Excel comes to mind.
Lake Shore Limited segment appears very expensive?
- Thread starter user 6862
- Start date
Help Support Amtrak Unlimited Discussion Forum:
- Status
I was curious about Amtrak's pricing practice....
I am aware of their 'bucket' system on each segment, as is being discussed in this thread.
But I am wondering if they ever consider giving some consideration to a passenger's full itinerary... In this case, the passenger is traveling from SYR to CHI to EMY. Would they ever say, drop down each segment one bucket, to give incentive to someone going almost coast-to-coast?
I am aware of their 'bucket' system on each segment, as is being discussed in this thread.
But I am wondering if they ever consider giving some consideration to a passenger's full itinerary... In this case, the passenger is traveling from SYR to CHI to EMY. Would they ever say, drop down each segment one bucket, to give incentive to someone going almost coast-to-coast?
jis
Permanent Way Inspector
Staff member
Administator
Moderator
AU Supporting Member
Gathering Team Member
Not clear at all why Amtrak should incentivize coast to coast travel by losing revenue. Would be relevant if they had problem filling a train like the LSL without the longer distance passengers who pay a lower per mile rate anyway.
One thing I've wondered about is how much, if any, the summaries provided in Post #22 would change if fares had been checked each of the 41 days prior to March 2015 - instead of just 7 times. Had that been done, instead of 552 opportunities to detect a fare change there would have been 41 X 92 = 3,772 of them. So only about 15% of the possibles were checked. Oh well, maybe next time - if there ever is a next time.
I studied the Empire Builder simply because it's the one we take to visit our sons in Seattle and Anchorage. Should I ever get bored enough to do this sort of thing again . . . .
Q: Anybody have any suggestions as to which train to study? One that might possibly provide data more typical of long distance trains?
I studied the Empire Builder simply because it's the one we take to visit our sons in Seattle and Anchorage. Should I ever get bored enough to do this sort of thing again . . . .
Q: Anybody have any suggestions as to which train to study? One that might possibly provide data more typical of long distance trains?
Ryan
Court Jester
One of these days I'll have sufficient time to write an Amsnag scraper that pulls data daily and does automated analysis.
The only way I can think of to quantify "popularity" is by how full of passengers a train is. That would be it's seat occupancy rate. And judging from the seat occupancy rate of the LSL, I don't think "ultra-popular" is a good description. That's based on the derivations shown in the right hand column of the chart below. Trains below the average of 0.59 for all long distance trains are indicated by the red dots:
Note that the division operation eliminates the Fully Allocated Contribution/(Loss) and Mile terms. And while the quantities shown as Totals below columns J and K seem a bit peculiar (way to low), they do equal 0.59 when divided - the same as the average of the 15 Occupancy Rate figures in the last column. Perhaps the LSL is ultra-popular with those who spend much of their lives riding around on trains, but it appears there aren't enough of them to get its occupancy rate above the fair-to-middlin' range. Based on its 0.58 rate, I'd have to describe its popularity as "about average".
Last edited by a moderator:
Dang, sure wish I knew how to do stuff like that! Only scrapers I've got are for smoothing wood, carving metal, removing paint and getting ice off of windshields. :blink:
Last edited by a moderator:
I see what you're saying, if you only considered revenue on the LSL. But if you looked at the entire system, perhaps it would be advantageous to give up a bit of revenue on the LSL to gain more total revenue by attracting someone wanting to travel much further....?
As another example, I recall when they used to limit seats available between Denver and Glenwood Springs on weekends on the CZ, so that there would be sufficient space for travelers going much further. As travel dates got closer, they would sometimes open up this segment, if long distance traveler's had not filled demand. So they initially turned away sales in hopes of getting more thru travelers, and avoiding having empty seats beyond this segment.
Not sure if they still do this today....
