OIG: Opportunities to Enhance Decision-Making for LD Equipment

[WoodyinNYC]

Link of Glory

Of note:

Thanks again for this link. Rich nuggets of info,

in the footnotes.

On page 3,

4 These initiatives are part of a larger plan that the business line developed to reduce

the operating losses on long-distance trains by about $200 million by the end of fiscal year 2018.

That's a helluva plan!

Cut $200 million operating losses within 4 years, by the end of Sept. 2018.
On page 7,

9 Amtrak ... also plans to add a sleeping car to an overnight regional train traveling

to Boston, New York, and Washington.

So Amtrak probably will add a sleeper to this venerable route. Keeping in mind

it was scolded for lacking a "final" plan to allot the new equipment. LOL.

On page 10,

13 This change could save the company about $0.2–$1.3 million on the total order.

The piddling costs savings of rejiggering the terms of the original CAF order for

130 cars suggests that cash flow was not a consideration at all.

It also suggests that the modular innards of the new cars are not costly (except

probably the diners, with kitchen and refrigeration equipment, HVAC, etc.).

Because when 15 bag-dorms became 15 baggage cars, the savings from doing

without the roomette modules was only $200,000 to $1,300,000, or not even

$100,000 saved per car.

++++++++++++++++++++++++++++

​Of course, this trifling OIG report itself might be a reason that Amtrak has

made no move on the option order for 70 more Viewliner II cars. Management

knew this "critique" was in the works, and would be used by haters to try to block

any attempts to improve Amtrak. So best to just lay low for a while longer.

 

[Anderson]

It looks like, absent the disruptions to the Silvers and Crescent, the "southern" trains for Amtrak just don't see a significant effect in the winter (at least, based on 2013...I'll start poking around in some other years to see how much this holds up year-to-year). I seem to recall running some similar numbers a few years back and finding something similar on the Silvers at the time; I'll admit, for me the biggest surprise might well be the relative stability on the CONO. It makes sense on further thought, but it's a surprise on first glance.

That being said, I'm willing to bet that if you could break out airline travel by market you'd see something akin to what is going on with Amtrak...witness the seasonal flights from New York to Florida that Virgin America adds, for example.

...

The question does come up of using a second diner on the Meteor as a table car and/or going 6/3 instead of 5/4 on the Florida trains, but that's a more complicated question.

Florida's population has grown so much over the past 40 years along with tourist travel for spring break, summer and winter seasons, that is not really a surprise that total travel flow to/from Florida has evened out to a large extent over the course of a year. There are still peak direction periods with the snowbirds heading south in October/November and north in March/April that probably offsets what otherwise would be declines in those months, but we don't the north-south breakdowns in the monthly totals. All of which is good news for keeping the Silver equipment fairly busy year round for better returns on equipment costs.
As for the CONO, Mardi Gras takes place in February or early March, so that must offset what would otherwise be a winter season decline in February. The Crescent must get a bump from Mardi Gras travel as well.

There are clearly seasonal changes in domestic airline travel patterns in Jan-Feb with more flights to FL and, for that matter, Hawaii, and less traffic between the smaller market northern cities. There is a lot of airline and airport travel data available, but I'm not inclined to spend much time digging it up.

If an extra long Meteor for the peak of the peak periods with 4+ sleeper cars and extra coach cars needs a table car between the diner and Amfleet II cafe/diner car for spillover crowds, an Amfleet I cafe car would likely be the most cost effective option. Amtrak still has spare Am I cafe cars and the cafe car could be used to store extra refrigerated food and supplies.

Well, and given that Amtrak will likely have an extra batch of diners (26 post-Viewliner II), one could be used for table space and/or food storage. Just because a second dining car is used does not mean that it has to be given "full" staffing in the kitchen.

 

[jis]

Provided a single kitchen is actually bale to serve 24 tables.

 

[Metra Electric Guest]

Regarding the CZ, I don't understand why Amtrak doesn't aggressively market it as the ski train in winter. If I were a skier living near Chi, it couldn't get any better than taking an overnight train to Winter Park and then using the free shuttle bus service to get around. Except maybe for a side trip to the hot springs in Glenwood Springs.

The compartments at the end of the superliner cars that are accessible from the outside used to be called ski lockers. But I haven't heard the term in years.

