It’s been a long summer for Metra. And it’s still just the middle of July.
Yesterday, Metra had their second board meeting since the new BNSF schedules took effect on June 11. While the woes of the BNSF were discussed at the previous board meeting on June 13, the change was still pretty fresh and it was generally understood that there’d be a few kinks to work out. (Well, a few kinks for riders to work out.)
Unfortunately for Metra, things on the BNSF haven’t really improved in the last month. Peak period trains are still packed on the best of days; throw in a fleet that isn’t getting any younger, aging air conditioning units on board, and a summer packed full of hot and humid days, and it’s been downright miserable for a lot of BNSF riders. (And, to be fair, riders on other lines have also been complaining about broken air conditioning throughout the system, so that particular issue isn’t unique to the BNSF. Old fleets break more often; unfortunately, it’s to be expected.) But the busiest commuter rail line in the region having chronic month-long overcrowding issues following a schedule rollout that was specifically designed to address existing overcrowding was eventuallty going to get some attention from local media, and local media did not disappoint.
So why is this such an issue now? Sure, the fleet’s older than ever before and the BNSF is by far Metra’s busiest line both in terms of passenger volume and number of trains operating, but ridership is stagnant or dropping and Chicago’s had plenty of hot summers before. No, there must be something different about this summer, but I just can’t put my finger on it…
Luckily Metra understands that the new BNSF schedule fails to accommodate the demands of the riders who rely on the service, and the agency said they would try harder to make changes to the new schedulhahaha, just kidding, Metra blamed riders again. From the Tribune:
“This line has grown and grown and grown over the years, and we’ve completely saturated this line,” Metra CEO Jim Derwinski told reporters after the meeting.
Derwinski said one reason why complaints have increased so much in the last month is that some riders have shifted to different trains because of the schedule change, so there is now crowding on trains not seen before.
On the bright side, the Metra board is starting to feel the heat. From the same article:
“I’ve heard more in the last 30 to 60 days from disgruntled passengers than I’ve heard in five years on the board,” said Metra board member John Zediker, who called for a “deeper dive” into the line’s problems. “What’s going on on that line is unacceptable.”
Good news, Director Zediker: your friends here at The Yard Social Club and Star:Line Chicago performed a deeper dive into the line’s problems. We crunched the numbers between the old and new schedules, and we found that Mr. Derwinski is right: riders are indeed shifting to different trains and showing different crowding patterns. It’s a shame no one saw this coming before the new schedule took effect, and it’s even more of a shame no one tried to warn Metra this would happen. While we don’t have ridership data by train, comparing and aggregating the old and new schedules provided some great insights into where riders are shifting their travel patterns and why some trains are getting crush-loaded while others are well undercapacity.
We broke down the morning and afternoon peak periods into fifteen minute increments to help compare between the two schedules, with the assumption that a change of 15 minutes or less is considered “reasonable” for most people.
For the morning commute, we worked backwards, looking at what time trains were scheduled to arrive in Union Station, then noting what stops that train made, regardless of whether the train ran express, local, or some combination of the two. In the below table, each light green cell represents one “station pair” that gets a rider into Union Station within the time period listed on top, so if you need to get to work by 8:30am and you work 15 minutes from Union Station, a green cell in the 8:01am-8:15am period indicates there’s a train from your station that will get you downtown and to work with time to spare. (Dark green cells indicate there are two trains from that station that get downtown in the particular time period.) The numbers in each cell indicate the change in trains between the old schedule and the new schedule, so a “1” indicates the new schedule has added a train serving that station pair at that timepoint, and a “-1” indicates a train at that timepoint was lost in the schedule change.
The chart shows a few interesting issues with the new schedule. First and foremost, Metra did do their best to minimize overall service levels: no stations lost more than one train in the morning; of the stations that did lose service, there’s a train that will still get you downtown within the next 15-minute period. As a LaGrange Road rider, I do appreciate that we were one of the four stations that actually added a new morning train.
But the chart also shows just what we warned way back when: to accommodate the longer times to flip trains with PTC, Metra tried to “flatten” the peak: early trains got earlier and late trains got later, and the hope was that riders would simply go along with their old train. The blue cells show where a train was moved forward and arrives in the earlier 15-minute interval; purple cells show where a train was pushed back to the following 15-minute interval. While this does keep the overall balance of trains the same more or less, it doesn’t take into account basic human behavior: no one wants to wake up any earlier than they have to, and generally speaking most office workers still have a somewhat firm starting time in the morning. In other words, many riders will find those early blue-cell trains unattractive (need to wake up earlier) and those late purple-cell trains are also unattractive (need to stay at work longer or take a shorter lunch to make up for coming in later), so riders will pile onto the remaining trains in the center of the peak, which makes the peak even “peakier”, exacerbating the very congestion Metra was trying to relieve.
Metra thought they could balance out the morning peak service by adding trains on the shoulders of the peak, and it’s backfiring. If you’re looking for one takeaway statistic, check out the bottom row, which is a sum of the changes by time period. In the peak-of-the-peak, which we defined as 6:30am-8:30am, a whopping 45 station pairs were removed while only six were added. In particular, check out all those red -1s in the 7:45am-8:30am periods for stations between Congress Park and Lisle. All those Zone C/D/E riders went from three trains every morning in that sweet-spot time range to only one or two. Sure, Metra made up for it by cramming a few late-peak express trains in there, but what’s the use of a train that arrives at Union Station at 9:00am when you had to be at work at 8:30am?
