Will Americans ever take sleepers again?

[PaulM]

The pertinent question is, would there be sufficient people using the train to make it pay or to be able to operate at an acceptable deficit?
So to be succesful, Amtrak does not need to grab the majority of the market, and does not need to be worried that there are plent of naysayers out there who would never set foot in a train no matter what.

A good point.

Why would I, and by extension the average traveler, want to ...

Congratulations, you're an outlier.

The term "outlier" suggests some knowledge of statistics. So then why use the meaningless term "average traveler". Travelers (or anyone else) have an average height, age, income, etc. But there is no such thing as an average traveler. Now if he had said "majority" or "significant percentage" of travelers, he would be saying something that we could debate.

But I agree with Cirdan, nowhere near a majority would not be needed to significantly change the situation, at least for an odd ball, oops, I mean outlier like me.

 

[neroden]

I doubt that Amtrak will ever maintain a large fleet of sleeping cars to cover those periods during the calendar year that demand dramatically exceeds to the scheduled sleeping space. Sure, there are times of the year that ALL of the trains sell out their sleeping car space, but there are also many months of the year when these same trains are running half-empty or less.

We (was it Anderson?) analyzed this carefully elsewhere. The result, IIRC:

* The Silver Service has practically constant loads year-round

* The Crescent has near-constant loads after accounting for trackwork closures

* The CONO and Cardinal have relatively small seasonal variation

* The LSL, CL, and Empire Builder have manageable levels of seasonal variation, with reasonably high demand even in February

* The real wild seasonal fluctuations are on the other western trains. Yes, the trains which perform worst financially overall *also* have the worst seasonal fluctuations...

 

[neroden]

I'm not sure how it would take "hours" (as in two or more hours) to get from either Chicago airport to downtown. Both airports have rapid transit service that makes the trip no more than an hour (including wait time),

They've completely eliminated the slow zones on the Blue Line then? It used to routinely be more than an hour.

and from MDW, considerably less.

Won't argue with that. Go ahead, buy the rare Midway-Buffalo direct tickets... if they aren't sold out!
I see Southwest is offering relatively cheap Midway-Buffalo tickets this year. *This has come and gone over the years*. Will it stick around? Maybe it will, maybe it won't. We'll see.

Two of the three New York airports have transit service, and even from LGA, Manhattan is about a 35 minute cab ride.

LGA is generally the *fastest* of the three NY area airports for trips to Manhattan.

My general comment is that this is not 1952.

Yeah, that's my comment too. You seem to be acting like the airplane situation is that of 1952. You're wrong.

 

[Ryan]

I'm not sure how it would take "hours" (as in two or more hours) to get from either Chicago airport to downtown. Both airports have rapid transit service that makes the trip no more than an hour (including wait time),

They've completely eliminated the slow zones on the Blue Line then? It used to routinely be more than an hour.

Didn't you see the video the CTA posted in the other thread? It took like 8 minutes to run the Blue Line from end to end! h34r:

 

[neroden]

(comment fail)

 

[Anderson]

I will agree on the cost point (though I know that flops around quite a bit and has gotten quite a bit more expensive in the last few years); a lot of that does vary based on destination, but the point is taken.

If the airline situation was still that of the 1960s, what you're saying would hold true. However...let's face it, it doesn't. JetBlue and Southwest are the only airlines that still offer any sort of free checked baggage service...and JetBlue is dropping that as far as I can tell. Looking at Google Flights for January (round-trip 8/1 to 15/1, Chicago (all airports) to Orlando (all airports)) the general prices run $155 on Frontier to $207 on Spirit to $260 on American/United. Breaking that down, however, you're going to probably face most of the difference being made up. Assuming a week's vacation in Florida, I figure one carry-on and one checked bag plus a seat assignment:
-On Frontier, that's $45 per person per direction (so $180 total), taking $310 for two up to $490 (and another $12 on top of all that to pick a seat).
-On Spirit, that is $65 per person per direction assuming no discounts (so $260 total); there's an additional chunk that gets applied on top of that for just booking a fare (I wound up with $20/person), which brings the total to $714 before seat assignment fees and the like.
-On American/United you're looking at an extra $100 in fees (so around $620 assuming no surprises).

Now that I've gone on about this, I'd point out that Chicago-Florida is a two-night trip and really not at the core of what is being discussed here. Atlanta-DC, Atlanta-Florida, Atlanta-New Orleans...all of those would fall under this umbrella, but Chicago-Florida is a much longer hike. While I agree that with robust service you'd see Amtrak get a modest amount of ridership, two factors limit it. One is the two-night issue; the other is the fact that Midwest-Florida passengers are competing with New York-Florida passengers on the Silver Meteor. I have little doubt that a routing that ran Chicago-New Orleans-Florida (or indeed Chicago-Atlanta/Birmingham-Florida) would be less simply because of the lack of a flood of NEC traffic to compete with.

