From Fred Kaplan's latest column:

...at the end of 2009, Obama sent an additional 33,000 troops to Afghanistan, a surge of nearly 50 percent above the 68,000 already there—and that he did so not to go after bin Laden and al-Qaida (a task that could have been handled with far fewer forces) but rather to pursue a counterinsurgency strategy, at least in the cities, particularly in the southern districts. This strategy involved not only killing and capturing bad guys but also helping to reform the Afghan government and providing the people with basic services—in short, nation-building...

...(Obama had) treated the strategy as an experiment; he gave it 18 months to work, and his generals assured him that would be enough time for the Afghan military to take the lead in a majority of the country’s districts, even though some of them knew very well it would take longer. They gambled that enough progress would be made to convince the president to give them more time and more troops. They gambled wrong. After 18 months, almost to the day, Obama announced that he would start pulling out all 33,000 surge troops—and not replace them with any new ones. This too he publicly presented as a victory, and by the same rationale: bin Laden had been killed, al-Qaida decimated, Taliban foot soldiers routed. But the goals of the surge—the goals of the counterinsurgency strategy—had not been accomplished. Obama simply—and wisely—rejected them; the experiment was over; he wasn’t going to double down.

Wise, maybe, but only after the fact. As Kaplan explains, it was Joe Biden who was against this course all along, and Biden's policy is now US policy (related: Biden may be the most influential vice president ever). Policy wonks are usually in favour of experimentation, but it's mildly disturbing to think of Obama using 33,000 troops as part of an elaborate policy 'experiment'. Let's be glad the experiment is over, and sorry that it did not end sooner.