Report by Noel T. Braymer
Imagine you are at the Albuquerque Train Station and you can catch a train there to Fresno or Denver, St Louis, Phoenix, Pittsburgh, Washington D.C. as well as Chicago and Los Angeles. How many “trains” would be needed to do this? Well 5 but they would all arrive at Albuquerque at the same time.
In this country there isn’t a school to learn how to market intercity passenger trains. There are plenty of places to learn how to build railroads or how to run trains. But there is no place and little is written on how to best operate passenger trains to draw more passengers and run them at a profit. Most of what I’ve learned about rail passenger service that works and brings in passengers I learned from Byron Nordberg and Adrian Herzog through our membership in RailPAC. Byron and Adrian were mostly self-taught but they also learned a great deal from a small network of dedicated people around the country also wanting to see Passenger Trains grow and prosper in this country.
Rather than reinvent the wheel, I’ve decided to look back at some of the work of my late friends to show what is possible and can still be done by concentrating on just the route of the Southwest Chief. Central to the work of Nordberg and Herzog was an accurate way to measure potential ridership for different types of rail service. Adrian or should I say Dr. Herzog (who was a Professor of Astronomy at Cal State University Northridge) used the University’s mainframe computer roughly 30 years ago to create a computer program to simulate ridership for rail passenger service. As a Professor he could use the computer for research of any subject of his choice. The first route that Adrian chose to simulate was that of the Southwest Chief. His program was calibrated so that it would independently produce results like the actual ridership of the Chief. From his research creating this computer simulation Dr. Herzog developed his Matrix Theory for Passenger Trains. Central to ridership is the number of markets a trains serves with station stops and connections. This can be expressed in an equation STATION PAIRS = N x (N – 1). What this means for an example if you have a train line with 3 stations. With those 3 stations there is a total of 6 combinations of potential markets that this line can serve. If you call the stations A, B and C you can go A to B, B to A, A to C, C to A, B to C and C to B. Now if you add a 4th station this jumps up to 12 possible trips you can make on this rail line. To get a real world view of what the implications of this Adrian looked at the Southwest Chief’s which now has 32 stations which using the formula N x (N-1) is 32 x 31 equals 992 city pairs or possible trip combinations. Now compare this to the Capitol Limited between Chicago and Washington and you have 16 stations which is 16 x 15 for 240 city pairs. Now at first glance if you combine the Southwest Chief and Capitol Limited into one extended train you would think if you added 992 city pairs to 240 city pairs this new extended trains would have 1232 potential trips combinations as one train right? Well the math actually is 16 and 32 which is 48 stations. Using Adrian’s formula which is 48 x 47 equals 2256 city pairs not 1232. To quote Dr. Herzog “Integration of the network is like starting a chain reaction: it ignites a growth cycle or upward spiral in utility, usage, output, and financial results.”
A good way to maximize markets at low cost is to run a train with multiple sections. This is an ancient railroad practice that was long ago proven productive. In Europe for years most intercity passenger cars have signs with the car’s destination since if you got on the wrong car you could easily end up in the wrong county! With sections it is almost like getting 2 trains for the price of one. As part of Adrian’s computer simulations he added a section of the Southwest Chief spliting off at Barstow to run up the San Joaquin Valley to the Bay Area and terminating at San Jose. The results astonished Adrian. Ridership on this “San Francisco Chief” section alone was greater than the present ridership on the Southwest Chief. The reason for this was there were more station stops and station pairs from San Jose to Chicago than from Los Angeles. Most modeling for ridership is based on population density and because of this the assumption is that the Los Angeles area should produce greater ridership. Adrian’s paper on his Matrix Theory has a mathematical explanation why this doesn’t always happen. But we can see this at work on the ridership of the Empire Builder which goes through some of the least populated areas of the country. Yet it produces the most income of any passenger train in this county. This is largely because the average trip is very long because the train’s stops are far apart but there still are plenty of stations on this long route which generate greater passenger miles hence the most income. A route which is long with connections to many city pairs can generate ridership and revenue. It should also be noted that the Empire Builder is the only Western Long Distance Train that still splits into 2 sections with this happening at Spokane to Seattle and Portland. Maybe there is something we can learn from that.
