It's a big mistake to think that just because prices dropped last year, they will drop this year or any binary analysis.
US has sophisticated software which looks at yield, not headcount. The # of pax on a flight does not matter. It's the revenue. One pax paying $2,500 in full Y is better than 5 pax, each paying $450 for an el cheapo discount. And so on.
Presuming that OP is not prepared to spring for a refundable/changeable ticket, he can either buy today at the current price or gamble that prices drop (as aviation fuel rises and the economy improves) and wait. If he's wrong, he will pay more.
Or, he can gamble that prices will drop by more than $150, but today and, if the price drops by $150+ cancel and rebook at the cheaper fare. But, to be frank, how much is the fare really going to drop to make this gambit pay?
US is betting that it can sell the seats you think it can't. Guess who has more metrics available?