Q: Engine Bob, Why does my morning train take 48 minutes, while my evening train—with the exact same stops—takes 42 minutes? Does it take longer for people to get on than get off? 

A: First, pat yourself on the back for picking up on a detail that 99.98% of commuters have never noticed. We’re going to have a little fun with your question because, first, I’m going to give you my speculation and then, second, we’re going to see what the MTA has to say about it. Last night I stopped into the Stationmaster’s Office to retrieve a customer-comment/question form, and on it I wrote down a more detailed version of your question, using two specific New Haven trains for comparison—ones that showed a seven-minute discrepancy between north and southbound runs. The Stationmaster promised me I’d receive a personal reply from the Public Affairs Office. We shall see. 

Meanwhile, here’s my take. Assuming that the trains you’re comparing share all the relevant parameters in common—both peak, same station stops, no transfers with one and no the other, etc.—what you’ve struck upon here is evidence of a practice that’s as old as train schedules themselves. 

It’s known as “schedule padding.” 

While it may take a wee bit longer to board a train than disembark from one, I doubt that such a thing would result in a schedule adjustment. And if it did, logic would hold that you’d see a greater amount of time allotted for station stops where the most people get on—and when I tore through some Metro North schedules last night, that’s not where I found the discrepancies. More on that in a minute. 

Here’s another interesting observation. While your trip home (northbound, I presume) is taking six minutes longer, I found that the opposite appears to be more often the case. For example—and these are all peak trains I’m referring to, with all station stops common to both directions—catch a morning express train at South Norwalk and you’ll get to Grand Central in 55 minutes. Homeward on the same run? 1 hour, 4 minutes. Hop an early local out of Stamford and you’ll be at GCT in 1 hour, 2 minutes. Grab the same run back and it’ll take 1 hour, 9 minutes. 

Is that just the New Haven’s quirk? Nope. Let’s get a morning-rush local out of Croton-Harmon on the Hudson Line. Train No. 704 will take precisely 1 hour to reach Grand Central. But at evening peak, the same run (Train No. 757) will take 1 hour, 4 minutes. 

I found a few other examples in which run times were exact, and a few, like yours, in which the return runs took less time. But overall, it seems far more common for the northbound runs to take longer. And this brings me back to the practice of schedule padding. 

What is “schedule padding”? Schedule padding is the practice of adding a few extra minutes to a run—time that a train does not really need to cover the allotted distance—merely to “help” a late train make up time on paper, and hence arrive “on time” even if, in fact, it’s not. In short, schedule padding helps boost a railroad’s on-time performance statistics. Before you shake a fist at Metro North for stooping this low, consider: New Jersey Transit does it, too. So does Amtrak. So does VIA Rail in Canada. Most every railroad uses schedule padding. 

But let’s take a closer look at how it works. We’ll go back to our local train between
Stamford and Grand Central, because the discrepancy is pretty dramatic—a full seven minutes more on the return (northbound) run. 

I made a chart and marked down all the station stops with the corresponding route-mile distance for each (Grand Central is Mile Zero; Pelham is at Mile 15; Harrison is Mile 22, etc.) With some simple subtraction and comparison, we see that both southbound and northbound trains take the same time to cover the same distances for the vast majority of the run. For instance, The southbound train travels the 14 miles between Mt. Vernon East and GCT in 26 minutes—and the northbound train takes just as long over that same stretch. It’s an even time match for the bulk of the route. 

But watch closely! Let’s take a look at the last two stops at the top of the run—Stamford and Old Greenwich. Those two stations are 2 miles apart. If you’re taking the train southbound out of Stamford, you’ll get to the next stop in only 2 minutes. Returning home, however—why, this is amazing!—to cover that same distance of 2 miles between Old Greenwich and Stamford, Metro North allots SIX MINUTES. 

Where did those four extra minutes come from? Does the northbound train pause to pick up a crate of caviar and champagne? Hardly. The schedule has been padded. So if northbound Train No. 1360 out of Grand Central is running a few minutes late, not to worry! That two minutes the train should take to traverse the final leg of the trip has been conveniently tripled. Think of it as loosening your belt by two holes. There now, don’t you feel thinner? This numeric hocus pocus means that a really late train will be less late, and a slightly late train will suddenly, magically, be “on time.” 

(Incidentally, the one other extra minute on the northbound run that adds up to the total extra northbound travel time of seven minutes comes from the stretch between Harrison and
Greenwich. That’s six miles of track; the southbound train does it in 10 minutes while the northbound takes 11.) 

As I said, schedule padding is not a secret that only Metro North knows. A number of years ago when Shirley Delibero headed up New Jersey Transit, an absurd amount of padding was added to the schedules—always between the second-to-last and the last stop, as in my example above—to keep on-time statistics high. But too many people noticed the trick, and the railroad’s loose schedules quickly became known as “Shirley Time.” 

In theory, if a railroad uses padding to improve on-time statistics, it should be able to use them in both directions, right? But in the case of Metro North—I’m making an educated guess, here—the four-mile stretch between 125th Street and Grand Central is so complex a choke point shared by so many trains at any given time, it’s just not a wise place to toy with the schedule. Uniformity rules on this stretch; I’ve seen no examples of trains taking longer in either direction. Plus, I just have this feeling that discrepancies would be easier to notice down on the GCT end of the timetable; it’s not a good place to hide stuff. 

But let us not close the book on this matter yet, and wait to see how the good people at Metro North explain these discrepancies—if, that is, they reply to my comment form.

Got a question for Engine Bob? Email it to trainjotting@gmail.com.