I would use it! As a kid the trek to Winter Park was train to Denver and bus or ride from friends to Winter Park. The station is is right at the base of the slopes as you leave the tunnel and a package of luggage transfer to lodging and lift ticket would be great, ski right off the platform, how cool would that be? I know the ski train failed, but that was only from Denver.

 

[Anderson]

Provided a single kitchen is actually bale to serve 24 tables.

I doubt they'd be using all 24 tables, or indeed anything super-close to it. Taking the wildest possibility the equipment could offer, a six-sleeper Viewliner train with no space in those six sleepers allocated to crew (so assuming a bag-dorm is used in some fashion) would have no more than 168 sleeper passengers (that assumes 2 per roomette and 2 per bedroom, with 11 roomettes and 3 bedrooms). That's 14 tables on a three-seating schedule. Adding in coach passengers dining as well suggests no more than 16-18 tables would be needed at a full load. Using what is probably a bit more realistic estimate, at full capacity (2/bedroom and 1.5/roomette) you'd have 135 sleeper passengers, or about 34 tables per dining period. This could be accomplished on paper with one diner and three seatings, but you'd need some of the space in the second diner to accommodate coach passengers and/or be more flexible on the seating schedule.

There are three reasons I like the idea of using a "spare" diner for this train instead of trying to pack folks into the cafe. One is that taking up half of one end of the cafe might get a little tight. The second is that you'd have two options as to what to do as far as the OBS situation. One would be to experiment with a PPC-like car for the sleeper passengers (the second diner serving as a lounge in the off hours). The other would be to use the second diner as a "diner-club" as some of the PIPs suggested might work well. The third is that...well, unless Amtrak is going to add a diner to another train, they should easily have the equipment to do this with.

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

Another few thoughts, if I might:

-While the CCC idea was, and to some extent still is, a dubious one...on a few of those trains where sleeper ridership nosedives in the middle of winter the concept might have some merit. I'd be cautious about yanking the SSL on any of the western LD trains, but would it be possible to de-staff the Zephyr's cafe and just run things out of the diner west of Chicago or somehow cut OBS back west of Denver?

-On the SW Chief, it looks to me like ridership holds up as long as the extra sleeper is on the train. Sleeper ridership holds up from May to October unusually well (the train simply doesn't have a falloff in September like the other highly variable trains do), so I'm wondering if the drop in ridership in the fall (which seems to line up with the equipment shift) isn't induced by the equipment cut. The train would likely suffer some in January/February, but that seems to be about the only guaranteed "bad" period.

 

[keelhauled]

Regarding the CZ, I don't understand why Amtrak doesn't aggressively market it as the ski train in winter. If I were a skier living near Chi, it couldn't get any better than taking an overnight train to Winter Park and then using the free shuttle bus service to get around. Except maybe for a side trip to the hot springs in Glenwood Springs.

The compartments at the end of the superliner cars that are accessible from the outside used to be called ski lockers. But I haven't heard the term in years.

I would use it! As a kid the trek to Winter Park was train to Denver and bus or ride from friends to Winter Park. The station is is right at the base of the slopes as you leave the tunnel and a package of luggage transfer to lodging and lift ticket would be great, ski right off the platform, how cool would that be? I know the ski train failed, but that was only from Denver.

Isn't the Amtrak stop somewhere west of Winter Park? I think that's what I found out when planning my trip west this winter. The ski train used to stop at the resort, but I don't think the CZ does. If Amtrak wanted to pursue the ski traffic more, I'm sure they could stop at the resort, but right now it's not quite that convenient.

 

[PRR 60]

Regarding the CZ, I don't understand why Amtrak doesn't aggressively market it as the ski train in winter. If I were a skier living near Chi, it couldn't get any better than taking an overnight train to Winter Park and then using the free shuttle bus service to get around. Except maybe for a side trip to the hot springs in Glenwood Springs.

The compartments at the end of the superliner cars that are accessible from the outside used to be called ski lockers. But I haven't heard the term in years.

I would use it! As a kid the trek to Winter Park was train to Denver and bus or ride from friends to Winter Park. The station is is right at the base of the slopes as you leave the tunnel and a package of luggage transfer to lodging and lift ticket would be great, ski right off the platform, how cool would that be? I know the ski train failed, but that was only from Denver.

