By William Lindley, Guest Commentator; June 16, 2022
Key points:
• Convenience of use is more important than train speed for passengers.
• Arrival and departure times play a major role in choosing train travel.
• Frequency of operation is critical to attract passengers.
• Terminal to terminal is inefficient; intermediate station stops increase utility and ridership
• For train operators, efficient equipment utilization dictates schedules and passenger convenience.
Convenience is more important than speed
There are some people ‒ including those declaring themselves experts on the subject ‒ who believe and preach passengers will not be attracted to passenger trains unless the trains have competitive travel times with private automobiles.
They are not correct. Here is the reality of the subject.
There has been much written about train speeds and how much that affects whether travelers will select a train trip over driving, but less on the real key question of convenience for the traveler.
Private car travel is the ultimate choice for complete convenience, assuming one is willing and able to use an automobile; passengers/drivers come and go at will and do not follow any set schedule. That is the big competition for trains in markets of under 500 miles, not the exact travel time.
Regardless of how fast a train is, if the departure and arrival times are inconvenient, people will not use it. The same is true for a bus, ship, or jet airplane. This is demonstrated in Europe daily with the high utility of overnight sleeper trains. The trains do not travel particularly fast but leave their departing terminal at a desirable hour and arrive the next morning at the arriving terminal again at a desirable time. Between terminals, passengers receive a comfortable night’s sleep in a real bed with suitable amenities.
The same holds true for daylight travelers; if a train departs at a desirable hour and arrives at a convenient time, passengers will choose the train over driving. But the train must offer amenities driving a private vehicle cannot, such as the ability to work or relax enroute, have sufficient food service, be clean and comfortable and have station facilities which are desirable, including available parking.
The question of longer distance, inter-regional trains is another discussion. Passengers choosing train travel over longer distances which includes trips of more than 15 to 18 hours still are more likely to value the convenience factor over the speed factor. They, too, want to know dependably when they will depart and when they will arrive. If the trip takes longer than the time to drive, but not significantly so, the train will still win because of the amenities outlined above.
Driving a private vehicle can be tiring, expensive and worrisome. The tradeoff of travel by train versus the various hazards of a long road trip in a car or truck for is not a contest, the train wins. But, again, convenience most likely will be the deciding factor.
Each traveler and trip is equally important
There are many types of trips such as work, shopping and leisure; traveling for each is equally important and we must look at them as a group, not individual travelers. Think of it like counting how many people are on a New York City sidewalk, riders on a subway, or cars on a highway, at various times of day. What are reasons someone chooses to ride, or not ride a train?
Imagine for a moment there is an imaginary “transporter” from classic science fiction that has infinite speed, but which only leaves on a fixed schedule. How often would you use this instant transit, regardless of where you are going, if its only departures were at 5:17 am, 2:17 pm, and 11:37 pm? Or only Sundays, Wednesdays, and Fridays at 1:40 am? Think of it like the current Amtrak schedule at Cincinnati.
Now, harken back to the 19th Century and the horse-drawn streetcar with a top speed of perhaps 10 miles an hour. How often would you use this if one came along every three minutes — miss one and the next one is already in sight — if you were only going a mile or two?
Note, too, that the horsecar would usually stop at each intersection, or could be stopped with a wave or a word to the conductor. If you were on the street where it ran, you never had to go far to get on board.
Avoiding the mistake of endpoint mentality
Main line trains generally stop only at stations or designated places several miles apart as accelerating and decelerating a train takes time and energy. The faster your top speed, the longer it takes to get up to or slow down from it at a station. The result is the further apart you will place stations because of this fact. This often leads to the “endpoint mentality.”
If you think of a train same as an airplane, you might plan a rail service like this, using Glendale and Tucson, Arizona as an example:
Leave Glendale 7:00 am
Arrive Tucson 9:30 am
because you are thinking only of top speed and avoiding station stops.
Surrounding areas influenced by a train; if you have to travel too far for a station you will select another transportation choice
Each station has a “catchment area” — the surrounding area, conceptually a circle but shaped by geography and city design. If you must travel too far to a station, you will likely choose another mode entirely.