 

[Paulus]

Chicago-Florida is rather unlikely to be any cheaper than airplane with fees for a sleeper (breaks down to about 23¢ per mile on the old Floridian timetable with $310 per passenger cost; on CONO+Sunset East it's 18.3¢: average sleeper yield is about 27.2¢). And honestly, there's a problem in that you're trying to justify heavily subsidized routes against profitable competition: How much would the sleeper fare be if they actually had to pick up the slack and make the train profitable to run? The incremental revenue of a sleeper has always, from the numbers I've seen published, been quite low, at best, and such a routing is unlikely to have significant coach ridership, which is where most incremental revenue is on a long distance train.

Edit: Actually, worse than that. Airfare is round trip, Amtrak is one-way. So you'd be looking at sleeper revenues of 9-12¢ per mile; less than current coach yields.

 

[Orie]

Personally, I travel from NYC to Chicago/Cleveland/New Orleans/Savannah quite a bit. In the hypothetical scenario described, I would absolutely take the train to those destination when I could, but that would be dependent on my travel either being for personal reasons, or business happening on a Monday/Friday meaning I could take the train one way on the weekend and fly the other. If I had to go to Chicago on a Tuesday for work, for example, it's just not feasible to take the train either way. Unless I could get away with "working from home (ie the train)", but that's not very likely. For personal travel, if I'm taking a week or so off every year from work for a vacation, I would definitely include a train to New Orleans or Savannah or maybe even as far as Miami as part of it (NOLA- I'm doing in January), and I have a number of friends who want to take similar trips with me. We're all aged 20-30.

But NY-NOLA/Miami is the farthest I would go on a train, unless the trip itself was the sole reason for the trip, and that isn't going to drive any sort of ridership numbers. Now don't get me wrong, I do intend to take those long trips occasionally one day. But most people wouldn't. Even my friends who want to go from NY-NOLA with me think I'm crazy for wanting to take a train to the West Coast. Even then, a full trip on the SWC or the Texas Eagle or Empire Builder is likely to be something I only do once in an extremely long while. I just can't imagine taking a trip from Chicago to LA on a train if my goal was LA and not the train ride itself, the exception being when I retire in 45 years and perhaps have the time to do that.

And just for the sake of talking about the real world and not the hypothetical scenario, I rely on my Amtrak MC to get points to spend for redemption trips. If I can get a sleeper trip relatively cheap then I consider that too, but I would never pay more than $300-500 for a one way trip, and even then only if I can pair the return trip or a future one way with a free redemption trip. It's not feasible to expect everyone to have an Amtrak MC to accumulate reward points.

 

[cirdan]

I doubt that Amtrak will ever maintain a large fleet of sleeping cars to cover those periods during the calendar year that demand dramatically exceeds to the scheduled sleeping space. Sure, there are times of the year that ALL of the trains sell out their sleeping car space, but there are also many months of the year when these same trains are running half-empty or less.

We (was it Anderson?) analyzed this carefully elsewhere. The result, IIRC:

* The Silver Service has practically constant loads year-round

* The Crescent has near-constant loads after accounting for trackwork closures

* The CONO and Cardinal have relatively small seasonal variation

* The LSL, CL, and Empire Builder have manageable levels of seasonal variation, with reasonably high demand even in February

* The real wild seasonal fluctuations are on the other western trains. Yes, the trains which perform worst financially overall *also* have the worst seasonal fluctuations...

An Interesting analysis.

This might be an argument for using single level cars on all services. This way cars could be re-shuffled between trains seasonally, or major maintenance work done between seasons.

Airlines also have seasonal highs, but somehow manage to work around that.

 

[cirdan]

I'm not sure how it would take "hours" (as in two or more hours) to get from either Chicago airport to downtown. Both airports have rapid transit service that makes the trip no more than an hour (including wait time),

They've completely eliminated the slow zones on the Blue Line then? It used to routinely be more than an hour.

Didn't you see the video the CTA posted in the other thread? It took like 8 minutes to run the Blue Line from end to end! h34r:

I have heard some trains don't even stop at the airport station any more but run straight up the escalator

 

[Anderson]

One thing that is being forgotten in all of this is that a train's endpoints do not make a train. With the exception of the Capitol Limited (and obviously the Auto Train) no train has more than about 15% of traffic going from endpoint to endpoint (and the LSL and EB, the trains with the highest endpoint traffic, have three endpoints instead of two). The Cap's high share comes from the fact that it is the only valid daily connection between the Midwest and the Florida trains, plus the lousy hours for a number of intermediate cities, while the Auto Train is sui generis.

The picture is a bit more complex when you have two "endpoint regions": The Silvers have the NEC and Florida, so something does get lost in the multitude of city pairs...NYP-ORL may be only a few percent of business, but add in NWK-ORL, TRE-ORL, PHL-ORL, WIL-ORL, BAL-ORL, and WAS-ORL and it's probably significantly bigger. By the same token, the Silvers only have about 90,000 boardings/discharges at Miami...but by the time one gets out of the Tri-Rail zone, you've tripled or quadrupled that number in an area where nobody is supposed to travel internally. The NEC-Florida is probably a much larger chunk of traffic for the Silvers...but even then, I can point to significant crowds for the Star north out of Richmond on a daily basis (a crowd of 20-25 is pretty common) while I'm rarely the only one getting off the Meteor. Likewise, the Silver Star is known to have a good deal of intrastate traffic in Florida.