The PM peak is a little more challenging to analyze, since there are now two time periods to track: what time riders board at Union Station, and what time they arrive back home. To analyze the PM peak, we took five sample stations and charted Union Station departure times and outlying station arrival times, then calculated the minutes of travel. Otherwise, the same assumptions apply: people generally want to get home from work as soon as they can, and it’s inconvenient to get home any later than the old schedule. Likewise, for some riders, earlier departures would also be inconvenient, since work schedules may not allow riders to make it to Union Station in time for the train.
We looked at five stations in five different zones: Route 59 (Zone G), Naperville (Zone F), Downers Grove (Zone E), Hinsdale (Zone D), and LaGrange Road (Zone C). Gray cells indicate the schedule pairs in the previous schedule; green cells indicate the schedule pairs in the new schedule. The numbers in each cell indicate scheduled travel times; bolded trains are expresses and italicized trains are locals. Where trains overlap between schedules, the new schedule is shown as a slightly darker font color.
The first observation is to note that, for four of the five stations, there were actually a few trains that straight-up improved: left downtown at about the same time and got back to the suburbs in the earlier 15-minute interval. (Reminder: that doesn’t mean it’s 15 minutes faster, just that those riders are getting home notably earlier.) Good on Metra for that.
Other than that, more of the same from the morning peak analysis: late trains get later. Average travel times for each of these five stations increased in the 5-10% range; unlike the extra PTC time required to flip trains, these longer travel times are based solely on in-service operations. While some of the delay is likely due to changes in stopping patterns (Naperville lost most of their super-express trains), this is hard to explain by anything other than atrophy of the system. Metra will occasionally make “schedule adjustments to better reflect operating conditions”, which usually means tacking on time to the end of the run to keep the train on-time, but sometimes they also go through and balance the schedule at each stop. The catch is that nothing is really changing out in the field: the same trains are running on the same tracks, making the same stops, so there’s no good excuse for why trains are gradually getting slower, unless ridership is increasing and it’s taking trains longer to load and unload passengers. (But again, ridership is flat or declining.)
The most interesting observation though is Zone D (Hinsdale) and Zone C (LaGrange Road) losing an afternoon peak train in the new schedule. While the loss isn’t terribly significant from a time management perspective — in both cases, there’s still a train making the same connection in the same time periods — it does shed some light on new potential crowding issues.
And there’s the rub: yes, the BNSF is crowded. But it was crowded before. Naperville riders have known all too well that the line is overcrowded; even with the old schedule it wasn’t uncommon for morning trains to essentially load-and-go, leaving passengers on the platform to hope for better luck on the following train. The difference is now, different parts of the line are feeling the pinch thanks to the new schedule. Zone C/D riders are getting pinched both in the morning and in the afternoon more than they were previously used to, and sure enough, the Trib‘s token complaining commuter comes from LaGrange Park:
Brian Pitts, a 48-year-old resident of La Grange Park who has commuted on Metra since 1998, said that both he and his wife, Carla Pitts, have seen “out of control” crowding on express trains with people “packed like sardines.”
“They lost thousands of $$$s by not collecting fares,” said Pitts in an email to the Tribune, referring to the problem of conductors not being able to get through the train aisles to collect tickets.
It’s clear that Metra is hoping BNSF riders are simply accepting that this will be the new normal and they’ll just change their work schedules to take advantage of the later, less-crowded express trains and just go back to complaining about fares and the fleet on Twitter. But this time riders are pushing back, going straight to the Metra board — or straight to the Tribune and Channel 2.
The new schedule stinks — and after the deep dive we can say that objectively, not just subjectively — and to date Metra and the BNSF aren’t exactly showing any urgency about fixing it. Oh, well, it’s crowded because the fleet’s old. It’s crowded because there are problems with the air conditioning. It’s crowded because passenger trains get delayed by freight trains. It’s crowded because the signals need upgrading. It’s always something, and Metra’s standard recommendation is to just throw more money at it:
But Metra is limited on how much it can do because it has a limited number of cars and old equipment, Derwinski noted. The ultimate solution is more money, he said, and the state legislature has not passed a capital bill to pay for transportation infrastructure work since 2009.
Derwinski noted that the Metra Electric District, which has new cars, is not having the reliability problems seen with older, diesel locomotives.
“One of the things we’re definitely going to need is a capital bill and start replacing a big chunk of our fleet,” Derwinski said.
But all those excuses miss a central point: the new schedule ignores how riders actually use the system. Yes, having an old fleet with crappy air conditioning isn’t helping anything, but the larger current capacity issues are pretty clearly a result of supply and demand: the schedule changed to offer less supply (fewer trains) at the highest period of demand (peak of the peak period). That’s all there is to it, and now there are more standing-room-only trains.
That brings up another thing we can learn from the above analysis: look at how basic and simplistic the schedule changes were. Trains were simply moved around within the old schedule. Sure, some trains added a few stops and some trains dropped a few stops, but otherwise it was just a matter of shuffling the same old trains around so they could have 15 minutes to flip instead of ten, then inexplicably changing consist lengths without trying to forecast how many people would take which trains.
What Metra should do is start with a clean slate and totally reimagine the schedule from the ground up. Pretend this is a brand new commuter line being started (except with decades of data about ridership patterns) and make a new schedule. Be bold.
Or Metra can keep trying to fit this square peg of a schedule into the round hole that is ridership demand, which honestly is the more likely path forward for the agency. But hey, the good news is, the overcrowding issues will take care of themselves as Pace keeps adding more I-55 express bus service and more people start working from home because, honestly, who wants to deal with the BNSF any more?