8 ms·

The government data I am able to find does not seem to support the idea that there is a massive (or any) increase in derailments. And this letter never provides

The government data I am able to find does not seem to support the idea that there is a massive (or any) increase in derailments. And this letter never provides any data supporting the premise either. According to the Bureau of Transportation Statistics, there were 1,056 train derailments (both cargo and passenger) in the US in 2021. This was the lowest number in the dataset going back to 1975. Ten years ago, it was 1,470 derailments. In the year 2000, it was over 2,000 derailments. In reality, the data suggest we are on a strong downward trend in derailments per year. Is this government data wrong? Or is this writer trying out a career in fiction? Data: https://www.bts.gov/content/train-fatalities-injuries-and-accidents-type-accidenta https://www.bts.gov/content/train-fatalities-injuries-and-ac...

beamatronic 4y ago
Perhaps there are different definitions of “derailed”. After all one of the communication problems in this pandemic was that even technical, medical professionals could not agree on the definition of “airborne” as it relates to the public’s best interest.
simulate-me 4y ago
Isn’t there a third option that both your interpretation of the data and the letter are wrong? The letter specifically mentions the derailment of monster trains. These trains have more cargo and so one derailment can lead to a more severe delay in shipping times.
- maxmcd 4y ago
  You can see the ratio here: https://www.bts.gov/content/us-vehicle-miles https://www.bts.gov/content/us-vehicle-miles. Breaks down car miles vs train miles. I hope the third option is a bit more of just trying to take the situation in. This is someone sharing their experience, the executives running the show have their perspective as well. I'm not sure why we have to jump so quickly to "I found data that refutes this, is this person lying or is the data wrong?". Just seems to leave out so much nuance. Feels like a very tech industry "see the data and make a decision". This situation is so human, it's about fear and whether those fears are being considered.
  - kbenson 4y ago
    > I'm not sure why we have to jump so quickly to "I found data that refutes this, is this person lying or is the data wrong?". I don't think it's that people are jumping to it as much as people are trying to identify what's actually going on. The solution for "this problem is happening more often" is different than for "this problem appears to be happening more, but is actually happening less". Applying a process change for what was a perception problem might actually make things worse. It's important to identify the problem so the correct fix can be determined. Even if this actually is a problem but the national data belies that, that could be useful in pointing towards the problem being more localized to Iowa, and differences in that area can be looked for.
  - mediaman 4y ago
    That data is great - thank you for posting. It looks like rail-car miles have held somewhat steady, down a little bit, except for 2020, which I assume was COVID related. I am being a bit facetious with the fiction comment. However, I do believe there are a lot of problems in the rail industry. With the introduction of "PSR" (precision-scheduled railroading), trains actually have no schedule, and only leave when they're full. This maximizes efficiency. It's very hard on rail staff, because it means they, too have no schedule. They work a train in one direction, do a staff change, and then go to a hotel while waiting for a return trip. They get as little as 90-120 minutes of notice for the return trip. So it's a lot of time away from home, with very little ability to control one's schedule. I imagine that this source of frustration emerges in a lot of different places, and it's possible that this letter was one of them.
    - doodlebugging 4y ago
      >It's very hard on rail staff, because it means they, too have no schedule. They work a train in one direction, do a staff change, and then go to a hotel while waiting for a return trip. They get as little as 90-120 minutes of notice for the return trip. So it's a lot of time away from home, with very little ability to control one's schedule. This is a major frustration in a nutshell but you have to know that in addition to the short notice that they are about to get out for the return trip back home all the time they spend off duty in the hotel room counts (to the railroad) as time off, in other words it counts as if it totally belongs to them to do with as they please going about all their normal business. In reality they are in a hotel many times with limited meal options, with no way to visit family or to help out around the house, to conduct any of the other normal household business. Their time in the hotels is not their time at all to the point where the actual time they spend on a run (a round trip) can be multiple days and only a part of that time is compensated. A recent text I had from my relative currently with a railroad gives the railroad's perspective quite clearly during a Zoom meeting - "Y'all