One of the five “trains” on the Southwest Chief route could be named the Southwest Chief which under Adrian’s proposal would run from Los Angeles to Chicago and continue on the route of the Capitol Limited. Since both trains use Superliner equipment either the Southwest Chief could be extended to Washington or cars could be transferred between the two current trains. As shown earlier connecting just these two services greatly expands both trains possible markets. What could be done with a San Francisco Chief is connect from San Jose and the San Joaquin Valley at Barstow. At Kansas City this section could be split off and head towards St Louis and end up in Chicago. This new service would require no new stations and little additional overhead but greatly increase the markets served by passenger rail. A new train we could call the El Capitan could start in Phoenix. Ideally it should start in Tucson or even Nogales but currently there isn’t a track connection today between those cities. Since this would be a slow route it would be best to have an overnight service to Williams Junction to connect with the main Southwest Chief. At Trinidad a Denver Chief could split off of the Southwest Chief and head up to Denver for possible connections with the California Zephyr and also serve Colorado cities north of Denver and Cheyenne. With these 5 trains in one, the market for the current Southwest Chief explodes with service to Phoenix, Denver, St Louis, Washington, San Jose and Pittsburgh plus all the other stations in between. This makes possible direct trip combinations like Los Angeles to Denver, Phoenix to Kansas City, St Louis to San Jose, and Washington to Bakersfield.
To be realistic this can’t happen overnight. To do what has just been described for the Southwest Chief route for all the Western Long Distance Trains will require thousands of new rail cars when now there are only hundreds. New equipment can be financed from the increased revenues of these new services but it will still take years to build more cars and locomotives. There are also problems with the condition of the tracks on many lines and getting the cooperation of the railroads to expand service. Right now there is the possibility than unless funding soon can be found to restore the tracks between Belen and Newton to passenger rail standards then the Southwest Chief could be rerouted at Belen via Amarillo to Newton. This would bypass several potential connections to cities like Albuquerque and Denver as well as smaller towns in New Mexico, Colorado and Kansas. Something that could be done in the near future is to combine the Heartland Flyer from Fort Worth to Oklahoma City with the Southwest Chief. There are proposals to extend the Flyer north of Oklahoma City to Kansas City with some possible connections to other trains at Kansas City. As a section of the Southwest Chief not only could same seat service be available on the current Heartland Flyer Route with the current Southwest Chief but also with a section of the Chief going to Chicago by way of St Louis from Kansas City could be served without changing trains. Extending the Flyer to Houston at the same time and connecting with the Sunset would add many more market combinations to both the Southwest Chief and the Heartland Flyer.
So why haven’t these simple fairly short sections been added to America’s intercity trains? One factor is oppositions from the operation department. Adding and breaking sections of trains is more work and works best when trains run on time. Yet railroads all over the world have and continue to add and take off cars and locomotives between trains several times on a single run. The main reason though is in this country train routes and service is determined more by politics than by economics. Corridor trains travel in populated areas with lots of voters and politicians to impress and can carry large loads of passengers on busy trains. A good example of this are commuter trains during rush hour. But commuter trains are big money losers. The distances and fares charged on commuter trains are low and outside of rush hour commuter equipment often sits idle or underused producing little or no revenue. Intercity corridor trains do better than commuter trains but still have problems with limited markets and underused equipment. On Amtrak long distance trains often run at over 60 percent occupancy while corridor trains with smaller consists generally have less than 40 percent occupancy. Plus corridor trains usually are idle overnight while long distance trains run day and all night in revenue service. Much can be done to improve the economics of corridor and commuter trains. Serving other markets besides rush hour trips in and out of downtown 7 days a week over 12 hours in a day improves the economy of commuter and corridor trains. Improving connections and extending the routes of commuter and corridor trains have a major impact on ridership and revenue. Look at the Surfliner trains and the ones to and from San Diego that have been extended to Santa Barbara and San Luis Obispo and you see these trains carry many more passengers and generate much more revenues then the trains just between Los Angeles and San Diego. The whole issue of trying to increase ridership with faster trains by skipping stations always lowers ridership unless there are connecting feeder trains because eliminating stops reduces the number of markets the express trains serves compared with all stop trains. If these simple principles of having maximum markets, route miles and good equipment utilization were better understood by politicians and the public we would have more and busier trains. But as long as train promoters and politicians are fixated with higher train speeds by skipping markets to reduce running times then we will keep seeing the same mistakes repeated again and again which gives trains the reputation of being wasteful and unnecessary.
To see video of Adrian Herzog explaining Matrix Theory and how it would work on trains go to Dr Herzog explains Rail Passenger Matrix Theory Part 1 and Dr Adrian Herzog explains Rail Passenger Matrix Theory Part 2