Isn't the Amtrak stop somewhere west of Winter Park? I think that's what I found out when planning my trip west this winter. The ski train used to stop at the resort, but I don't think the CZ does. If Amtrak wanted to pursue the ski traffic more, I'm sure they could stop at the resort, but right now it's not quite that convenient.

The Amtrak Winter Park stop is actually in Fraser, about six miles north of Winter Park (railroad west) and the ski slopes.

 

[PaulM]

Isn't the Amtrak stop somewhere west of Winter Park? I think that's what I found out when planning my trip west this winter. The ski train used to stop at the resort, but I don't think the CZ does.

That's why I mentioned the free shuttle bus.

 

[keelhauled]

Isn't the Amtrak stop somewhere west of Winter Park? I think that's what I found out when planning my trip west this winter. The ski train used to stop at the resort, but I don't think the CZ does.

That's why I mentioned the free shuttle bus.

Yeah but the person between our posts referred to the stop being at the base of the resort.

 

[neroden]

2) As to the present Viewliner order, the allocation for Viewliners that I see working would be as follows:

Summer:

-LSL (x3): 3 sleepers NYP, 1 sleeper BOS (12 sleepers total)

-SM (x4): 4 sleepers (16 sleepers total)

-SS (x4): 3 sleepers (12 sleepers total)

-Cre. (x4): 3 sleepers (12 sleepers total)

-Car. (x3*): 2 sleepers (6 sleepers total)

-Cap. (x3): 2 sleepers (6 sleepers total)

This would use 62 sleepers (of 75 theoretically available), allowing for spares and what-have-you. 2 more are technically free with a non-daily Cardinal (hence the asterisk).

Too many.
You want to allow for 1 in the shop for every 5 active at any given time -- at least. That would be 83%. Currently Amtrak keeps 41 Viewliners active out of 50 (82%), counting the extra one on the Cardinal -- but Amtrak is often short on protect cars for the Viewliners.

Keeping the same usage level, you could deploy 61. But that level is already too high and isn't allowing for reliable placement of protect cars or sufficient maintenance. Perhaps the maintenance can be improved by RCM, but the need for protect cars will continue.

You want to have one protect car permanently stationed at each of Miami, New York, Chicago, New Orleans, and Boston (or maybe a second one at NY). 75 - 5 = 70.

Putting 83% of the remainder in service, that means you really only want to commit 58 cars into regular service. And Amtrak still intends to trial a sleeper on #66/67, which requires 2. So 62 + 2 = 64 is about 6 too many.

There are a number of ways to adjust your proposal to allow for this.

A cut-off car plan for the Crescent (1 sleeper to New Orleans x4 + 2 to Atlanta x2 = 8) could free up 4 cars. Only running 1 on the Cap/Pennsy would free up 3. Only running 3 on the Silver Meteor would free up 4. Only running 2 on the NY section of the LSL would free up 3. Some of this might be done in concert with the deployment of bag-dorms.

Or, finally, more maintenance could be done in the winter, allowing for more cars total to be deployed during the summer, with fewer total deployed in January/February. So, you pull 6 cars from the northern trains in the winter and send 'em to maintenance rather than putting them on other trains. The Silvers have *flat* ridership in the winter; they aren't actually spiking in the winter and so don't need *extra* cars. (Yet.)

I still conclude that Amtrak needs yet more Viewliner sleepers.

 

[neroden]

On page 10,

13 This change could save the company about $0.2–$1.3 million on the total order.

The piddling costs savings of rejiggering the terms of the original CAF order for

130 cars suggests that cash flow was not a consideration at all.

Eh... I'm not so sure about that. It certainly doesn't save money overall, but that's quite different from *cash flow*. Delivering the baggage cars before the dining cars might change what month the progress payments must be paid in, even though it doesn't change the total cost of the order *at all*.

This does finally give us a ballpark on the costs of the roomette modules. The savings from replacing 15 bag-dorms with 15 baggage cars is $13333 - $86666 per car. Much of this is going to be the "overhead" of having HVAC and plumbing vs. not having it, so the modules really don't cost very much.

I wonder what it will cost for Amtrak to order the extra bathroom modules to retrofit the Viewliner Is; I'm guessing less than $1 million total for all 50. The labor of reconnecting the plumbing will probably cost more than that.