hired out to work 24/7/365, we are going to make good employees out of you all." I am part of a fourth generation railroad family. Three of the last four generations, beginning in 1942, have worked on the railroad. It skipped me because I went a different direction. EDIT: I forgot to mention that the ultimate goal of the railroads is to decrease train crew size to 1. Each of these trains will be driven by a one-man train "crew". A couple decades ago normal crew size was three - engineer, brakeman, conductor. Brakeman positions were eliminated leaving the engineer and conductors rolling the trains. The railroads were supposed to be using PTC, positive train control a long time ago but were able to dodge the upgrades to track and infrastructure in the typical American corporate way by delaying using a standard set of excuses. PTC would bring about a single man crew by automating almost everything. That would mean huge loss of jobs and a huge investment in infrastructure upgrades since many thousands of miles of tracks have speed limits that are in place due to trackage that can't handle higher speeds safely. There's a lot going on here and it's always deeper than one single argument.
joshstrange 4y ago
I think the other important piece of data you would need to find is average train length over that same period of time. Fewer derailments could have just as much or a larger impact if the train length is growing.
- foobarian 4y ago
  Right, given same car count and derailment rate, 2x train length could appear as 1/2 the derailments.
  - AnimalMuppet 4y ago
    Depends on if the probability of a derailment is K per train mile, or K per car mile. And I suspect that the answer is "some of both". When a car has a mechanical failure that causes a derailment, that kind of derailment is going to be proportional to the number of car miles. Some other causes (misaligned switches, say) will be proportional to the number of train miles.
- mediaman 4y ago
  I did find a GAO study that said that trains increased in length by 25% since 2008. But the number of derailments dropped by more than 25% during that time, so the probability of a given car being involved in a derailed train still decreased. There's been other coverage of the ironically named "precision scheduling" in the rail industry causing a lot of labor problems, because it makes for nightmarishly unpredictable schedules for workers. They never know when a train is going to leave until it is full, so after working a shift, workers stay in a hotel waiting for a return trip that they won't know about until 90-120 minutes ahead of departure. I am guessing that the claims around increased derailment may actually be about frustration with the work environment.
- simulate-me 4y ago
  The average length isn’t enough because the likelihood of derailment might be a function of train length. It’s possible shorter trains have gotten safer and monster trains have gotten more dangerous due to their increasing length.
pas 4y ago
Is there comparable data on other countries? What is this a proxy for? (Track upkeep problems? Trains are too fast for the load? Signaling issues? Mechanical issues?) Rail safety seems a lot less complex and pretty much as regulated as aviation safety, how come there aren't just ~1-2 per year? Is it just because fatalities are luckily fewer (due to the nature of these derailments) so no one really gives a shit?
- redshirtrob 4y ago
  I don't know if there's comparable data, but I can tell you the mining companies in north west Australia would routinely run 300+ car trains without incident. But these were very much a straight line from the mine to the port without much civilization in between.
- r_hoods_ghost 4y ago
  Well in the UK there were 11 derailments total in 2020-21. The UK has roughly ~10,000 miles of railway and the USA ~150,000. Probably not a good comparison tbh. EU wide would probably be more appropriate but I couldn't find anything. https://dataportal.orr.gov.uk/media/1999/rail-safety-2020-2021.pdf https://dataportal.orr.gov.uk/media/1999/rail-safety-2020-20...
  - hardware2win 4y ago
    Indeed, 5 derailments per day sounds like a lot
- brnaftr361 4y ago
  It's just the forces of the trains. When it's not a derailment it's breaking them. They're doubling the running room for slack action^, for one. That's the least of the concern, but there's more potential for breakage because there's more potential for greater differences in velocity from one end of the train to the other. Topology also matters, previously consists here average 135 cars, but range from 115 to 150, they're 53' each. Usually they'll have 4-5 motors at 74' each. Suffice it to say, depending on the territory there can be huge force differentials generated which can break trains or rail, think cresting a grade on undulating hills. And you can "stringline" trains too, literally pull cars off the tracks, and it's exactly those forces that'd do it. And then there's the human element. Nobody knows when they're going to work, how long it's going to take, what kind of bullshit they're going to run into. You can get stuck for long periods of time away from home terminal, and the Class 1 doesn't care. And there's