So let's consider a Chicago-Florida train in two iterations:
-The CONO extended to Orlando. This was mentioned in the Sunset East restoration report, so it's worth a look. The CONO presently has about 250k passengers per year (up about 25% from when the report was done). Based on the Sunset's traffic on the eastern segment and the effects of daily operation on that, you'd probably add at least 100k for non-through traffic NOL-ORL. Through traffic (counting anyone passing through NOL for any significant distance) would probably add another 50-75k or so, bringing the train to 400k (of which only about 10-15% would be going from the Midwest to Florida, presuming capacity was made available for them).
-A hypothetical restored Floridian would be in the same boat: CHI-IND, CHI-Louisville, CHI-Nashville, CHI-Birmingham...and a bunch of other pairs...come to mind as traffic generators. Again, I'd look for about 15% of traffic to run through.

It's also worth noting that short-hop passengers pay more per mile than end-to-end passengers. For example (all tickets Value fare for the Star and Zephyr and Flexible for the Cap [Value sold out PGH-CHI], using trains 91(19), 29(19) and 5(19)):
-On the Star, a ticket NYP-RGH ($92), one RGH-SAV ($71), and one SAV-MIA ($78) runs about $241. A straight NYP-MIA ticket runs $189 (a difference of 27.5%).
-On the Capitol Limited, a ticket WAS-PGH ($98) and one PGH-CHI ($134) costs $232 versus a WAS-CHI ticket, which costs $189.
-On the California Zephyr, booking tickets CHI-DEN ($146), DEN-GJT ($51), GJT-SLC ($48), and SLC-EMY ($97) totals $342. A through ticket CHI-EMY runs $204 (or about the same as CHI-DEN+CHI-GJT and less than CHI-DEN and SLC-EMY; the difference is 67.6%).

Granted the latter exercise is somewhat academic; the former two, however, are not and represent turning those seats over at major destinations for those trains.

Finally, I'd like to point out that making a straight comparison between sleeper fares and coach airline fares is not entirely fair given the inclusion of meals. Ignoring any other costs that might be attached there for consideration, meals for two CHI-ORL will get you two dinners, two breakfasts, and possibly a lunch per person.

 

[cirdan]

Frankly, if I was going to spend $1700 on travel ticket, I'd go to Tahiti or Seychelles, not Orlando

I've spent more than that on an Amtrak trip before, and don't regret it one bit.

i wasn't going to Orlando, mind you.

 

[jis]

Frankly, if I was going to spend $1700 on travel ticket, I'd go to Tahiti or Seychelles, not Orlando

I've spent more than that on an Amtrak trip before, and don't regret it one bit.

i wasn't going to Orlando, mind you.

Whatever rocks ones boat. I have never spent $1700 on any single train ride ever, and don't intend to start now. OTOH, airline tickets, every year there is at least one $5000+ RT in Business Class half way around the world, net RT mileage around 18,000.

 

[tonys96]

OP asked "Will Americans ever take sleepers again?".

We just got back from a trip from DAL to CUS on the TE, with all rooms in sleeper and Transdorm occupied from DAL to STL, and only two open on to CUS.

Then we took LSL from CUS to NYP. Since we had to walk through all sleepers to get to diner, and through sleepers and three coaches to get to lounge car, we saw all rooms full from TOL to ALbany split, and on NYP section all full yo Rochester.

So, to answer the OP question.........evidently, yes.

 

[rrdude]

I can never take more than a week off work at a time. ...

Some folks are so eager to say, NO! that they don't care what's the question?

...

There is a 6-day or 7-day difference between overnight and a week.....

If you insist on posting irrelevant replies about week-long trips, you are

in the wrong thread.

Then what's the point? Are you going from DC to Chicago (a single overnight that burns a day and a half, which turns into three total days lost) and immediately turning around?

Who is immediately turning around? If you arrive before 10 a.m. and leave after 5 p.m. can't you get in the meetings you need? If not, stay overnight in a hotel to have two days for work.

DC to Chicago is not a good example, as Neroden explains above. These are NOT overnight trips, they are overnight and then some. But that train could give you a fairly good ride Pittsburgh to South Bend, departing midnight, arriving before 8 a.m. (under normal conditions, that is, not with the on-going NS meltdown on this route). The return would be 6:40 p.m. out of Chicago, arriving Pgh at 5 a.m., so dawdle over a nice breakfast and your laptop until the office opens. But what to do instead, get up at 4 a.m. to take a 6 a.m. flight to O'Hare, get to Pgh Intnl Airport, arrive in the Triangle, um, much much later than 5 a.m., probably well after the office is open.

Or let's leave Greeneville, SC, at 11 p.m., or Atlanta at 8:15 p.m., and arrive in D.C. by 10 a.m., having had breakfast in the diner.

