men. Craven understood their reluctance, understood how difficult it was for the admirals to believe they might have been somehow responsible for a mistake that had caused the loss of so many people’s lives.

Despite the admirals’ reluctance, Craven wasn’t about to give up, not now that he was convinced he had enough information to find the Scorpion and prove what killed her. He began to mathematically construct a map of the ocean bottom in the area that interested him. He then began trying to place the Scorpion on that map by using Bayes’s theorem of subjective probability, the same algebraic formula he had employed during an earlier deep-sea search for a hydrogen bomb. [The theorem quantifies the value of hunches based on a series of bets that experts make and combines them to produce a prediction that may be more likely than one based solely on hard facts. The idea is to factor in the subconscious knowledge people may have.]

Few of the officers involved in the search for the Scorpion had taken much note of Craven’s earlier successes. And by the time Craven was finished explaining that he was going to factor hunches into his data, some of the operational commanders were convinced he had gone completely over the edge. To them, it sounded like he was talking about extrasensory perception.

Still, Craven pushed on, asking a group of submarine and salvage experts to bet on the probability of each of the different scenarios being considered to explain the Scorpion’s loss. To keep the process interesting, and in line with previous wagers, the men bet bottles of Chivas Regal on aspects of the accident that would help pinpoint where the sub ended up.

By the time the bets were finished and Craven sat down to draw a probability map, the calculations had become so complicated that he had to rehire the group of mathematicians who had helped him with the hydrogen bomb. They concluded that the Scorpion was less than eight miles east of Point Oscar, 400 miles from the Azores, on the edge of the Sargasso Sea.

Years later, the mathematicians would write a book based on their work with Craven, titled Theory of Optimal Search. The U.S. Coast Guard would adopt the method for search and rescue, and the Navy would use Craven’s interpretation of Bayes to help Egypt clear sunken ordnance from the Suez Canal. But in the Scorpion search, naval officers just shook their heads at Craven’s acoustic evidence and his probability map. Craven may have been convinced that the Scorpion lay farther east, but the Mizar had found the three scraps of debris to the west, and that’s where the Navy wanted to keep searching.

Weeks passed. Craven waited, trading messages nearly every night with the man in charge of the Mizar, Buck Buchanan. Buchanan was so emphatic about intending to find the Scorpion that, on the day he left port, he started growing a Vandyke and swore he wouldn’t shave it until he located the sub. But, by late August, nothing new had been found and the jubilation within the Navy that had accompanied the Mizar’s find of the supposed umbrella and the monkey’s fist knot diminished. By September, all of the likely spots between Point Oscar and Norfolk were almost ruled out. By October, the weather was getting so bad the Navy decided it would halt the search by the end of the month.

But the Mizar still hadn’t thoroughly searched east, and it had never searched the site Craven had pinpointed. By now, Buchanan was willing to point the Mizar east.

Almost as soon as the Mizar passed east of Point Oscar, its long-range sonar registered iron, and lots of it. The Mizar steamed ahead full speed, right past Craven’s point of highest probability, and then lowered its cameras for a look. All it found was rock filled with iron ore.

That was it. The end. The top brass had had about all the disappointment they could take. The decision was made: It was time to give up. Time to call Buchanan and the Mizar home.

Buchanan, pugnacious and stubborn, refused to accept their decision. He flashed a message to Craven: "Can’t you get the Navy to let us stay out another month, or a week or two weeks? Tell them I need to calibrate the area for future operations."

Craven knew there was nothing left to "calibrate." But Craven also knew if Buchanan wanted to stay out, it could mean only one thing. The oceanographer was going to take the Mizar to the spot Craven and his team had pinpointed. Craven went to the admirals and began mixing his rapid-fire logic with pleas. By the time he was finished, he had won two more weeks.

Exactly one week later, Craven received a one-line missive from the survey ship: "Buchanan shaved his beard."

Craven didn’t need any translation. The Scorpion had been found. It was Oct. 29, almost five months to the day after she had been declared missing.

The Mizar found the Scorpion, 11,000 feet under water, within 220 yards of where Craven, his mathematicians, and a group of experts betting for bottles of Scotch had said she would be.

From the book Blind Man’s Bluff: The Untold Story of American Submarine Espionage by Sherry Sontag and Christopher Drew, published by PublicAffairs, a member of the Perseus Books Group. Copyright © 1998 by Sherry Sontag and Christopher Drew.