things like rules on how to walk. The environment that's been constructed isn't meant for human consumption and it's taking a toll on the workforce. There's reams of other shit to whinge about. And the derailments always matter, it's a total clusterfuck. They're not easy to fix, you've got to "bad order" and set out the car, at least by regulation. This involves rerailing, finding a set-out track, and making all the moves to get everything back in order. You've also got to wait for carmen (who are understaffed) and management, which depending on the route, could be a long sit. If it's bad enough it'll shut down traffic for a couple of days, requiring significant diversions around it while the Class 1 pays out the ass for a third party to come clean it all up and rerail shit. None of the current infrastructure is built to deal with these train lengths, either, so everything moves slower. The fact the Surface Transportation Board had a hearing with the Class 1's, I think, is a pretty good indicator they're about as interested in moving goods as GE is in manufacturing. ^: Each coal car has two draw bars with +/- 2-3" about the same for shipping container cars (from eyeing it) of run built in. This varies with car, auto racks have way more run out. Opposing forces in the trainline will pull it apart, the more run to build the faster the opponent acceleration... This typically results in broken knuckles. But it is possible to just push cars off the rail with these forces, which is exacerbated by the doubled length.
scythe 4y ago
The gross number of derailments isn't very meaningful, what you need is the derailments per train-miles-traveled, analogous to the collisions per vehicle-miles-traveled that is used for road safety. If we're running fewer trains, obviously we should expect fewer derailments. As some other commenters have noted, it might be even more useful to get the total weight of derailed cargo, per (mass times distance), I suppose.
temp8964 4y ago
foxyv 4y ago
I think the author was trying to explain the risks they are taking and saying that it's only a matter of time until we start to see disasters similar to the one in 2004 but 100 times worse. https://cen.acs.org/articles/82/i27/TRAIN-DERAILS-CHLORINE-LEAKED.html#:~:text=Two%20freight%20trains%20collided%20early,and%20killed%20two%20more%20people https://cen.acs.org/articles/82/i27/TRAIN-DERAILS-CHLORINE-L....
wheelerwj 4y ago
a few other people have made a few suggestions about interpreting the data, but i would also like to add that the writer is specifically citing "Monster Trains" and that may or may not affect the interpretation of data. still, awesome that you went to check the source.
scottlamb 4y ago
> According to the Bureau of Transportation Statistics, there were 1,056 train derailments (both cargo and passenger) in the US in 2021 The key word here might be "US". The letter writer is specifically talking about Iowa; are state-by-state statistics available? If you've never spent time in the Midwest, you probably have never seen a train that's a mile or longer. Endless cars of agricultural products, often high fructose corn syrup. I saw this often growing up in Iowa, never in California. In the Bay Area, freight sometimes moves along the same tracks as Caltrain at night, but the trains are tiny in comparison to those of Iowa. For a while, I think trains in Iowa were limited by how long a train was allowed to block a roadway (IIRC 10 minutes). I think trains were sized so at least in theory they could comply with the law when at speed. The law might have changed since, and/or maybe there aren't as many at-grade crossings anymore. (A painful intersection in my home town recently got grade separation.)
- dhosek 4y ago
  Other than a few years in L.A., I've lived my whole life within a mile of a rail line. It’s only in the last month that I have ever seen a mid-train locomotive. Something is definitely changing.
  - brnaftr361 4y ago
    They started this process a long time ago. 2018 was when they first started talking about it, locally they did the first pilots about 2019 if memory serves.
- AnimalMuppet 4y ago
  Out here in the west, I have seen longer trains in the past few years, but I've been seeing mile-long trains for years. I never saw trains with more than 150 cars before the last few years, though.
TimPC 4y ago
Derailments could be getting worse even with fewer of them. The author of the letter is talking about the trend towards monster trains which suggests fewer trains being run. Is the number of derailments per train going up? Is the impact of the larger trains derailing leading to more overall impact because each individual derailment is larger even though there are few of them? If I run half as many twice as large trains I'd expect fewer total derailments even if these monster trains derail more often than the more normal trains of yesteryear.
exmicrosoldier 4y ago
it's easier to have less total derailments, but higher numbers of cars derailed. Put every train car in the country on one train and derail it. Now 100% of the capacity was derailed, but it's an incredible success, because the number of train derailments shrank from 1056 to 1! So much safety!

