Comment by swiftcoder

Given this keeps happening in Spain, maybe we'll need to call this "Normalization of normalization of deviance".

Humans never learn.

You also miss appointments and are late if you are dead due to a train crash.

Jenga effectively. 3 others and Bob spent time destabilizing but Bob over here pulled the block before it collapsed. It's his fault.

(Un)Fortunately, finding someone who is "at fault" generally helps neither in investigating root causes nor in preventing similar problems in the future.

Usually it's detrimental, hence (criminal) investigations of personal liability in such cases are often only allowed after systemic investigation has concluded.

In court, you are allowed to withhold information that makes you look bad, which is simply lethally unhelpful when we want safer trains.

I don't mean to absolve people for negligence, but when margins are very small and paychecks may have become a necessity to survive on this planet, sometimes within the complex process the collective behaviour gets pretty weird.

There are not that many people who (care to) see the big picture(s) and the consequences their semi-small actions may have in the long run.

Also the train driver could have pushed back, in the radio recording with the dispatcher, he tells the driver to "proceed normally", the driver asks "proceed normally?" and the dispatcher confirms it. Here's were I would expect that better training or more availability of information to the driver would've made him push back on that decision.

The automatic system will fail at some points so manual dispatch is needed, which is how this all happened

Not OP but I'm imagining we'd ask the question why the "dispatcher ordered manual bypass ignoring the rest of the guardrails". They're saying it's all slices of cheese in a system that failed here.

It seems that one of the train conductors did know that it was not the procedure and tried to warn of the problem.

Trains take a long time to stop even at moderate commuter train speeds because they are so heavy, let alone high speeds of 300km/h+.

At least one of the failures here is that the dispatcher authorized such a high speed, in other systems I am familiar with, generally the procedure with a signal fault is to crawl at a speed such that the train can stop in line of sight distances.

In fact, that is the Spanish protocol, which for some reason was bypassed by dispatch even after the train driver had questioned it:

> Under normal protocols, the driver would be expected to continue moving at a maximum of 40km/h according to their line of sight, and not at normal speed on a track designed for speeds of up to 200km/h.

If a stretch of road continues to have frequent crashes even though a sign is present, it is worth examining why that sign alone is not solving the issue.

Air travel is as safe as it is because we don't stop the investigation at "pilot error" but instead ask why the pilot made the mistake, and why a single mistake was able to lead to loss of life.

I don't think trains are the same as cars. Trains are supposed to run at high speeds under normal circumstances and the system is supposed to keep everyone safe. It clearly said they wouldn't be able to drive over 40 km/h if they had to operate by line of sight.

Cars otoh don't have someone overseeing everything and directing traffic. They have to operate by sight.

I think it is wrong for you to make an analogy to someone driving a car.

I don't mean to pick, but it seems like you have near zero knowledge of how train systems work.

You are 100% dependent on your safety and monitoring systems because almost every problem that could come up is beyond the horizon of your human senses. Stopping distances are in thousands of feet to miles depending on the train size and weight. Add any curves and that's far beyond your visibility.

Then you add in management complaining that you're too slow and you need to do things faster and cheaper or you'll lose your job. Things turn from solution culture to blame culture.

Then you add in false errors and warnings in a system that teach you to ignore them and an upper management that doesn't fix them.

Then you have the day like hundreds of others where warnings are going off, management is yelling, but this time there is a stopped train around the corner and everyone dies.

This is a systemic problem.

I wish this were true, but there is basically no consequences for drivers in the situation you described. The most egregious example I can think of near me was in 2022 when a kid was crossing a flashing crosswalk with his bike near Tacoma, and truck had stopped to let him cross. A woman ignored the crosswalk signal and used the adjacent lane to drive past the stopped truck and hit the kid, killing him. The woman was never criminally charged. She didn't even experience any consequences for her driver's license.

https://komonews.com/news/local/13-year-old-boy-killed-while...

> Someone was driving the train who couldn't estimate the stopping distance relative to speed.

The stopping distance for a train this size and this fast is around 2km. Good luck estimating that by sight. There is a reason the normal protocol has a 40km/h max speed for traveling according to line of sight, which the driver was told wasn't needed.

Indeed, car crashes tend to just blame one (or more) of the drivers. Nobody excuses a foolish driver with systemic excuses.

On a surely unrelated note, car travel is by a substantial margin the most dangerous mode of transport most of us ever use, and the most dangerous activity most of us routinely engage in.

"The cause", singular, shows you're not getting the the swiss cheese model still. Trying to find a singular cause is a big mistake which prevents other causes from being addressed.

Every hole in the cheese slices, including operator error, is important. And operator error is an important kind of error in particular, not because we need to prosecute somebody but because many forms of operator error can be mitigated using better system design, but you won't figure that out if you blind yourself to operator errors.