 

[neroden]

Provided a single kitchen is actually [able] to serve 24 tables.

A kitchen definitely can. It's a question of waitstaff and storage space.

Taking the wildest possibility the equipment could offer, a six-sleeper Viewliner train with no space in those six sleepers allocated to crew (so assuming a bag-dorm is used in some fashion) would have no more than 168 sleeper passengers (that assumes 2 per roomette and 2 per bedroom, with 11 roomettes and 3 bedrooms). That's 14 tables on a three-seating schedule. Adding in coach passengers dining as well suggests no more than 16-18 tables would be needed at a full load.

Underestimate. I hate to keep banging on about this, but half the LSL diner patronage is coach. So at such high loads, you'd need 28 tables at a three-seating schedule.
But of course you'd do more seatings. A five-seating schedule would then only need 17 tables, if people would tolerate it. A more realistic four-seating schedule would need 21 tables. People's preferred dinner times range from 5 PM to 8 PM, and people eat lunch any time from 11 AM to 2 PM, so it might be possible.

Using what is probably a bit more realistic estimate, at full capacity (2/bedroom and 1.5/roomette) you'd have 135 sleeper passengers, or about 34 tables per dining period.

Again, double that to get 68.
This doesn't really affect your conclusions.

 

[neroden]

I know the ski train failed, but that was only from Denver.

The Ski Train was pretty close to a break-even operation, but it couldn't survive

(a) the decision of the original operator to shut it down

(b) the sale of the rolling stock

© the inability of the proposed new operator to reach agreement with Amtrak and UP

As a result it would require capital costs to restart it (buying new equipment, paying for track access, etc.); it would probably be breakeven again.

 

[Paulus]

Meaning no offense, but could you please use the multi quote function instead of separate posts for everyone that you're responding to?

 

[neroden]

-On the SW Chief, it looks to me like ridership holds up as long as the extra sleeper is on the train. Sleeper ridership holds up from May to October unusually well (the train simply doesn't have a falloff in September like the other highly variable trains do), so I'm wondering if the drop in ridership in the fall (which seems to line up with the equipment shift) isn't induced by the equipment cut. The train would likely suffer some in January/February, but that seems to be about the only guaranteed "bad" period.

I think you're probably right. On the SW Chief, extra sleeper should probably be cut after the first weekend in January and restored in late February.
I find myself wondering if there's actually any way to take advantage of the seasonal drop in traffic during January and (early) February. I might imagine that Amtrak could ship lots of vehicles to Beech Grove for maintenance in those weeks... but that would imply extra car maintenance workers or extra overtime for existing car maintenance workers during those months. The overtime would probably be viable, the extra workers probably not.

 

[neroden]

Meaning no offense, but could you please use the multi quote function instead of separate posts for everyone that you're responding to?

It's actually pretty hard to do that when I'm reading 'em and replying to 'em one at a time. Like I just did. I try to keep single topics in one post, though.

 

[Paulus]

Meaning no offense, but could you please use the multi quote function instead of separate posts for everyone that you're responding to?

It's actually pretty hard to do that when I'm reading 'em and replying to 'em one at a time. Like I just did. I try to keep single topics in one post, though.

So read through all the new posts and then go back and reply Or just hit the multi quote button as you finish reading each post. But it can be really frustrating with all of the different posts and it's been considered a major etiquette breach on most boards I've been on to have so many posts right after each other.

 

[neroden]

Really does not work. The editor can't handle the quote-trims well when trying to sort through 6 or 7 page-long text blocks. When I try it I get misattributions and quote-nesting failures... which is definitely worse. You can see some of those in older threads.

Also, this way the moderators can actually split these posts out into the separate threads which they probably belong in... and if you're replying to one of my comments about sleeper allocation, do you really want to have to wade through the process of chopping out the stuff about the Ski Train, the stuff about the diners, the meta, etc...?

 

[Anderson]

Provided a single kitchen is actually [able] to serve 24 tables.

A kitchen definitely can. It's a question of waitstaff and storage space.