Yield management is practiced in some form by almost everyone in the transportation and hospitality business. It can be very complex and difficult to follow at times, but it is a key component to how they operate. For the passenger/customer, confusing and/or frustrating at times, but not going away.
neroden
Engineer
It is a very good description.
I suggest you look at the seat occupancy rates -- load factors -- on some of the state-sponsored services.
Absent a very high proportion of end to end traffic, you generally don't get load factors higher than 60%, period. Auto Train is aberrant due to having only two stations.
So 58% is a very high load factor. It's notable how high the numbers are for some of the other trains in the long-distance division. Some of them could also accurately be described as ultra-popular. Most of them, however, are achieving their high load factors by having lower ticket yields.
There is, however, one fundamental problem with using load factor: it doesn't keep track of how many seats are running. The LSL regularly runs six coaches and three sleepers, more seats than any other single-level train.
The correct metric for popularity is arguably simple total ridership. From FY2015:
Coast Starlight 455845
Empire Builder 438376
Silver Star 383347
California Zephyr 375342
Southwest Chief 367267
Lake Shore Limited 356898
Silver Meteor 346097
It is worth noting that FY2015 contained an enormous disruption to LSL service for the first six months of the fiscal year, thank you Norfolk Southern. If we look at FY2014, the LSL is ahead of the Chief and ahead of the Zephyr.
On Amtrak, these ridership numbers are only exceeded by some of the state-sponsored corridors with at least three frequencies each way each day (and the NEC with even more frequencies).
I think it is fair to say that in addition to the LSL, the Silver Star, Coast Starlight, and Empire Builder are also ultra-popular.
Last edited by a moderator:
Ryan
Court Jester
On capacity constrained trains like the LSL, total ridership isn't the best metric.
The best way to do it is probably some combination of load factor combined with average trip length (as a proportion of total route length). Someone smarter than me should play with the math and see what falls out.
The best way to do it is probably some combination of load factor combined with average trip length (as a proportion of total route length). Someone smarter than me should play with the math and see what falls out.
I agree that total ridership is not be best metric. But ridership is nowhere to be found on that chart in Post #32. Ridership is the total number of passengers that rode a train in a given period of time. As an example, NARP data for 2014 (unsure if it's for the FY or the CY) shows ridership for the LSL as being 367,195 - 367,195 passengers. The figure in the right hand column for the LSL is 0.58 - so that's obviously not 0.58 people or ridership.
Those figures in the right hand column ("K ÷ J,Passenger/Seat,[Occupancy Rate]) are therefore not ridership or total ridership. They are, instead, the portion of time any seat was occupied Those figures are the Occupancy Rate (OR) for each of the trains in decimal form. If all seats were always occupied (there were never any empty seats anywhere along a trains route) the OR would be 1.00. The 0.58 OR for the LSL shows that any given seat on the LSL was occupied 0.58 of 1.00 (the whole time) or 58% of the time.
You said the LSL is a "capacity constrained" train as if to set it apart from some other group of trains. That other group of trains, by default, must then be those with capacities that are not constrained. If a train is not capacity constrained it must necessarily be a train with unconstrained capacity. Does that mean that a train with unconstrained capacity can handle an unlimited (infinite) number of passengers? Have an unlimited (infinite) number of seats? Require an unlimited (infinite) number of passenger cars to contain all those seats? If so, the occupancy rate of any such train would always be 0.00 - zero. That's because the limit of the quotient of any finite quantity (people in the world) divided by another - as that other approaches infinity - approaches zero.
Besides all that, any train with an unlimited (infinite) number of passenger cars and would probably be too long for any existing passing siding. I'm sure host railroads would never put up with such a thing.
Please take the time time to study and comprehend what was done with Columns J & K and the meaning of their quotient. However, if you would like to use some other metric (or even non-metric) to judge a trains popularity, please feel free to do so. I'll stick with occupancy rate.