PZB is from 1934. Arguably always prioritizing deploying the most modern system on the highest profile lines, is what prevents us from properly rolling out the basic safety systems everywhere.

I'm pretty sure we had modern computer systems around 1990.

> Trains are still not able to assess properly where they are, even using satellites and inside train sensors. Of course there is research on that, but it's still an unsolved problem

Throwing up our hands about how hard a problem this is seems like a ridiculous position to stake. If we can design self-driving cars, that operate safely in far less controlled environments, we can figure out the linear offset of a large vehicle traveling along a physical rail running along a fixed pre-determined path.

Yeah and that's fucking ridiculous and a damning indictment of the glaring incompetence of train companies.

If I can have reasonably accurate gps readings of my phone while I'm on the train, there is zero reason for train companies to not have systems that are at least as accurate.

Yes they are, those balises on tracks are the standard technique for both metro and mainline signaling systems at this point. Those balises, often combined with radio communication, are how fully autonomous trains have been operating for decades. (It's possible to use things other than balises, like track radio loops as in LZB or more recently beacons or LIDAR, but balises are standard and pretty affordable).

> Trains are still not able to assess properly where they are [...]

Isn't the ability to do so the basis of operation for e.g. the German LZB train protection system, which has been around since the 1960s?

If you mean that train protection systems don't (exclusively) rely on self-reported positions, your statement makes a bit more sense, but even that is changing with e.g. newer ETCS levels which do rely on self-positioning more and more.

Would you mind sharing some more of the why it is still an open problem? (I know I know little about this subject.) I understand satellite positioning isn't definitive and did not suggest it should be the /primary/ risk mitigating factor. But even as a tertiary or lower risk mitigating factor, this would have been a worthwhile investment. (Sorry for the bleak introduction of economics.)

And some problems are just _open_ to outside looks. My assistant tells me that the room for error in trains on one dimension ('left - right') is mere 10s of centimeters. I understand that problem. But what track you are on, isn't that obvious to the human in the cabin? (Again, simplifying, not flaming. Production ready safety is _hard_.) So could a vision based solution help count tracks from left to right? (Assistant: "US PTC, on about 54,000 route-miles, uses GPS for the lead locomotive plus crew-entered track information and switch position".) So you use the dimension that matters most in crashes (forwards / backwards) from satellite, augmented with vision or human based left-right positioning. Some trains already have sensors to sense things like leaves on the track to adjust speeding and braking... So the sensor based approach is already there.

Perhaps I am read as suggesting that the problem is easy, and that is not my intention. But for cars this was solved decades ago for routing and years ago for self-driving. I use that daily. Flying is organised with a very clear 6 sigma approach. Crashes are nearly always situations where events / risks stack. Trains (in Spain)...?

Tl;dr - Safety is a function of (human; process; technology) and there is clearly something lacking for trains, when compared to other types of transport.

GPS-based PTC absolutely does what you describe. It has long been solved.

(The following is speculation:) The issue is that there was still a fallback system operating (ASFA), so they likely assumed that there is still a safety system running, while actually it needed to be ignored, because due to virtual blocks of LZB there is then more than one train in a ASFA block. This means that the next signal shows you "clear", so you accelerate thinking you are back into normal operation, but there is actually a train in front of you, just not in the next block, but in your very own block, which is a situation that is not supposed to happen. So the issue is kind of different train control systems interacting disagreeing with each other. Ironically it would have been actual if there would not have been another safety system, because then they would have known, that they need to operate manually right away.

There are audio recording from this incident where you can hear the "train orders" being given and acknowledged.

They were the wrong orders apparently.

> Why is that?

What exactly?

A conductor always just gets a movement order. They are not able to assess whether the track is clear, that's the reason why the signalling system exists in the first place. How they get that movement order doesn't necessarily tell what the security steps were to back it.

Of course an attentive conductor can notice and mention there something being weird ("Doesn't the other train normally arrive first?"), but that's a rare lucky occasion, nothing to be relied on. And they just need to swallow the answer ("No, you go first today...") anyway.

> and conductor just wouldn't move unless green

The conductor won't ever move unless they get a movement authority, which most of the time is a green signal.

Many train systems have a trip-stop on signals that will automatically activate the trains's emergency brake if one passes a red signal.

It cannot be bypassed, if an operator must go through a red they have to manually reset the e-brake on each car once it passes the signal.

So that's why my train slowed to 40kmph when the Norwegian Telecom died for a few hours 3 weeks ago.

Depends. For example German Zs1 (https://en.wikipedia.org/wiki/Railway_signals_in_Germany#Zs1...) would give them the full clear outside of stations.