Taking the wildest possibility the equipment could offer, a six-sleeper Viewliner train with no space in those six sleepers allocated to crew (so assuming a bag-dorm is used in some fashion) would have no more than 168 sleeper passengers (that assumes 2 per roomette and 2 per bedroom, with 11 roomettes and 3 bedrooms). That's 14 tables on a three-seating schedule. Adding in coach passengers dining as well suggests no more than 16-18 tables would be needed at a full load.

Underestimate. I hate to keep banging on about this, but half the LSL diner patronage is coach. So at such high loads, you'd need 28 tables at a three-seating schedule.
But of course you'd do more seatings. A five-seating schedule would then only need 17 tables, if people would tolerate it. A more realistic four-seating schedule would need 21 tables. People's preferred dinner times range from 5 PM to 8 PM, and people eat lunch any time from 11 AM to 2 PM, so it might be possible.

Using what is probably a bit more realistic estimate, at full capacity (2/bedroom and 1.5/roomette) you'd have 135 sleeper passengers, or about 34 tables per dining period.

Again, double that to get 68.
This doesn't really affect your conclusions.

Several factors you're forgetting:

(1) Breakfast/lunch have higher attendance from coach passengers, partly due to meal cost. Per the LSL PIP, the average coach passenger spent something like $12-13 versus something like $20 for a sleeper passenger's (included) purchase per meal.

(2) We're not adding any coaches to the trains at present, so you'd basically take present diner levels for coach and bolt them onto the sleepers. You're not going to get to 68 tables needed on the Meteor...34 tables needed for coach passengers in the diner (136 seats) would imply approximately 40-60% of coach passengers on the train at any given time getting meals in the diner (the train, at its peak, runs with 295 coach seats)...at which point I would argue that you should simply axe the cafe car and run two diners.

(3) The LSL has a somewhat higher-than-normal share of coach passengers using the diner...likely due to the eastbound train normally only serving breakfast and lunch (on top of a decent slug of Buffalo-New York coach passengers getting added to the mix) if I am not mistaken.*

Basically...the average share of coach passengers using the diner at any given meal does not generally seem to exceed 20% of total coach passengers on board (my best estimate is actually somewhere around 10-15%). This is actually in keeping with the LSL and SW Chief observations: The Chief only has 252 coach seats at any given time (3 coaches at 84 each) while average sleeper patronage runs around 83/train all year, while if I'm not mistaken the LSL has 295 coach seats (5 coaches at 59 each) and averages 51 sleeper pax/train.

A train with a lot more folks in the sleepers is going to have a lower coach share of diner use if the coach section is not extended simply due to relative volume. To use a hypothetical six-sleeper Meteor versus the present three-sleeper Meteor, the present train has 236-295 coach seats and around 60-ish sleeper slots. Doubling the sleeper space (adjusting for the lost roomettes) roughly doubles the sleeper space while not affecting coach space (so now the train goes from 1/6-1/7 sleeper to a little under 1/3 sleeper).

*In general, I believe the share in the diner tends to be in the 35-40% range of traffic (on the WB SW Chief, excluding breakfast into LA, average sleeper traffic in the diner per meal was around 55 while total meals served ran from 75-90, putting the coach share at 25-40%; though it is possible that in some cases coach pax were locked out by traffic load, the diner on the Chief ought to be able to handle at least 120 pax/meal). My experience is that "diner jams" only seem frequent on the single-level trains and a few segments of the Western trains.

 

[afigg]

This does finally give us a ballpark on the costs of the roomette modules. The savings from replacing 15 bag-dorms with 15 baggage cars is $13333 - $86666 per car. Much of this is going to be the "overhead" of having HVAC and plumbing vs. not having it, so the modules really don't cost very much.

I wonder what it will cost for Amtrak to order the extra bathroom modules to retrofit the Viewliner Is; I'm guessing less than $1 million total for all 50. The labor of reconnecting the plumbing will probably cost more than that.

This late in the contract schedule, Amtrak may be paying a penalty for the change order of 15 bag-dorms to 15 baggage cars. A total $ number in the OIG report is not going to tell us much about the cost difference between a bag-dorm and a baggage car.
A proper retrofitting the current Viewliner Is to match the Viewliner II configuration will also require replacing the roomette and bedroom modules. Would not surprise if the complete retrofit/upgrade and overhaul cost per Viewliner I sleeper was on the order of $1 million each. The FY14 budget allocates $7.6 million for Viewliner overhauls with 13 Viewliner overhauls scheduled, so a regular 4 year period overhaul of a Viewliner appears to run around $500K to $600K.