Last edited by a moderator:
jis
Permanent Way Inspector
Staff member
Administator
Moderator
AU Supporting Member
Gathering Team Member
Why is it necessary to setup the strawman of an unlimited capacity train? Ryan was saying a very simple thing. It really does not take this strawman to argue about it. If a train in any part of its run is running to capacity thus making it impossible for additional people to be transported in that segment then the train is capacity constrained, even if it may be running empty elsewhere. Not just the LSL, but many of the single level LDs have that problem in some specific segments, typically either just out of New York or Albany or just out of Washington southbound and just into Washington northbound.. That fact is impossible to capture in an overall rolled up occupancy rate number for the entire run. You'd need separate OR numbers for each segment to capture such information. That is the danger of depending too much on gross rolled up statistics to analyze something that actually consists of many parts with vastly different characteristics.Let us just accept that fact and carry on.
I have no problem with you or anyone sticking to OR, but will happily discount any claims of popularity of the train in one specific segment or another based on a gross rolled up number for its entire run. just because that is the only number one can calculate based on numbers ready at hand should not be a reason to claim more than the limited thing that it actually says.
I have no problem with you or anyone sticking to OR, but will happily discount any claims of popularity of the train in one specific segment or another based on a gross rolled up number for its entire run. just because that is the only number one can calculate based on numbers ready at hand should not be a reason to claim more than the limited thing that it actually says.
CoachSlumber
Lead Service Attendant
- Joined
- May 9, 2013
- Messages
- 266
Gosh, at some point won't they finally restore a second route Chicago-New York, with a sane Chicago departure/NY arrival time?Believe it or not, that's middle of the road pricing. It's a quite popular train, so supply and demand is hard at work.
image.png
He may have been saying a very simple thing, but I simply didn't understand what he meant by "constrained". Now that you've explained it to me I can see that it would be impossible to tell (even using NARP data) at which points along a trains route it became constrained (had no more seats for passengers). But even if a train had no more room for additional passengers (OR =1.00) there's no way of knowing how many were "left behind on the platform" or wanted to ride but couldn't. And if that needs to be included in any metric equating to popularity, then there's no way to compare the actual popularity of trains.
In essence, it would seem a train is popular simply when it's said to be popular.
neroden
Engineer
I realize I didn't fully explain the reason why trains practically never have load factors above 60%.
Lake Shore Limited: Chicago-Buffalo-NY.
Train is typically full from Chicago-Buffalo.
More people want to travel Chicago-(Buffalo,Rochester, Syracuse,Albany, NYC), but can't because train is full Chicago-Buffalo.
Train is rarely full Buffalo-NY. (In this case NY-Albany,Syracuse,Rochester,Buffalo traffic is mostly distributed onto the Empire Service.)
This is a pretty typical dynamic. For a train with many stops, you can't get the load factor up too high because the train is always stronger at one end than the other. (In the case of most of the so-called "long-distance" trains, it's the Chicago end.) 60% is an empirically discovered number which is roughly as high as you can get the load factor on a typical route with lots of stops. You can get the load factor higher by having fewer stops (Auto Train, or even Capitol Limited which has no stops with significant population between DC and Pittsburgh) but that is generally actually bad for the train's overall ridership and revenue.
The Texas Eagle has an exceptionally high load factor considering its typical many-stop route.
Load factor is mainly useful for figuring out what equipment to assign to the train. If the load factor is very high, add more cars. If it's very low, run fewer cars. You may want to figure out how to put cut-off cars at some intermediate point. (If your load factor indicates that you should run significantly less than one car, of course, then you have a route which doesn't have enough ridership to run a train.)
Total ridership is probably the best metric for popularity (uh, by definition of "popularity"). You could argue however that you can *buy* popularity with cheap tickets, so for an underlying "base popularity" you probably want to look at some combination of ridership and ticket yield. The LSL, Starlight, Empire Builder, and Silver Service all run long trains and fill them up on the peak segments, and do so with high ticket prices. They are ultra-popular... compared to, for instance, most of the state-supported corridors, which run short trains, don't come anywhere near filling them up, and do so with lower ticket prices.