Atlanta will give you a non-stop flight to National Airport, but Greenville will probably have you changing planes in ATL or Charlotte, with four or more exhausting hours in airports and airplanes.

Let's do take the sleeper from Chicago at 8 p.m., to arrive in Memphis at 6:30 a.m., then after some BBQ and Beale Street jazz, return the same day on the 10:40 p.m. out of Memphis to arrive in Chicago at 9 a.m. We'd better make our reservations in advance. That segment tends to be full.

At present that aren't enuff trains with city pairs where one-day go-and-return overnights are practical. But there are some. More where it's two day, go overnight train-then overnight hotel in end city-overnight train return.

Continued investment in corridors could add more good overnight connections. A Chicago-Cleveland corridor at 110 mph speeds could cut 3 or 4 hours out of the trip times Chicago-D.C. so that one could work in the future. Of course, it wouldn't be "that one". Passenger use would demand more frequencies, with at least morning, afternoon, and evening departures from each end city, making evening departures with good morning arrivals for more city pairs, like Chicago-Pgh and Toledo-D.C., along the route.

Remember that end-to-end riders are usually 15% of LD passengers. To see where the riders are, you have to look at the midpoint cities like Pgh, Cleveland, Toledo, or on Neroden's train, look at Syracuse, Rochester, Buffalo, Cleveland, and Toledo.

But looking at D.C.-Chicago or NYC-Chicago is almost like looking at NYC-L.A. You'll be missing the 85% of the riders on board going hundreds of miles for 5 hours, or 10 or 12 hours, daylight or overnight, but certainly not going thousands of miles for a week. That end-to-end trip taking days on end hardly exists in the real world, less than 15% of the LD trains' riders

There is however, a HUGE problem "traveling to/from mid-point cities...." OTP. I used to take the CAP 2-3 X a year DC-CHI, primarily bc of the points the OP made. Sworn off for now, due to OTP. I have tried the Silvers for DC-RGH, with HORRIBLE luck. There is plenty of padding in schedule for end-to-end travel, but you take the chance of getting SCREWED and missed meetings if you opt for mid-point cities, on LD trains.

I did DC t Huntington twice too, same result screwed on timing. Now, on business travel, I can only rely on taking Amtrak HOME, where it matters less, if I am 2,3,4,5 hours late. I'm still PISSED OFF at being late, but my home hasn't "packed it's bags and left", yet.

 

[Paulus]

One thing that is being forgotten in all of this is that a train's endpoints do not make a train. With the exception of the Capitol Limited (and obviously the Auto Train) no train has more than about 15% of traffic going from endpoint to endpoint (and the LSL and EB, the trains with the highest endpoint traffic, have three endpoints instead of two). The Cap's high share comes from the fact that it is the only valid daily connection between the Midwest and the Florida trains, plus the lousy hours for a number of intermediate cities, while the Auto Train is sui generis.

The picture is a bit more complex when you have two "endpoint regions": The Silvers have the NEC and Florida, so something does get lost in the multitude of city pairs...NYP-ORL may be only a few percent of business, but add in NWK-ORL, TRE-ORL, PHL-ORL, WIL-ORL, BAL-ORL, and WAS-ORL and it's probably significantly bigger. By the same token, the Silvers only have about 90,000 boardings/discharges at Miami...but by the time one gets out of the Tri-Rail zone, you've tripled or quadrupled that number in an area where nobody is supposed to travel internally. The NEC-Florida is probably a much larger chunk of traffic for the Silvers...but even then, I can point to significant crowds for the Star north out of Richmond on a daily basis (a crowd of 20-25 is pretty common) while I'm rarely the only one getting off the Meteor. Likewise, the Silver Star is known to have a good deal of intrastate traffic in Florida.

...

It's also worth noting that short-hop passengers pay more per mile than end-to-end passengers.

This is all the more reason to keep the trains short enough for high frequencies and good OTP.

So let's consider a Chicago-Florida train in two iterations:

-The CONO extended to Orlando. This was mentioned in the Sunset East restoration report, so it's worth a look. The CONO presently has about 250k passengers per year (up about 25% from when the report was done). Based on the Sunset's traffic on the eastern segment and the effects of daily operation on that, you'd probably add at least 100k for non-through traffic NOL-ORL. Through traffic (counting anyone passing through NOL for any significant distance) would probably add another 50-75k or so, bringing the train to 400k (of which only about 10-15% would be going from the Midwest to Florida, presuming capacity was made available for them).

-A hypothetical restored Floridian would be in the same boat: CHI-IND, CHI-Louisville, CHI-Nashville, CHI-Birmingham...and a bunch of other pairs...come to mind as traffic generators. Again, I'd look for about 15% of traffic to run through.

CONO-East would cost at least another $23 million in running costs, likely closer to $35-40 million. That's not exactly a cost that it's going to be recouping. Nova Floridian will be around $44 million in direct costs, probably around $75 million in allocated expenses. Both of these are going to have significant capital expenses as well. Chicago-Florida just isn't a profitable idea for Amtrak.