beamatronic 4y ago

Perhaps there are different definitions of “derailed”. After all one of the communication problems in this pandemic was that even technical, medical professionals could not agree on the definition of “airborne” as it relates to the public’s best interest.

simulate-me 4y ago

Isn’t there a third option that both your interpretation of the data and the letter are wrong? The letter specifically mentions the derailment of monster trains. These trains have more cargo and so one derailment can lead to a more severe delay in shipping times.

maxmcd 4y ago
You can see the ratio here: https://www.bts.gov/content/us-vehicle-miles https://www.bts.gov/content/us-vehicle-miles. Breaks down car miles vs train miles. I hope the third option is a bit more of just trying to take the situation in. This is someone sharing their experience, the executives running the show have their perspective as well. I'm not sure why we have to jump so quickly to "I found data that refutes this, is this person lying or is the data wrong?". Just seems to leave out so much nuance. Feels like a very tech industry "see the data and make a decision". This situation is so human, it's about fear and whether those fears are being considered.
- kbenson 4y ago
  > I'm not sure why we have to jump so quickly to "I found data that refutes this, is this person lying or is the data wrong?". I don't think it's that people are jumping to it as much as people are trying to identify what's actually going on. The solution for "this problem is happening more often" is different than for "this problem appears to be happening more, but is actually happening less". Applying a process change for what was a perception problem might actually make things worse. It's important to identify the problem so the correct fix can be determined. Even if this actually is a problem but the national data belies that, that could be useful in pointing towards the problem being more localized to Iowa, and differences in that area can be looked for.
- mediaman 4y ago
  That data is great - thank you for posting. It looks like rail-car miles have held somewhat steady, down a little bit, except for 2020, which I assume was COVID related. I am being a bit facetious with the fiction comment. However, I do believe there are a lot of problems in the rail industry. With the introduction of "PSR" (precision-scheduled railroading), trains actually have no schedule, and only leave when they're full. This maximizes efficiency. It's very hard on rail staff, because it means they, too have no schedule. They work a train in one direction, do a staff change, and then go to a hotel while waiting for a return trip. They get as little as 90-120 minutes of notice for the return trip. So it's a lot of time away from home, with very little ability to control one's schedule. I imagine that this source of frustration emerges in a lot of different places, and it's possible that this letter was one of them.
  - doodlebugging 4y ago
    >It's very hard on rail staff, because it means they, too have no schedule. They work a train in one direction, do a staff change, and then go to a hotel while waiting for a return trip. They get as little as 90-120 minutes of notice for the return trip. So it's a lot of time away from home, with very little ability to control one's schedule. This is a major frustration in a nutshell but you have to know that in addition to the short notice that they are about to get out for the return trip back home all the time they spend off duty in the hotel room counts (to the railroad) as time off, in other words it counts as if it totally belongs to them to do with as they please going about all their normal business. In reality they are in a hotel many times with limited meal options, with no way to visit family or to help out around the house, to conduct any of the other normal household business. Their time in the hotels is not their time at all to the point where the actual time they spend on a run (a round trip) can be multiple days and only a part of that time is compensated. A recent text I had from my relative currently with a railroad gives the railroad's perspective quite clearly during a Zoom meeting - "Y'all hired out to work 24/7/365, we are going to make good employees out of you all." I am part of a fourth generation railroad family. Three of the last four generations, beginning in 1942, have worked on the railroad. It skipped me because I went a different direction. EDIT: I forgot to mention that the ultimate goal of the railroads is to decrease train crew size to 1. Each of these trains will be driven by a one-man train "crew". A couple decades ago normal crew size was three - engineer, brakeman, conductor. Brakeman positions were eliminated leaving the engineer and conductors rolling the trains. The railroads were supposed to be using PTC, positive train control a long time ago but were able to dodge the upgrades to track and infrastructure in the typical American corporate way by delaying using a standard set of excuses. PTC would bring about a single man crew by automating almost everything. That would mean huge loss of jobs and a huge investment in infrastructure upgrades since many thousands of miles of tracks have speed limits that are in place due to trackage that can't handle higher speeds safely. There's a lot going on here and it's always deeper than one single argument.

