Hacker News new | past | comments | ask | show | jobs | submit login
How I Almost Destroyed a £50M War Plane and the Normalisation of Deviance (2016) (fastjetperformance.com)
473 points by dddddaviddddd 60 days ago | hide | past | web | favorite | 181 comments

It's really nice to have a term to pin to a vague concept that had been floating around in my head for at least a decade. I've come across instances of this repeatedly but never knew quite how to describe it properly to others. It's a treacherous slope where the first step is usually completely un-noticed and by the time people do notice it is way too late. We've always done it like this is the oft heard excuse but on closer examination that usually isn't true. Typically at some point in the past there was a much tighter regime and the present day situation would have definitely not passed muster. But by inching towards it with very small steps critical thresholds can be exceeded without notice.

Thanks for posting this. It will help me professionally to formulate my opinions on such situations more clearly and will serve as a reference point for future discussion.

I think for every 100 people who talk about Vaughn's "Challenger Disaster" there are maybe 2 or 3 who actually read it.

Like many books in sociology it comes at the problem from multiple perspectives (in different chapters) such that you can't really hold the author to any one clear conclusion.

Big picture, the space shuttle was "unsafe at any speed" from the very beginning. It was estimated that it had a 3% or so chance of blowing up on any flight which turned out to be just about right in hindsight.

One point of the book was that the Space Shuttle had many things wrong with it: they had a committee who's job it was to review all the safety risks and decide which ones they could accept. "Normalization of deviance" was not a bottom-up matter of bad practices, but a formal part of the organization.

If things had gone a little different with the O-Rings then perhaps the "Challenger Disaster" wouldn't have happened, but then it would have been something else. I remember in the early days the thermal tiles were falling off, but after they had done a few flights they convinced themselves it was OK. It was the Columbia disaster that proved that they weren't OK and couldn't ever be OK.

"...they had a committee who's job it was to review all the safety risks and decide which ones they could accept."

Having a committee dedicated to assessing risks seems like a good idea. Having a group consciously deciding between risks you can live with, risks you have to live with, risks you can mitigate, and risks that stop everything is a fundamentally good idea.

The problem is taking risks without thinking about them, or taking risks without knowing exactly why they haven't killed you before. And even more unfortunately, this is all on an experimental project That Cannot Ever Be Allowed To Fail, Ever™.

A agree that having such a comittee is essential. But (at least in hindsight) the risk assessment for the space shuttle seems very strange. The only rocket ever built that requires human pilots is also the least safe human-carrying rocket ever built, both in track record and design (no launch escape system, no abort possible for large parts of the ascend).

In a comittee maybe you can justify the human deaths as nessesary risk, after all people die in traffic all the time. But that ignores the political dimension. The Challenger Disaster was a giant blow to the entire shuttle program: The space shuttle was only funded because of support by the military. With the entire fleet grounded for 2 years and much needed performance improvements canceled that support vanished (see for example: the last DoD shuttle mission was in '92).

It seems like Apollo had a worse track record even if you consider 13 to have been a successful mission rather than an incredibly near miss.

Do committees who do such reviews equally ask themselves - if this does fail - can we live with it and justify our initial risk assessment and stand by it?

So often - risky decisions are made - many pass without incidence and yet, problems happen. Often such risks are reviewed in hindsight and the weight of the realisation of risk that pass into actuality. Though when they don't fail, nobody passes a second thought and over time, such risks get normalised.

That's a fair point, but the fundamental fact is that at some point, someone/something/some process needs to accept a level of risk, or perform some type of cost/benefit analysis. Sometimes the analysis is straight forward or easy to criticize. For example, when the costs are high, and benefits relatively low. It's super easy to criticize the shuttle disasters, because the cost of cancelling the launches are so obviously lower than the lives of astronauts (at least in popular discourse).

The issue is that all mechanisms to do so are susceptible to some variant of the problem you highlighted. As so highlighted in the OP's submission, individuals can make the exact same mistakes.

I feel that a lot of things comes from the fact that we are constantly encouraged to over-quantify risk. That is we take things where REALLY we don't have a good grasp of the real numbers, and then make something up. If we're good, they'll even be in the right ballpark. If we happen to choose to follow the course of action, and everything is working "just fine", we go back to our mental risk model, which we always knew was kinda made up, and then say "well, I guess I guessed wrong", and assign a lower risk.

Yes, it does get down to risk factors and the parallels with the reinsurance/insurance market with actuaries, is an aspect not lost upon me. As we know you can't put a price on a human life - but they do in the end.

When I've been involved in such things, the phrase "Dear Mister Coroner" has come up regularly.

That's a really good summary / perspective. The more I read the Challenger Launch Decision, the more I was convinced that mistakes like the Challenger are unavoidable. We can try to reduce the risk, but given that we're always operating with limited resources, we just can't eliminate all the risks (from the design and implementation as well as the decision making and operation).

But "normalization of deviance" is also partially how the science / the engineering works. The science and engineering is pushing the boundary and turning what used to be unknown into known and tested through experiment / experience. Learning from experience is fundamentally how the science and the engineering work. So it's really difficult to tell when it's a normalization of deviance and when it's a normal scientific/engineering process.