Granted the latter exercise is somewhat academic; the former two, however, are not and represent turning those seats over at major destinations for those trains.

Finally, I'd like to point out that making a straight comparison between sleeper fares and coach airline fares is not entirely fair given the inclusion of meals. Ignoring any other costs that might be attached there for consideration, meals for two CHI-ORL will get you two dinners, two breakfasts, and possibly a lunch per person.

While also taking at least another day to get there. Free meals don't exactly offset having to burn another couple days of PTO or reducing the amount of time spent at my destination.

 

[Anderson]

Something else worth a thought: A travel market that involves 8-15 hours en route will attract a different crowd than one that involves 16-24 hours, and when you start climbing up towards 40 hours or more the market shifts again. The Midwest-Florida market is very much in the third category while the NEC-Florida markets are in the first two (WAS-JAX is about 14-16 hours and the Meteor times out beautifully for this market). Only the extreme endpoints stretch over more than about 24 hours (a night and a day for the most part).

Comparing a lot of those markets mentioned earlier in the Rust Belt or along the East Coast with far longer markets involving two nights en route isn't exactly a fair match. I will concede that two-night trips are going to be limited to a relative subset of situations (trips where the train is a key component of the vacation, trips where the traveler has reasons for not flying, and/or trips where the air market is fouled up [either by weather or by issues in a specific market]). As a case in point, the damage to Amtrak in January from the freeze-up in Chicago was limited by the fact that a normally slow week in January saw a decent number of passengers diverted from cancelled airline flights to the LD trains (anecdotally, the SW Chief was packed with them). A lot of trains weren't running, true, but those that were running were slam full.

Swinging back around, I can anecdotally back up both sides of this discussion with my own experience getting into taking Amtrak:
-My first overnight trip was a side-effect of trying to fly into a second-tier airport at the last minute in a peak season. Namely, I looked up PHF/ORF-DAB for right after Christmas, got a rather horrid one-way fare, and decided to check Amtrak on a lark. I would have come out ahead in the sleeper, but I didn't even think to book that. The return leg DLD-RVR (I still have the stub) was $150 right after the New Year, and the ticket was booked pretty much at the last minute. This would have been 2006/7 if I'm not mistaken.

-My second overnight trip was on 66/67. No sleeper available, though I would probably have taken one had there been one. This was the return from a messy Model UN conference at Yale (which was conveniently located a quick cab ride from NHV while William and Mary is literally about three rather short blocks from WBG); to put it plainly, by Saturday evening I had no shot at an award and everyone at the conference was miserable. The only reason to stay was for the partying (which I was known for not partaking in), so my request to leave was granted with full understanding (I also wanted to avoid the van ride home, to be fair, and the train was an efficient means to that end). This was a last-minute booking again; this would have been late 2007.

-My third overnight trip was to Iowa at Christmas in 2008. Here a couple of factors came into play, but the big one was that airline tickets to Des Moines were always notoriously expensive, security had gotten increasingly obnoxious over the previous 7 years, it was winter, and since it was Christmas I had a bit of time to kill. I routed through NYP onto the LSL outbound and onto the Cap back and, I believe, came out slightly ahead on the fare. I lost a day due to weather (Amtrak put me up at the Hyatt for the night), but would have lost at least a day had I tried to fly (there was a nasty blizzard). On the way back, things went smoothly. As a note, after Amtrak did well and the airlines went into meltdown mode I decided to take the train exclusively at Christmas because of the absurd cost of flying (when I checked the next year the r/t cost was over $800 in coach to Des Moines; this time I remember the r/t on the Cap being only about $600).

-Finally, we have my fourth trip, to Chicago for a PoliSci conference in the spring of 2009. This was induced by a string of good experiences mentioned above as well as one or two good daytime trips on top. I either broke even or came out ahead on the costs once taxis, parking, etc. came into the picture (I actually ran the numbers after being bugged by a professor as to why I wasn't flying like everyone else was) and the Cap timed into and out of Chicago nicely (I had a presentation at something like 1030 or 1100, spent a night in Chicago, then caught the train home the next evening and got home Sunday night). The only nail-biter was getting into Washington (I got a late start and basically had to run down the stairs at WAS to make the train).

What do these trips have in common?

(1) All were single-night trips. None of the total trip times ran more than about 36 hours one way. Of them, very probably only the Iowa one would have happened had it been more than one night, and even then it's a close-run thing. Florida might have happened, if only because of the relative insanity of the airfare at the time, but a trip cancellation is also plausible.

(2) In two of the four cases, they were in the face of rather crazy airfares. In a third (New Haven), flying didn't even get a glance (if only due to the quirky situation).

(3) In three of the four cases the booking was last-minute. I think Iowa was done a few weeks out, but the others were all a week or so ahead of time. FWIW, when I went back to do Iowa again the next year the booking check was done about four months out.

