In brief, I said Assad had to contract his realm to a Core Syria that his depleted ground forces could secure; expand his ground forces to reclaim land outside of that core; and engineer his own "awakening" among some of his enemies by exploiting fear of jihadis.
Assad needs to do something that offers his troops hope of victory by giving them an objective within reach. Assad needs to abandon large parts of Syria to the rebels and prepare to rebuild his forces to retake the country. ...
Assad would have to similarly [NOTE: I compared his situation to our position in Iraq] hold his core, rebuild his army's numbers, use air power and ground raids into rebel territory to keep the rebels off balance, and then begin expanding areas of control as he gains the numbers to do so.
Assad has a problem in that he can't count on a foreign patron to supply a surge of trained forces as Iraq's government could count on America in 2007. He also has a problem in that his forces rely on a minority of the population rather than the majority (80-90% Shia and Kurd) that the Iraqi government could rely on.
Assad would need to engineer at least a partial awakening by using divide and conquer diplomacy with Sunnis who fear al Qaeda more than they fear a deal with Assad. Perhaps the Kurds could be won over with promises of autonomy. ...
Assad could still win this fight. But he has to retreat until the correlation of forces can be swung back in his favor. Assad just can't win the way he is fighting, now.
Assad, with a lot of help from Iran and Russia, has achieved much on the first objective of securing western Syria and is working on the second goal of expanding his forces.
Strategypage has one more piece of reclaiming lost ground in the east with the third requirement to win:
Meanwhile the Assads keep lines of communication open with Sunni tribal leaders in eastern Syria. Many of these tribal leaders know that ISIL is likely to burn out sooner rather than later and they want to be ready to quickly switch sides.
Assad didn't give up. He worked the problem. Of course, giving up meant slaughter of his people and exile, poverty, and death for the rulers. So there was that incentive.
Not that Assad must win because he is winning.
The Kurds of Syria will not give up the autonomy that they have gained. Iraqi Kurds are helping Syria's Kurds.
And the west is not locked down even though Assad's forces have denied rebels the ability to control large chunks of the region. Insurgency can still plague the Assad forces if the rebels don't lose heart (and if they get foreign support).
Hezbollah is tiring of casualties and alienating Sunnis in Lebanon for assisting Assad.
And the Iraqis who dominate the Shia foreign legion that Iran sent to help Assad may well want to come home to defend Iraq.
And even if Assad can engineer his own awakening among those eastern tribes tired of al Qaeda rule, Assad still needs to fight and capture the region.
Assad's troops have taken heavy casualties and are tired and demoralized. There has been no rotation home for these troops. Will they be capable of being sent east in sufficient numbers with enough morale to fight to reclaim eastern Syria? Many are militias that are strategically immobile and are really only effective in defending their neighborhoods.
And will the dispatch of sufficient troops east allow still-resisting rebels in the west to regain the initiative there?
Assad certainly has a chance to win. But let's not use the fact that Assad is winning as an excuse to do nothing on the assumption that the end is inevitable. Remember, we once assumed that the fact that Assad was losing as proof that Assad was doomed and so we had to do nothing to hasten his end.