Lake Shore Limited: Chicago-Buffalo-NY.
Train is typically full from Chicago-Buffalo.
More people want to travel Chicago-(Buffalo,Rochester, Syracuse,Albany, NYC), but can't because train is full Chicago-Buffalo.
Train is rarely full Buffalo-NY. (In this case NY-Albany,Syracuse,Rochester,Buffalo traffic is mostly distributed onto the Empire Service.)
This is a pretty typical dynamic. For a train with many stops, you can't get the load factor up too high because the train is always stronger at one end than the other. (In the case of most of the so-called "long-distance" trains, it's the Chicago end.) 60% is an empirically discovered number which is roughly as high as you can get the load factor on a typical route with lots of stops. You can get the load factor higher by having fewer stops (Auto Train, or even Capitol Limited which has no stops with significant population between DC and Pittsburgh) but that is generally actually bad for the train's overall ridership and revenue.
The Texas Eagle has an exceptionally high load factor considering its typical many-stop route.
Load factor is mainly useful for figuring out what equipment to assign to the train. If the load factor is very high, add more cars. If it's very low, run fewer cars. You may want to figure out how to put cut-off cars at some intermediate point. (If your load factor indicates that you should run significantly less than one car, of course, then you have a route which doesn't have enough ridership to run a train.)
Total ridership is probably the best metric for popularity (uh, by definition of "popularity"). You could argue however that you can *buy* popularity with cheap tickets, so for an underlying "base popularity" you probably want to look at some combination of ridership and ticket yield. The LSL, Starlight, Empire Builder, and Silver Service all run long trains and fill them up on the peak segments, and do so with high ticket prices. They are ultra-popular... compared to, for instance, most of the state-supported corridors, which run short trains, don't come anywhere near filling them up, and do so with lower ticket prices.
jis
Permanent Way Inspector
Staff member
Administator
Moderator
AU Supporting Member
Gathering Team Member
To me, a good way to gain some insight into the popularity of a train is to look at the total ridership per route mile and also look at the revenue per rider per route mile. Grossed up OR is a pretty useless number except when it is way off in the fringes - like exceedingly low or exceedingly high, which are usually cause by other extenuating circumstances which needs to be understood before blindly using it for anything.
In attempting to use such grossed up numbers for fine scale decisions we are making the same mistake that Amtrak makes in its accounting exercise, and perhaps for the same reason too. Amtrak of course goes one step further. It takes the grossed up number and then using some essentially arbitrary criterion then reallocates those numbers back to the piece parts. This is even worse than trying to claim that the grossed up OR is the OR for each segment, just because we forgot to or were unable to collect the number to compute the actuals. It basically allows one to concoct numbers to support any story line that you want to support. That's is what highly paid consultants excel in.
In attempting to use such grossed up numbers for fine scale decisions we are making the same mistake that Amtrak makes in its accounting exercise, and perhaps for the same reason too. Amtrak of course goes one step further. It takes the grossed up number and then using some essentially arbitrary criterion then reallocates those numbers back to the piece parts. This is even worse than trying to claim that the grossed up OR is the OR for each segment, just because we forgot to or were unable to collect the number to compute the actuals. It basically allows one to concoct numbers to support any story line that you want to support. That's is what highly paid consultants excel in.
The grossed up nature of Amtraks Fully Allocated Contribution/(Loss) data is understood. But unless there's something grossed up about Amtraks Passenger or Mile data, when the grossed up data in Column K is divided by the Grossed up data in Column J, the grossed up portions are eliminated. What happens in the division process goes like this (with $ substituted for Fully Allocated Contribution/(Loss) for brevity):
• J = $/Passenger X Mile
• K = $/Seat X Mile
• K ÷ J = ($/Seat X Mile) ÷ ($/Passenger X Mile) = $ X Passenger X Mile =$ X Passenger X Mile = Passenger = Passenger Per Seat = Occupancy Rate, Seat Fill Rate, etc.