=========================

I'm going to say that these trips all happened before I became a diehard fan of taking Amtrak. Not that I didn't already have a rather engrained loathing of what the airlines were becoming (airport security never failed to unimpress me), but previous to this I would have been inclined to drive (and indeed I did drive quite a bit...I think I drove to Florida 3-4 times between 2005 and 2010, and I did a number of other roadtrips in that timeframe as well...when I went to NH to work for McCain, I never considered flying but ruled the train out as impractical for several reasons as well) rather than take the train (I came under a lot less pressure not to drive on long trips throughout college; witness NH in 2008).

 

[tonys96]

Granted the latter exercise is somewhat academic; the former two, however, are not and represent turning those seats over at major destinations for those trains.

Finally, I'd like to point out that making a straight comparison between sleeper fares and coach airline fares is not entirely fair given the inclusion of meals. Ignoring any other costs that might be attached there for consideration, meals for two CHI-ORL will get you two dinners, two breakfasts, and possibly a lunch per person.

While also taking at least another day to get there. Free meals don't exactly offset having to burn another couple days of PTO or reducing the amount of time spent at my destination.

Simple solution. You do not take Amtrak.

Others might, though.

We CHOSE to take Amtrak to New York, and fly back. We consider the two days on the trains to be part of the enjoyment of the trip. We got to visit two separate "destinations" Chicago and New York. Was not PTO "burned", instead it was PTO enjoyed immensely.

BTW...I hated the flight back along with the rigamarole surrounding it. Three hours of my PTO was most certainly burned sitting around LaGuardia. and four hours stuck belted to an itty bitty metal seat with darned near zero legroom was a torture to the body and to the soul.

 

[Paulus]

That they may, but I was pointing out that you can't claim sleeper vs airfare is an unfair comparison because of meals without also counting the effect of extra travel time for the train.

 

[tonys96]

That they may, but I was pointing out that you can't claim sleeper vs airfare is an unfair comparison because of meals without also counting the effect of extra travel time for the train.

True. We choose to enjoy that extra travel time, whenever possible. It is all about the attitude.

Amtrak is a perfect fit for some situations. It is contraindicated in other situations.

 

[Anderson]

That they may, but I was pointing out that you can't claim sleeper vs airfare is an unfair comparison because of meals without also counting the effect of extra travel time for the train.

True. We choose to enjoy that extra travel time, whenever possible. It is all about the attitude.

Amtrak is a perfect fit for some situations. It is contraindicated in other situations.

Agreed. I can actually give an example on this front:

I got stuck, due to the OTP situation, f**ing much of the way to/from the NARP conference (I was able to make something of a vacation of the trip, but there was a limit). I will admit that I enjoyed much of the experience in question, though that was down to flying Virgin First. However, the flight was an inconvenient one insofar as Virgin doesn't go to SLC, so I flew to SFO and caught the next morning's Zephyr. Yes, I spent about a day and a half in transit, minimum...but it was one direct flight and then a train with an overnight in Oakland, not a string of connecting flights. Note that I picked this in no small part because when I went to look at flying, everything heading to SLC either had a connection tactically placed to cut the flight short of meal service (the Houston connections seemed to do this) or involved flying on a third-party "express" flight with nothing but very tight coach seating (which was the case in Denver).

Again speaking personally, a lot of why Amtrak won out over flying is that I simply got used to having legroom and whatnot. Something is screwed up when an airline's "first class" product is roughly on par with Amtrak's long-haul coach. The airlines that ran passenger rail out back in the 50s and 60s had baseline seating in the mid-30s in terms of seat pitch and included a lot more amenities in many respects (as well as less hassles). In many regards I strongly suspect that, caveats about pricing notwithstanding, the airlines of today wouldn't have had nearly such an easy time beating the railroads, particularly on the not-super-long routes. NYC-LAX is one thing, but again I turn back to the shorter routes and that's where Amtrak is at least contraindicated a lot less.

 

[Anderson]

Giving this a bit more thought:
-For business trips in that 10-14 hour range, if a sleeper option is available and the pairing allows a departure and arrival outside of business hours, I believe that Amtrak is at least not contraindicated. The indication is not overwhelming and it can be offset by a sufficiently cheap and convenient flight...or supported by a badly inconvenient flight routing with known delays, etc. Below 12 hours, something like a 1900-0700 option becomes available, increasingly negating the chances of anything but a major delay interfering. With a genuinely early arrival (again, 0700 comes to mind) the odds of missing a morning meeting start lining up with the chances of a massive tarmac delay...generally until recently, endpoint arrival within two hours of scheduled arrival was pretty high.
-Above around 14 hours, business travel is going to fall off with certain oddball exceptions (usually down to bad air service in some fashion). At this point, leisure travel is still going to be a component, particularly if a good service could be set up at the destination (i.e. baggage transfers to hotels and shuttles to attractions; yes, I'm thinking Disney's Orlando-area operations, but I'm also thinking cruise lines as well). This is probably viable up to somewhere around 20-24 hours, and is particularly strong if one is likely to need a decent amount of baggage.
-Beyond somewhere in the range of 20-24 hours, passenger rail will only prevail as a preferred option if the airline market is royally screwed up. This happens in some cases, but those are somewhat isolated.

