Talk about politicized intelligence. At least that’s what it would be called if the president in office were a Republican.

At the request of the White House, in 2012 or 2013, the CIA did a review of the agency’s long history of supporting insurgencies abroad and found “that it rarely works.” Now, as Seth noted, the result has been leaked to the New York Times. Would it be cynical on my part to imagine that CIA analysts are telling the president what he already thinks–that the U.S. shouldn’t do much to back moderate Syrian rebels?

As a historian, I’m all for studying history. But let’s not cherry-pick historical examples to support a predetermined conclusion. Because based on the Times’s reporting of the CIA study (which needless to say I have not seen) the “dour” conclusions need a lot of qualification.

It’s true that in its early days the CIA failed in supporting would-be rebels in places like Poland, Albania, North Korea, and Tibet. But that’s because they were fighting against totalitarian police states that had great intelligence on U.S. plotting thanks to the information provided by traitors such as Kim Philby. The Bay of Pigs operation was similarly hare-brained and ill-fated.

But there have also been notable successes such as the U.S. support for the mujahideen in Afghanistan in the 1980s–one of the CIA’s biggest coups ever even if there was a lack of follow-up which allowed the Taliban to rise out of the succeeding vacuum of authority. The U.S. had just as much success backing the Northern Alliance to overthrow the Taliban after 9/11 and, earlier, helping the KLA to overthrow Serbian authority in Kosovo, in both cases with American air support. Croatia also succeeded in rolling back a Serbian offensive in the early 1990s with informal American help. Let’s remember too that the U.S.-backed rebels in Libya succeeded in overthrowing Muammar Gaddafi with NATO airpower. As in post-Soviet Afghanistan, there was nothing inevitable about the resulting chaos, which occurred because President Obama failed to support the governmental forces attempting to impose order.

The CIA’s support for the contras in Nicaragua in the 1980s was also successful, contrary to the CIA report and despite the halting nature of the program (due to congressional opposition), because even though the contras didn’t seize power at gunpoint, they pressured the Sandinistas into holding elections, which they lost. U.S. support for anti-Communist rebels in Angola and Mozambique was less successful but at least tied down Cuban and other Soviet bloc forces in defending those regimes. During the Vietnam War, too, the CIA had considerable success supporting anti-Communist fighters in Laos who prevented for a decade a takeover by the Communist Pathet Lao at low cost to the U.S.

The U.S. has had even more success in supporting governments fighting communist insurgencies in countries such as Greece, the Philippines, El Salvador, and Colombia.

So the historical record of U.S.-backed insurgencies (to say nothing of counter-insurgencies) is certainly not one of unalloyed failure. But while it’s good to learn from history it’s also important to understand differences between historical examples and present-day dilemmas. And the situation in Syria today is nothing like the situation the U.S. confronted in the Communist bloc in the early Cold War days. The Free Syrian Army is not fighting a powerful totalitarian regime. It is fighting a multi-front struggle against a weak dictator (Bashar Assad) who has already lost control of two-thirds of his country and against Islamist insurgent groups, the Nusra Front and ISIS, which have filled some of the succeeding vacuum but are a long way removed from the Stalinist or Maoist states in their ability to control their terrain. In such circumstances U.S. backing for the Syrian rebels was–and is–the best available option for the U.S. even though the Free Syrian Amy’s odds of success decline the longer we refuse to provide them with serious backing such as American airpower to impose a no-fly zone and take away Assad’s murderous air force (Which even the CIA study seems to concede would raise the odds of success).

Ultimately responsible policymakers cannot retreat into inaction by citing studies of historical examples where support for insurgencies has failed, while seemingly ignoring contrary examples. The relevant question to ask in Syria or any other hard case is: What is the least bad option? Sure it’s possible that serious support for the moderate rebels would have failed–but what’s the alternative? Actually we’re seeing the alternative today: letting ISIS and Assad run wild, slaughtering tens of thousands of people and destabilizing neighboring countries. Obama made a horrible decision by taking a hands-off attitude toward Syria and he can’t take refuge in a slanted view of the historical record to justify his inaction.