CIA studies, commissioned in 2012 and 2013, found that arming "rebels" in civil wars usually fails. When such operations do somewhat work, like in Afghanistan against the Soviets, the later blowback is hard to avoid. The Obama administration leaked this story now to reject criticism against its current policies in Syria where it has given up on the Free Syrian Army and wants to create another one.
Political scientists have know for quite a while that arming "rebels" is nearly always a bad policy:
In general, external support for rebels almost always make wars longer, bloodier and harder to resolve (..). Worse, as the University of Maryland’s David Cunningham has shown, Syria had most of the characteristics of the type of civil war in which external support for rebels is least effective.
But if the administration knew that arming rebels was bad policy why did the U.S. start in June 2012 to arm them and why does it continue do so? Why does it still allow Israel and Qatar to do so?
Dan Froomkin suggests that it is all about electoral politics. Not arming the "rebels" ..
.. probably would have been cast by the elite media — not to mention Fox News — as surrender, costing the Democrats another few House and Senate races.
It could also have been a policy driven by the neocon/liberal-interventionist urge to just "do something" – i.e. to achieve some self-satisfaction.
Or the plan was never to win. If the aim was and is the "destruction of the infrastructure, economy and social fabric of Syria" then arming all kinds of insurgents was and is a sane and successful policy.