Note that in all cases I am not presuming that rail will dominate travel selections. I'm thinking of rail holding an air-rail market share of 10-25% (a good portion of which would actually be highway diversions)...which in many cases would still generate shattering amounts of traffic. Per the DOT's Consumer Airfare Report for Q4 of 2013, Washington-Orlando was averaging 4,237 passengers per day. Philly-Orlando was generating 2,671. New York was generating 8,850. Smaller markets in the general area (Harrisburg, Richmond, Norfolk) and ones that attach (Hartford) generate a decent additional flow. So using Orlando, 10% penetration of passenger rail into the NEC-Orlando markets south of NYP alone would generate 1575 passengers per day (or 5.75 million per year). Assuming 25% sleeper penetration among that market (or 2.5% of the total travel market), you're looking at 1.4 million sleeping car passengers (probably enough to, on its own, support at least a half-dozen daily trains in each direction). Note that this doesn't assume any travel to Fort Lauderdale or Miami.* Even a lousy 2% penetration by rail would kick out 1.15m passengers to/from Orlando (and 25% sleeper penetration on this, or about 0.5% of the whole market, would generate 288k sleeper passengers on this run).

So the answer, ultimately, is that you don't need massive penetration to achieve a very successful situation with "Americans taking sleepers".

I'll run some other numbers (I'm going to try and tally up full-year numbers for a bunch of pairs) and see what spits out, but the penetration levels needed to support a massive, competent, and flexible rail system compared to what we have now are surprisingly small in a lot of cases. With this in mind, I do feel the need to make a quick note: The population of a lot of the areas in question (especially Florida) is far higher today than it was in the mid-60s when the "wheels came off".

*Those two can be accessed by rail a lot faster than Amtrak currently does...IIRC, Amtrak loses about 3-5 hours over what the FEC line should be able to allow (9:07-10:50 vs. 6-ish hours). The Meteor does NYP-JAX in 18:08, and that should drop to about 17:30 with the SEHSR improvements in Virginia should those come to pass. JAX-ORL is about a 3:00 trip, while JAX-MIA should be something in the 6:00 range, so the markets would be fairly similar.

 

[Anderson]

Ok, I promised I would run a bunch of numbers, and here they are. As a note, I used Newark as a base for checking seasonal load variation, and Q3 was the closest to the average on daily traffic to Orlando. I therefore went with Q3/2013 for my baseline report. I defined two general segments for this, which I am using as imperfect measurements of traffic:
NEC Core: EWR, JFK, LGA for New York; PHL for Philly; and BWI, DCA, and IAD for Washington
NEC Rim: RIC, MCO, PHF for Virginia; ALB, BDL, and ISP for north-of-New York, and BOS, PVD, and MHT for Boston

Daily Air Traffic per market
NEC Core-Jacksonville: 4,045/day
NEC Rim-Jacksonville: 744/day

NEC Core-Orlando: 11,753/day
NEC Rim-Orlando: 5,139/day

NEC Core-Tampa: 6,842/day
NEC Rim-Tampa: 3,031/day

NEC Core-Fort Lauderdale: 10,739/day
NEC Rim-Fort Lauderdale: 4,045/day

NEC Core-Palm Beach: 3,371/day
NEC Rim-Palm Beach: 1,277/day

NEC Core-Miami: 6,781/day
NEC Rim-Miami: 1,096/day

Total, NEC Core-Florida: 43,531/day (15,888k/yr)
Total, NEC Rim-Florida: 15,332/day (5,596k/yr)

Basically, we have an NEC Core market of 15.9m/yr and a surrounding market of somewhere around 5.5m/yr. That is a /ton/ of passengers, and Amtrak currently holds about 2% of the market here (largely on the Auto Train). Let's assume that with a robust system (including multiple-daily trains, slightly better times, solid timekeeping, and direct service on the FEC down to Palm Beach, Fort Lauderdale, and Miami) Amtrak could get to 10% of the core market and 2% of the rim market, of which 25% (2.5% of the core market, 0.5% of the rim market) would take a sleeper:
Core: 1,582,311
Rim: 111,923
Total: 1,694,235/yr
Sleeper: 423,558/yr

For the record, the Meteor and Star currently run about 750k/yr, including local traffic in Florida and elsewhere. Note that sleeper traffic on these trains is often in the 15-20% range, but there's also a lot of shorter-hop traffic that we're ignoring here (Florida intrastate, traffic to/from NC, SC, and GA, and traffic between VA and the Northeast). I've also excluded a bunch of airports (Harrisburg, Atlantic City, etc.) from consideration in the rim, though it is quite probable that you could double the rim's traffic on folks who would connect in from places like those.

