Oh, I absolutely agree with you! However, I cannot fathom calling a 50%-60% load factor, over time, (at very best) anything but "struggling to fill trains"
You're just wrong. 50%-60% load factor over the course of a long run is basically full for a train.
This is how it works, using a stylized example. Suppose you have a train which stops at a bunch of cities, crosses a mountain pass, and then stops at a bunch more cities. The train is empty at one end of the run. It accumulates passengers as it stops at intermediate stops. When it crosses the mountain pass, the train is completely jam-packed full, standing-room-only. Then the train starts draining passengers out as it heads towards the other end of the run.
When you average the load factor across the length of the run, it ends up being 50%-60% load factor. But the train is completely full across the mountain pass. And the travel demand is "across the pass", so anyone who wants to cross the pass -- regardless of what city they start or finish at -- sees a sold-out train.
Some trains have more favorable geographical distribution of passenger demand (and they can get higher load factors), but the situation I just described is *typical*. The same is true with roads, incidentally.
Since it doesn't cost anything significant to drag along empty coaches behind a train, you size the length of the train according to the demand across the mountain pass, and just haul the additional cars empty the rest of the way. This "lowers your load factor" but it doesn't *cost* anything significant.
Airlines don't do this because nearly all their flights are point-to-point, because it's expensive for an airplane to stop. (It's relatively cheap for a train to stop en route.) So airplanes end up with people taking transfers to smaller and smaller puddle-jumpers as they head further away from "the mountain crossing". The train might as well keep going straight through.
This causes overall train load factors to appear lower than for airplanes, but it doesn't mean anything economically. (Well, it means discount tickets should be offered for trips which stay out of the peak segment, which "don't cross the mountain pass".)
The interesting number is the load factor at the *peak occupancy part of the route*, which is hard to measure from Amtrak statistics. If that's low, then you do have an issue. But for Amtrak, on the vast majority of trains I've taken end-to-end, there's been one route segment where the train typically fills up -- showing that the trains are pretty much full.
-- For Empire Service, it's the final New York City approach.
-- For Pacific Surfliner, it's Oceanside, interestingly. (Because people who aren't crossing Oceanside take Metrolink or Coaster.)
-- For Empire Builder, it's MSP-CHI.
-- For the Lake Shore Limited, it's South Bend - Elkhart and again on the final New York City approach; having two separate peak segments helps the LSL. It's pretty full overnight too. It has good characteristics.
-- For the NE Regionals, it's New York City to Metropark (unsurprisingly).
-- For the California Zephyr, it's actually Denver to Winter Park. (But the whole Denver-Chicago run is pretty crowded too. It *empties out* west of Grand Junction and doesn't start filling up again until Reno.)
-- For the Southwest Chief, it's Galesburg to Kansas City. (Because there are other trains Galesburg - Chicago; this is mostly Chicago-Kansas City traffic.)
-- I couldn't actually tell for the Capitol Limited since I basically slept through the entire trip, so didn't see the other passengers, but maybe it's an exception.
Consider yourself educated about this basic characteristic of train service.
It would be nice to improve the load factor by filling up the empty seats on the emptiest section of the trip, but think about how to get more people to ride from Niagara Falls to Buffalo, or from Grand Forks to Spokane, and you start seeing that it's a hard problem. Many routes have a "big draw" on only one end of the line. On Empire Service, for instance, nearly everyone is travelling to or from New York City, so the Niagara Falls end is inevitably half-empty. You try to put a "big draw" on both ends of the line, and you get a New York - Chicago train like the LSL -- but then it's emptier in the *middle*. The NEC is lucky to have a lot of different major cities fairly evenly spaced, but it's still most crowded from NY to Philadelphia.