$ X Seat X Mile$ X Seat X Mile Seat
Amtrak could gross up the FAC/(L) by trillions of dollars and it wouldn't make any difference as long those figures for each train are grossed up by equal amounts - which they appear to be.
In Neroden's opinion, ridership is the best indicator of popularity. In Ryan's opinion, it isn't. And I think my occupancy figures are the best. And that's probably OK because exactly what the heck indicates the most popular train anyway?
• Gross number of riders?
• Occupancy Rate?
• Number of folks disappointed because there was no room for them on the train?
• Passengers per route mile?
• Revenue per rider per route mile?
• Percentage of passengers that wave hello to motorists at grade crossings?
• Total of tips received by the OBS?
• Number of repeat riders?
• Number of passengers that need to be forcibly removed from the train at the end of its route?
Maybe it's all of the above. So take your pick or add your own to the list. I picked mine yesterday. This is, after all, still the land of the free and the home of the brave!
• J = $/Passenger X Mile
• K = $/Seat X Mile
• K ÷ J = ($/Seat X Mile) ÷ ($/Passenger X Mile) = $ X Passenger X Mile =
$ X Seat X Mile
Amtrak could gross up the FAC/(L) by trillions of dollars and it wouldn't make any difference as long those figures for each train are grossed up by equal amounts - which they appear to be.
In Neroden's opinion, ridership is the best indicator of popularity. In Ryan's opinion, it isn't. And I think my occupancy figures are the best. And that's probably OK because exactly what the heck indicates the most popular train anyway?
• Gross number of riders?
• Occupancy Rate?
• Number of folks disappointed because there was no room for them on the train?
• Passengers per route mile?
• Revenue per rider per route mile?
• Percentage of passengers that wave hello to motorists at grade crossings?
• Total of tips received by the OBS?
• Number of repeat riders?
• Number of passengers that need to be forcibly removed from the train at the end of its route?
Maybe it's all of the above. So take your pick or add your own to the list. I picked mine yesterday. This is, after all, still the land of the free and the home of the brave!
Last edited by a moderator:
jis
Permanent Way Inspector
Staff member
Administator
Moderator
AU Supporting Member
Gathering Team Member
And I actually don't plan to spend any further time on this exercise trying to extract juice out of a rock with insufficient information. So have fun.
So have fun.
Ryan
Court Jester
ainamkartma
Service Attendant
- Joined
- May 26, 2015
- Messages
- 137
(Notice that the TE _effectively_ (in terms of served population and boardings) has just two intermediate stops between San Antonio, TX and Palm Springs, CA: El Paso and Tucson. This is a distance of ~1300 miles, or ~equal to NYC to Miami.)
Ainam "I have a theory that Benson, AZ is a fictional copyright testing town made up by the Southern Pacific" Kartma
The one time I rode the TE (a points run from FTW-DAL) I couldn't help but notice the large number of PBF seatchecks in my car (even more than LRK/STL)...I didn't even know what station that stood for until I looked it up...
At 12:11am today, received an AmSnag Fare Alert saying the roomette fare had decreased by US$113 to $269:
However, Arrow now (about 1:20pm) shows the upcharge for a Roomette is $382 - the same as it previously was. Wonder if VV reacted quickly enough to have his fare modified to that momentary lower one. Or perhaps this was simply Arrow having a spasm with the lower price lasting only a few seconds - and then that square electron dislodged itself from the P-N junction in which it was stuck.
However, Arrow now (about 1:20pm) shows the upcharge for a Roomette is $382 - the same as it previously was. Wonder if VV reacted quickly enough to have his fare modified to that momentary lower one. Or perhaps this was simply Arrow having a spasm with the lower price lasting only a few seconds - and then that square electron dislodged itself from the P-N junction in which it was stuck.
Last edited by a moderator:
- Status