joshstrange 4y ago

I think the other important piece of data you would need to find is average train length over that same period of time. Fewer derailments could have just as much or a larger impact if the train length is growing.

foobarian 4y ago
Right, given same car count and derailment rate, 2x train length could appear as 1/2 the derailments.
- AnimalMuppet 4y ago
  Depends on if the probability of a derailment is K per train mile, or K per car mile. And I suspect that the answer is "some of both". When a car has a mechanical failure that causes a derailment, that kind of derailment is going to be proportional to the number of car miles. Some other causes (misaligned switches, say) will be proportional to the number of train miles.
mediaman 4y ago
I did find a GAO study that said that trains increased in length by 25% since 2008. But the number of derailments dropped by more than 25% during that time, so the probability of a given car being involved in a derailed train still decreased. There's been other coverage of the ironically named "precision scheduling" in the rail industry causing a lot of labor problems, because it makes for nightmarishly unpredictable schedules for workers. They never know when a train is going to leave until it is full, so after working a shift, workers stay in a hotel waiting for a return trip that they won't know about until 90-120 minutes ahead of departure. I am guessing that the claims around increased derailment may actually be about frustration with the work environment.
simulate-me 4y ago
The average length isn’t enough because the likelihood of derailment might be a function of train length. It’s possible shorter trains have gotten safer and monster trains have gotten more dangerous due to their increasing length.

pas 4y ago

Is there comparable data on other countries? What is this a proxy for? (Track upkeep problems? Trains are too fast for the load? Signaling issues? Mechanical issues?) Rail safety seems a lot less complex and pretty much as regulated as aviation safety, how come there aren't just ~1-2 per year? Is it just because fatalities are luckily fewer (due to the nature of these derailments) so no one really gives a shit?

redshirtrob 4y ago
I don't know if there's comparable data, but I can tell you the mining companies in north west Australia would routinely run 300+ car trains without incident. But these were very much a straight line from the mine to the port without much civilization in between.
r_hoods_ghost 4y ago
Well in the UK there were 11 derailments total in 2020-21. The UK has roughly ~10,000 miles of railway and the USA ~150,000. Probably not a good comparison tbh. EU wide would probably be more appropriate but I couldn't find anything. https://dataportal.orr.gov.uk/media/1999/rail-safety-2020-2021.pdf https://dataportal.orr.gov.uk/media/1999/rail-safety-2020-20...
- hardware2win 4y ago
  Indeed, 5 derailments per day sounds like a lot
brnaftr361 4y ago
It's just the forces of the trains. When it's not a derailment it's breaking them. They're doubling the running room for slack action^, for one. That's the least of the concern, but there's more potential for breakage because there's more potential for greater differences in velocity from one end of the train to the other. Topology also matters, previously consists here average 135 cars, but range from 115 to 150, they're 53' each. Usually they'll have 4-5 motors at 74' each. Suffice it to say, depending on the territory there can be huge force differentials generated which can break trains or rail, think cresting a grade on undulating hills. And you can "stringline" trains too, literally pull cars off the tracks, and it's exactly those forces that'd do it. And then there's the human element. Nobody knows when they're going to work, how long it's going to take, what kind of bullshit they're going to run into. You can get stuck for long periods of time away from home terminal, and the Class 1 doesn't care. And there's things like rules on how to walk. The environment that's been constructed isn't meant for human consumption and it's taking a toll on the workforce. There's reams of other shit to whinge about. And the derailments always matter, it's a total clusterfuck. They're not easy to fix, you've got to "bad order" and set out the car, at least by regulation. This involves rerailing, finding a set-out track, and making all the moves to get everything back in order. You've also got to wait for carmen (who are understaffed) and management, which depending on the route, could be a long sit. If it's bad enough it'll shut down traffic for a couple of days, requiring significant diversions around it while the Class 1 pays out the ass for a third party to come clean it all up and rerail shit. None of the current infrastructure is built to deal with these train lengths, either, so everything moves slower. The fact the Surface Transportation Board had a hearing with the Class 1's, I think, is a pretty good indicator they're about as interested in moving goods as GE is in manufacturing. ^: Each coal car has two draw bars with +/- 2-3" about the same for shipping container cars (from eyeing it) of run built in. This varies with car, auto racks have way more run out. Opposing forces in the trainline will pull it apart, the more run to build the faster the opponent acceleration... This typically results in broken knuckles. But it is possible to just push cars off the rail with these forces, which is exacerbated by the doubled length.