As an example, if you are in the metropolitan Phoenix town of Glendale, Arizona and must find your way to the nearest train station which is Maricopa, a quaint but remote clearing in the desert 45 miles away to board a train for Tucson, you might as well just stay in your car and drive the remaining hour or so to Tucson. This distance from Maricopa to Tucson is only an additional 84 miles.
“First and last mile” rule: A huge part of the convenience factor which makes usage soar
Catchment area also depends on transit connections and the critical “first and last mile” rule: Tucson’s depot today is served by SunTran trolleys, and Glendale is served by Valley Metro bus lines.
Looking at the original two terminals ‒ Glendale and Tucson ‒ but adding intermediate station stops, convenience for users soars.
Leave Glendale 7:00 am (Valley Metro bus service)
Leave State Capitol 7:15 am (Connection, Valley Metro west line trolley [being built])
Leave Phoenix Union Station 7:30 am
Leave Phoenix Sky Harbor International Airport 7:40 am (Connections, Valley Metro trolley, SkyTrain)
Leave Tempe 7:55 am (Connections, Valley Metro trolley, Tempe Streetcar, ASU main campus)
Leave Mesa 8:10 am (Connection, Mesa Streetcar [planned])
Leave Gilbert 8:25 am
Leave Gateway Airport 8:35 am (Arizona State University East [former Williams Air Force Base])
Leave Coolidge 9:05 am (bus to Casa Grande National Monument)
Arrive Orange Grove Road, Tucson 9:30 am
Arrive Tucson Depot 9:45 am (Connection, SunTran trolley to University of Arizona)
Arrive Tucson International Airport 10:00 am
We go from a single origin/destination combination of two possible trips in the first case (Glendale to Tucson and return), to 132 possible trips (from each of 12 stations to the other 11; 12 x 11 = 132). Furthermore, we add connections to residential zones, business centers, universities, and airports.
Sixty times more useful at small cost to endpoint-to-endpoint time
The entire system becomes sixty times more useful at a small cost in endpoint-to-endpoint time. And by serving both sides of each endpoint city, we achieve closer to the enviable “one-seat ride” which so increases ridership (and passenger-miles, which result in revenue).
Convenience also depends on flexibility and critical early morning, midday and late at night service
For another example: A student, a nurse, a paralegal and a machinist each need to be somewhere at a certain time, but if your transportation choice does not allow for late starts, half days, “come home your child is sick,” or an evening of bowling, a movie theater, or bar talk afterward, most folks will find a single way of transportation that doesn’t require intricate planning.
Consider the schedule of such a system. Showing only the terminal stations, we use two trainsets. To start, we have:
Trainset A.
Leave Glendale 7:00 am, Arrive Tucson International Airport 10:00 am;
Depart Tucson International Airport 4:30 pm, Arrive Glendale 7:30 pm.
Trainset B. (Same schedule in reverse; starting with Tucson International Airport 7:00 am departure)
This brief example gives a poor return on investment; each train operates only six hours and less than 300 miles a day. Worse, as a passenger, you have only one possible path of one departure to each city down the line. From Glendale, for example, you can ride to Tempe in the morning and back in the afternoon, that’s it; there are no earlier or later times available.
Now consider adding a single midday trip with the same equipment.
Trainset A.
Leave Glendale 7:00 am, Arrive Tucson International Airport 10:00 am;
Leave Tucson International Airport 10:30 pm, Arrive Glendale 1:30 pm;
Leave Glendale 4:30 pm, Arrive Tucson International Airport 7:30 pm.
Trainset B. (likewise in reverse)
From Glendale you could catch the 7:00 am train in time for an arrival in Tempe before 8am, and then catch either the midday or the evening train home. This 50% increase in train-miles has made the system twice as useful. In short, the Network or Matrix Effect works as much for time as it does for place.
Adding new time slots adds new origin/departure combinations creating greater ridership
Just as adding a single station to a route adds twice as many origin/departure combinations as stations on the route (from each existing station to the new one; and back), each new time slot (up to some limit) adds origin/departure time combinations.
The reason why slow attrition of branch line passenger service in the 1950s led to faster unraveling of the intercity rail network in the 1960s
Conversely, cutting a single primary time slot, or a single station, from a route’s schedule reduces the time or place combinations by twice as many as there are stations or departures remaining. That explains why the slow attrition of branch line passenger service in the 1950s led to faster and faster unraveling of the intercity rail network in the 1960s: each single cut (whether of a train departure or of a station served) had a far larger system-wide impact that might have been obvious.
