This consensus held largely unchallenged for decades.
—
## The 2014 Dutch Study
In December 2014, researchers from Delft University of Technology and the Dutch research institute Deltares presented findings at the American Geophysical Union’s annual fall meeting that challenged the conclusion.
Olivier Hoes and Rolf Hut of Delft University of Technology, working with Fedor Baart of Deltares, used a hydraulic simulation program called 3Di to model the specific conditions of San Francisco Bay on the night of June 11-12, 1962. The program incorporated historical tide and current data for that specific date and time.
Their methodology involved simulating fifty virtual boats — representing the escapees’ raft — launched from various points along the Alcatraz shoreline at thirty-minute intervals between 8:00 PM and 4:00 AM. They ran the simulations both with and without paddling, to model the difference between active propulsion and passive drift.
The results were not uniform encouragement. The majority of simulated launches fared poorly; many of the virtual boats were swept into the Pacific, consistent with the FBI’s earlier assessment. The conditions during most of the night did not favor survival.
However, the research identified a window. If the men had launched after 9:00 PM — which the evidence suggests was likely given the lights-out time and the time needed to complete their cell escape and reach the water — and if they paddled northward, the simulations showed that reaching Angel Island was physically possible. The best scenario in their modeling involved a launch around 11:30 PM, with the raft arriving at Horseshoe Bay on the Marin County shoreline rather than at Angel Island.