scythe 4y ago

The gross number of derailments isn't very meaningful, what you need is the derailments per train-miles-traveled, analogous to the collisions per vehicle-miles-traveled that is used for road safety. If we're running fewer trains, obviously we should expect fewer derailments. As some other commenters have noted, it might be even more useful to get the total weight of derailed cargo, per (mass times distance), I suppose.

temp8964 4y ago

foxyv 4y ago

I think the author was trying to explain the risks they are taking and saying that it's only a matter of time until we start to see disasters similar to the one in 2004 but 100 times worse. https://cen.acs.org/articles/82/i27/TRAIN-DERAILS-CHLORINE-LEAKED.html#:~:text=Two%20freight%20trains%20collided%20early,and%20killed%20two%20more%20people https://cen.acs.org/articles/82/i27/TRAIN-DERAILS-CHLORINE-L....

wheelerwj 4y ago

a few other people have made a few suggestions about interpreting the data, but i would also like to add that the writer is specifically citing "Monster Trains" and that may or may not affect the interpretation of data. still, awesome that you went to check the source.

scottlamb 4y ago

> According to the Bureau of Transportation Statistics, there were 1,056 train derailments (both cargo and passenger) in the US in 2021 The key word here might be "US". The letter writer is specifically talking about Iowa; are state-by-state statistics available? If you've never spent time in the Midwest, you probably have never seen a train that's a mile or longer. Endless cars of agricultural products, often high fructose corn syrup. I saw this often growing up in Iowa, never in California. In the Bay Area, freight sometimes moves along the same tracks as Caltrain at night, but the trains are tiny in comparison to those of Iowa. For a while, I think trains in Iowa were limited by how long a train was allowed to block a roadway (IIRC 10 minutes). I think trains were sized so at least in theory they could comply with the law when at speed. The law might have changed since, and/or maybe there aren't as many at-grade crossings anymore. (A painful intersection in my home town recently got grade separation.)

dhosek 4y ago
Other than a few years in L.A., I've lived my whole life within a mile of a rail line. It’s only in the last month that I have ever seen a mid-train locomotive. Something is definitely changing.
- brnaftr361 4y ago
  They started this process a long time ago. 2018 was when they first started talking about it, locally they did the first pilots about 2019 if memory serves.
AnimalMuppet 4y ago
Out here in the west, I have seen longer trains in the past few years, but I've been seeing mile-long trains for years. I never saw trains with more than 150 cars before the last few years, though.

TimPC 4y ago

Derailments could be getting worse even with fewer of them. The author of the letter is talking about the trend towards monster trains which suggests fewer trains being run. Is the number of derailments per train going up? Is the impact of the larger trains derailing leading to more overall impact because each individual derailment is larger even though there are few of them? If I run half as many twice as large trains I'd expect fewer total derailments even if these monster trains derail more often than the more normal trains of yesteryear.

exmicrosoldier 4y ago

it's easier to have less total derailments, but higher numbers of cars derailed. Put every train car in the country on one train and derail it. Now 100% of the capacity was derailed, but it's an incredible success, because the number of train derailments shrank from 1056 to 1! So much safety!