Three important points
Let us define three things:
• The total perceived travel time is the actual travel time plus half the time elapsed since the previous departure.
• The perceived speed is the distance traveled, divided by the above perceived travel time.
• The perceived convenience is directly related to its perceived speed, weighted against the perceived speeds of the alternatives.
The slower ferry gets passengers to their destination quicker than a train, bus, or car; “top speed” does not matter
Boston’s MBTA offers commuter ferry service from several towns around the Bay to Long Wharf right on Boston’s waterfront, even though most of those towns have rail and bus service.
The ferry, being more direct, gets there quicker despite a much lower top speed. Highways, which go around the Bay and are plagued by backups and slowdowns, are a less-attractive alternative. The “perceived speed” of the ferry is equal to or slightly faster than the “perceived speed” of the train or driving: “top speed” matters not at all.
Looking at the results in a graph
If we graph the number of daily departures beyond which adding more trains does not add significant ridership (i.e., the line of maximum return on market share) we get something like the chart above..
Frequency of departure does not matter as much for longer trips
As the trip length increases, the number of daily departures, or the frequency of departures, becomes less and less important in the aggregate. This is called a logistic curve, and is often seen in the natural sciences.
The New York Central Railroad in 1958 provides a good example
As for a train [1] from Chicago to Cleveland via Cincinnati: As just one example, the New York Central in the July 1958 timetable lists the overnight Train 306, The Midwestern, with a schedule departing Chicago at 5:10 pm, arriving at Cincinnati 11:00 pm and Cleveland at 7:30 am.
The train had a sleeping car from Chicago to Asheville, North Carolina, handled south of Cincinnati by the Southern Railway; a sleeper from Cincinnati to Cleveland; and another sleeping car from Cincinnati to Buffalo, carried in No. 306 to Cleveland and then in No. 50, the Empire State Express, to Buffalo. Coaches were operated from Chicago to Cleveland via Cincinnati. The through-cars left Cleveland at 9:15 am and arrived Buffalo at 12:45 pm.
[1] Actually a car-line handled in two trains but advertised to the public and operated as if it were its own train.
True, you could have chosen No. 358, the Niagara, departing Chicago at 8:50 pm and arriving Buffalo at 9:45 am, cutting six hours off the travel time, and with more convenient departure and arrival times. The New York Central, however, operated this train to serve not just Chicago and Buffalo but Kankakee, Lafayette, Indianapolis, Cincinnati, Dayton, Springfield, and Columbus — and all the combinations of those city-pairs added to the Cincinnati–Buffalo train. On all these routes, the New York Central offered three or more daily trains; there were four each way daily between Cincinnati, Indianapolis, and Chicago.
That was then, this is what could be now
Today, just the routing into Chicago from Kankakee north and west of Lafayette, Indiana as an alternative to Dyer — on a rail line that still exists — would open a fair number of city-pair combinations even in the present Amtrak network. If we imagine three daily trains departing Indianapolis for Chicago, the morning train might take the current route through Dyer (with suburban Chicago stops added for METRA and CTA connections); the second train could travel via Kankakee, and the third overnight from Indianapolis to Fort Wayne, Elkhart, South Bend, and transiting the South Shore Line making limited stops and terminating at Millennium Station.
The sum of it all
Altogether, we see that:
• A matrix of arrival and departures is important not just in space but also in time.
• Serving more stations with more schedules results in much more usability, and resultant revenue, than might be obvious.
• Increasing speed, or adding frequencies, has a huge impact at the low end but eventually with diminishing returns.
• Trains, being not airplanes, have tremendous flexibility both in serving intermediate stations and in exchanging cars with each other.
It is true, competition from superhighways and jet aircraft hurt the railroads in the 1950s. At the same time, streetcars were removed from most cities, and bus services declined; little effort was made to connect trains to airports.
Yet it was eliminating service to small towns and cutting secondary trains — and the railroads’ failure to work with each other on a coherent network of what remained — that resulted in entire intercity passenger network falling apart, ever faster, through the 1960s.
Today, it is possible to use the same mathematics in reverse to build connectivity and unleash the latent demand for regional and intercity passenger trains.
