Even A Stopped Clock's Right Twice as Often As the CIA

They've been wrong before on nuclear development---invariably:

Is Iran closer to testing a nuclear weapon than we think? Last week officials from the International Atomic Energy Agency (IAEA) revealed that samples of machinery taken from a former Iranian research center showed traces of weapons-grade uranium. On Sunday, Tehran rejected out of hand a package of incentives offered by the EU in exchange for Iran halting its enrichment of uranium. But intelligence agencies have assured us that Iran is years from testing a nuclear weapon. The latest publicly known estimate is that they are about a decade off. So nothing to worry about yet, right?

Consider the track record of these estimates. When have they ever been correct? Usually when a country tests a nuclear weapon, the event shocks the world. This was true of India in 1974 and Pakistan in 1998. As well with China—an August 1964 National Intelligence Estimate of the chances of a Chinese nuclear detonation noted that a test site was being prepared at Lop Nor, and would be ready in two months. However, the CIA stated that the Chinese would not have the necessary fissionable material to finish a bomb, so they doubted anything would happen for the rest of the year. Sure enough, two months later, on October 16, 1964, the Chinese successfully tested a nuclear weapon. Something to keep in mind when the “lack of fissionable material” argument comes up with respect to Iran.

The most noteworthy failed atomic forecast was the Soviet case. The CIA’s Office of Reports and Estimates (ORE) was given the task of making this prediction. ORE’s earliest analysis, in 1946, saw the Soviet bomb coming sometime in the 1950-1953 timeframe. Over a series of subsequent reports, the ORE settled on mid-1953 as “the most probable date” for a Soviet nuclear test. This estimate was published August 24, 1949; five days later, the Soviets tested their first a-bomb.