1.7 million passengers per year would therefore probably support at least six daily trains; moreover, the sleeper load would itself probably end up supporting multiple trains on its own (that traffic would support two daily trains of 12 Viewliner sleepers and have leftovers for the other trains even if there was absolutely no local traffic carried between Richmond and Jacksonville). More realistically, each train running with six coaches for through traffic would be able to take about 175-200k passengers per year (I'm not going to assume 100% load factors, though the theoretical capacity of such a train is 258,420/yr...yes, I counted both directions), so you would likely need seven trains two more in-season). Each train would, in turn, need 3-5 sleepers to deal with the load...though in all honesty, you'd probably end up with 5-6 sleeper trains with 5-6 sleepers apiece, plus 2-4 mostly daytime trains that don't have them but which take up a lot of intermediate-destination business.

This penetration level is not unrealistic in my opinion, given the equipment and the slots. You would probably see almost as much traffic added again on local-ish traffic (these trains would get a respectable share of the traffic between VA and the NEC, for example, as well as make a dent in the NC-NEC traffic; if FEC allowed it, the Jacksonville-Miami trains would likely pick up a solid share of traffic as well).

 

[Paulus]

How is that level of penetration "not unrealistic" given the average one-way airfares?

DC-Orlando: $149

Philadelphia-Orlando: $171

New York-Orlando: $175

Chicago-Orlando: $188

Albany-Orlando: $204

Boston-Orlando: $145

All numbers Q3 2013

 

[neroden]

There is however, a HUGE problem "traveling to/from mid-point cities...." OTP.

Certainly. The bane of Amtrak. And caused almost entirely by two things: lawbreaking dispatching practices at the Class I "freight" railroads (coded as "freight interference" in Amtrak's reports), and derelict maintenance at said railroads (coded as "slow orders" in Amtrak's reports).

This is all the more reason to keep the trains short enough for high frequencies and good OTP.

There is no evidence that shorter routes will have better OTP. Look at the Hoosier State!
The evidence is that routes have better OTP if they are over tracks owned by Amtrak, state governments, commuter agencies, or shortlines with state support, and worse OTP if they are over Class I freight tracks.

This is why I keep going on about the importance of owning the tracks. It is probably unrealistic to imagine a passenger-operated-owned corridor across the Rockies and the Prairies. However, a passenger-operator-owned corridor from New York to Chicago is entirely realistic.

I'm going to say that these trips all happened before I became a diehard fan of taking Amtrak. Not that I didn't already have a rather engrained loathing of what the airlines were becoming (airport security never failed to unimpress me), but previous to this I would have been inclined to drive (and indeed I did drive quite a bit...

Again, I believe Amtrak is primarily competing with driving, not flying. There are already a huge number of people who are strongly disinclined to fly, for myriad reasons. Most will drive. (There seem to be far fewer who are strongly disinclined to drive.)

Thinking about it, I don't think my family has ever flown from Ithaca (or Syracuse) to Chicago when we were actually going to Chicago (though we've done so to catch connecting flights). However, we drove Ithaca to Chicago a *lot*.

The sleepers on the LSL, when it's running close to on time, are quite competitive with driving, unless you enjoy the No Sleep Maniac style of driving. They are certainly at the luxury end of the market compared to the drive.

Again speaking personally, a lot of why Amtrak won out over flying is that I simply got used to having legroom and whatnot. Something is screwed up when an airline's "first class" product is roughly on par with Amtrak's long-haul coach. The airlines that ran passenger rail out back in the 50s and 60s had baseline seating in the mid-30s in terms of seat pitch and included a lot more amenities in many respects (as well as less hassles).

...and arguably at the luxury end compared to the current airplanes, too.
The luxury end of the market is a real market. Not everyone wants the cheapest, crappiest experience available.

Is any of this relevant to you, Paulus? Maybe not -- you live in California! People have made solid cases for sleeping car service in the extended NEC, and in the "Heartland" triangle from there to Chicago, and described sellout conditions running down the Mississippi River corridor on the Texas Eagle. In these areas, if the tracks are made faster, making more cities suitable for daytime corridor routes, it'll just mean that the sweet spot for sleeping car service is a slightly more distant set of city pairs. I don't think I'll see enough speed-ups in my lifetime to make everything out here close enough together to make sleeper service irrelevant -- you'd have to get NY-Chicago below 9 hours, for example, which seems highly unlikely.

.. all of which has absolutely nothing to do with California. There's simply nothing in the sleeper sweet spot in California, or at least nothing with enough population.

Someone's going to bring up Europe. In Europe, the sleeper services are having a fundamentally stupid problem: the arrival of high-speed rail means that the sweet spots for sleeper service are now nearly all international. The administrative situation makes new international routes hard to start up, and international routes (particularly crossing language barriers) always have less travel than you'd expect from the city sizes alone. By contrast, in the centralized Russian Empire and Chinese Empire, the sleeping car services are alive and well.

An Interesting analysis.

This might be an argument for using single level cars on all services. This way cars could be re-shuffled between trains seasonally, or major maintenance work done between seasons.

Unfortunately, all the routes have pretty much exactly the same strong months and weak months, so shuffling cars between trains isn't that effective.
The idea of doing maintenance work in the winter and pushing the cars into intensive service in summer is an interesting one. But Amtrak could do that right now with the Superliners (which run on all the most seasonally-variable routes).