The best predictor for future performance is the past record. So let's have a look:
late Feb or first week of March (Bobert predicting the start of the Iraq war on 7th of February)

Two weeks off a few weeks before the event. If that is the usual standard, we can expect Bobert to be two months off the actual date here, so I go the Bobert way, adjust for error and say first week of December.

Sorry that this prediction doesn't fit in a political agenda.

(On a serious note, I don't consider at this moment the USArmy or CIA to be able to keep their mouths shut about a capture for longer than two weeks. Not for the lack of ill will but for the lack of ability.)

Wolfgang (grin)