Very interesting. I'll be curious to see the Q4 numbers for the Cardinal.
Looking at the numbers next to one another (sleeper share as a part of the whole):
California Zephyr: 53.5% (!)
Empire Builder: 47.3%
Sunset Limited: 46.2%
Coast Starlight: 45.7%
Capitol Limited: 45.2%
Southwest Chief: 44.8%
Auto Train: 38.9%
Texas Eagle: 34.1%
Lake Shore Limited: 32.0%
City of New Orleans: 31.8%
Silver Meteor: 30.6%
Crescent: 25.8%
Cardinal: 22.3%*
Silver Star: 22.3%*
Auto Train in italics to represent its odd nature.
*If you go out about three decimal places, the Cardinal's sleeper share is slightly larger.
Alright, so what's going on here? The Western trains are all clearly a lot more sleeper-focused. In the case of the Starlight, some of this goes down to the SSL and sleeper turnover in CA. On the others, this is partially a function of the trains' long runs (one does not usually book coach for a two-night trip). It's also a function of bad load factors overall in the middle of the routes (Reno-Grand Junction is an infamously bad segment if I'm not mistaken, likely accounting for a good portion of the Zephyr's situation...let's go back and have a discussion about why Western Pacific was the one begging their way out of the Zephyr's operation, shall we?)...you simply don't have the scads of intermediate destinations that you do in the east.
Interestingly, the Eagle operates as more of an "eastern" train than not...but this is perhaps not surprising: It has the STL cutoff car as well as daylight service in Texas.
As to the eastern trains, most of those are reputed to be rather capacity constrained (either demand-constrained by surging costs or capacity-constrained outright). The Cardinal is the best example of this (add a sleeper and watch it sell out immediately); a comparison between Q2, Q3, and Q4 would be interesting here.
The Star is notable in that it has an unusual amount of coach turnover...it gets a lot of this in Florida (MIA-TPA being a major market for it, and Tampa-based intrastate routings account for at least 20% of the Star's ridership) and another major chunk on the northern end (RVR-north is also a decent-sized market for the train; it's not uncommon to see 30+ pax boarding there, while RGH-WAS actually makes the top ten pairs). The Crescent falls into the same general category due to the ridership offload in Atlanta plus the fact that ATL-NOL and BHM-NOL, both daylight markets, are two of the top three markets for the train...and there's often not much reason to ride sleeper there.
That being said, the Crescent is prone to capacity constraints north of ATL. The Meteor and LSL are also prone to these constraints in spades as well, and I'd expect the sleeper numbers on these trains to jump with the Viewliner IIs (albeit not to the heights of the western trains). These also do a decent amount of "corridor-ish" business (though not on the scale of the Star). The CONO is similar: It overlaps with an IL corridor and has a lot of short-haul business on the southern end, and CHI-MEM/MEM-NOL (IIRC Memphis is the biggest turnover point for the train) are both doable in coach.
Finally, a note on the Cap: While it is an eastern train, it has a mass of through traffic second only to the Auto Train, not to mention a lot of connecting traffic between Chicago and the east coast. It also has relatively low turnover otherwise (PGH is the main such point, and WAS-CHI, WAS-PGH, and PGH-CHI account for over 3/5 of the train's business). This odd ridership pattern increases the sleeper share substantially vis-a-vis the other eastern trains.