 

[afigg]

2) As to the present Viewliner order, the allocation for Viewliners that I see working would be as follows:

Summer:

-LSL (x3): 3 sleepers NYP, 1 sleeper BOS (12 sleepers total)

-SM (x4): 4 sleepers (16 sleepers total)

-SS (x4): 3 sleepers (12 sleepers total)

-Cre. (x4): 3 sleepers (12 sleepers total)

-Car. (x3*): 2 sleepers (6 sleepers total)

-Cap. (x3): 2 sleepers (6 sleepers total)

This would use 62 sleepers (of 75 theoretically available), allowing for spares and what-have-you. 2 more are technically free with a non-daily Cardinal (hence the asterisk).

Too many.
You want to allow for 1 in the shop for every 5 active at any given time -- at least. That would be 83%. Currently Amtrak keeps 41 Viewliners active out of 50 (82%), counting the extra one on the Cardinal -- but Amtrak is often short on protect cars for the Viewliners.

....

I still conclude that Amtrak needs yet more Viewliner sleepers.

The OIG report lists the April 11 plan was to assign 60 sleeper cars to be in service with 15 cars for "shop count". Figure two consists for a 3 day a week Cardinal and 1 sleeper car each assigned to the overnight #66/#67 Regionals. Then work back to the allocation of the remaining sleepers to the LSL, Silvers, Crescent, Cardinal.
One option to free up 3-4 sleeper cars would be to switch the Silver Meteor operation back to 3 consists, but the Meteor service would need a reasonably reliable OTP to keep delays from cascading into the next train departure.

I agree that Amtrak could use a few more new Viewliner sleepers to provide spare capacity to support all the trains. They should have ordered 30 sleepers instead of 25 when the order was placed with CAF while Amtrak was expecting that more annual federal money was coming. With the order set at 130 cars, likely to be difficult in the current federal funding picture to persuade the Amtrak board (and the FRA in their oversight role) to exercise any of the CAF options, even just for 5 additional sleepers.

 

[neroden]

Basically...the average share of coach passengers using the diner at any given meal does not generally seem to exceed 20% of total coach passengers on board (my best estimate is actually somewhere around 10-15%). This is actually in keeping with the LSL and SW Chief observations: The Chief only has 252 coach seats at any given time (3 coaches at 84 each) while average sleeper patronage runs around 83/train all year, while if I'm not mistaken the LSL has 295 coach seats (5 coaches at 59 each) and averages 51 sleeper pax/train.

Nice piece of calculation there. I would run with a high-end assumption, because you never want to *undersupply* dining service; people who expect it and can't get seated get really mad. (Another note: the percentage of coach passengers is probably higher than the percentage of coach capacity, because coach usually does *not* sell out... at this point. As coach capacity fills up, we will see more coach passengers in the dining car even if the percentage of coach passengers using the dining car stays the same.)
The percent of coach passengers using the dining car probably varies partly based on how far coach passengers are going, since people on short hops are extremely unlikely to use the dining car. I seem to remember that the average distance travelled by the average coach passenger is actually quite variable between trains, with some having much more short-haul traffic than others.

I am not surprised if coach passengers are preferentially eating breakfast (rather than lunch or dinner) in the dining car; the breakfast options in the cafe are *terrible*, while the lunch and dinner options in the cafe are OK.

 

[neroden]

A proper retrofitting the current Viewliner Is to match the Viewliner II configuration will also require replacing the roomette and bedroom modules.

I don't expect this to happen. It would probably be cheaper to simply disconnect and screw shut the toilets in the roomettes, while removing the worst-condition roomettes from service; it would realize most of the same savings, and standardize the fleet in terms of bathrooms.

 

[Paulus]

I wonder how much of an impact adding vending machines with a decent breakfast selection (like breakfast burritos) would make to any coach patronage of dining cars.

 

[neroden]

None. In the minds of many, including myself, there is no such thing as a decent vending machine breakfast.

People are sitting down and getting

(1) fresh-cooked eggs

(2) omelettes

(3) cereal & milk

(4) oatmeal...

These are all deeply resistant to the "prepackaged" model of food sales.