Basically, on most of the LD trains, you've got about a carload of space that gets filled up at or around WAS/ALX. What that means, in turn, is that a car is running empty NYP-WAS when that car could be used for, at the very least, "backup" capacity for the NEC (where ridership is running into blocks). It doesn't hurt that a few of the trains run at pretty good times for NYP-WAS traffic. Moreover, fill that car NYP-WAS at $150/seat and you're looking at $8400/trip (for 56 seats). Achieve a 20% load factor WAS-NYP on that car at that rate over the course of a year and it's half a million to a train's bottom line; a 50% load factor would give you $1.25m. Moreover, some even shorter sales (NYP-BAL, NYP-WIL, and NYP-PHL) can add even more.
However, this move also smacks of shenanigans: Doing this will likely allow Amtrak to pump up ridership on the Star, Crescent, and Cardinal, all of which are off noticeably for FY2013 as far as May (5.4%, 4.6%, and 3.6% respectively). ...