Things could have been done differently with the space shuttle.

Note that the crew of the Challenger wasn't killed by the "explosion", they were killed when the vehicle compartment they were in hit the water.

If the shuttle had an escape mechanism the way every other manned space vehicle had, they might have survived. Had they built in an escape mechanism, there wouldn't have been any room for cargo. From a safety perspective launching cargo + passengers on the same rocket was a mistake.

Also the shuttle was designed to do a military mission that it never actually did, which would be launching, orbiting the Earth once, and then landing at an airstrip close to the launch point. That requires a large "cross-range" (ability to turn when reentering the atmosphere) which in turn requires larger wings which are more vulnerable to something like what happened to Colombia.

One of the reasons I really like Challenger Launch Decision is because it does a very good job of avoiding the hindsight bias as much as possible, by getting the readers to understand the perspective and the cultural norm of the organizations and participants, and tracing the decision making process itself and influences on those, instead of technical decisions themselves that were / might have been wrong in hindsight.

In hindsight, we can see mistakes in terms of specific design and technical decisions. But I don't think those are the biggest lessons from the disaster. We will continue making such mistakes, since technical misjudgement itself will always happen. I think Challenger Launch Decision's main point, that the organizational structure and culture have large impact on individual decision making is the lasting lesson. The book's closing sentence is excellent:

"The lingering uncertainty in the social control of risky technology is how to control the institutional forces that generate competition and scarcity and the powerful leaders who, in response, establish goals and allocate resources, using and abusing high-risk technical systems".

> "The lingering uncertainty in the social control of risky technology is how to control the institutional forces that generate competition and scarcity and the powerful leaders who, in response, establish goals and allocate resources, using and abusing high-risk technical systems".

Seems to me that the relationship between Boeing and its various 737 MAX customers could be seen in a very similar light. Abusing high-risk technical systems as a tool to extract resources to cover for artificial scarcity in budgets or market expectations.

You seem to be implying that something near complete safety is practical with enough work. I'm not sure this is the case. A big problem with all endeavors, and not just space, is that things that work (or don't work) in theory somehow often end up doing the exact opposite in practice. This is why things become more safe at scale: you get to see those abnormal conditions play out in reality, and work to prevent such failures in the future. In rocketry we're still babies. We've launched, in total, something like 8000 rockets - hundreds have failed. And we've only launched around 600 humans to space on a small fraction of those 8,000 launches.

Just consider something like a software launch. No matter how much effort you expend trying to make things go perfectly, or how many months of beta testing you engage in - something's going to go wrong. And that's in a nearly completely deterministic environment. Space travel involves countless external nondeterministic factors. Another big difference is that in the software world we get instantaneous scale - up to millions of instantaneous independent trials. Spaceflight advances one costly trial at a time. In the entire history of rocketry, we haven't even begun to approach scale and so all the pretext of safety we could ever elevate is just that - a pretext.

It's for this reason that being open about risk and danger is important. It's also why I think aiming for complete safety can be paradoxically harmful. It creates scenarios where open discussion of danger and risk becomes taboo since it can result in negative consequences. The most critical factor in all of this should be informed consent. Making risk taboo goes directly against that.

Disagree. It's possible to launch software without things going wrong - I've been involved in one. Certainly NASA was not taking the absolute safest approach possible - they had people internally advocating for safer practices (which is not to say that there weren't legitimate other pressures).

what you are describing is not normalization of deviance, but rather proper risk management. that they made an incorrect assessment is a different issue.

normalization of deviance is when a process isn’t followed. when the reality of the process is different than the understood, risk-managed, process.

People writing about normalization of deviance in healthcare and aviation describe something very different from what is in Vaughn's book.

In this interview, Vaughn said "Social normalization of deviance means that people within the organization become so much accustomed to a deviant behavior that they don't consider it as deviant, despite the fact that they far exceed their own rules for the elementary safety", which seems to me to be pretty close to how people in healthcare and aviation are using it.


Shifting of the Overton window is the closest concept I know of, but it still seems too specific to fit this pattern. https://en.wikipedia.org/wiki/Overton_window

Yeah, another expression is "slippery slope", but I feel the term "normalization of deviance" has a lot more gravity to it, a formality that the other expression is missing. I've read a few articles now using this phrase, it's often about the space shuttle or aircraft, and that may just be the mindset that the software development industry should be striving to equate themselves with.

I think the phrase "defining deviancy downward" more closely captures the concept than "normalization of deviance." The slippery slope of redefining deviance to have lesser thresholds than the prior iteration, and it takes a series of those iterations to truly end up in a situation where the new normal would have been utterly out-of-bounds compared to the original.

Normalizing deviance is more akin to deciding that things that were obviously out of the norm are now accepted as a new normal without the gradual degradation, at least in my way of thinking.

I think one term for it is Creeping Normality.[1]

[1] https://en.wikipedia.org/wiki/Creeping_normality

Complacency seems to fit, but it doesn't sound interesting enough for people to read a story.

Also known as the boiling frog.

I would say this is the engineering incarnation of the Broken window theory of the early eighties, if you want a term which is much more known: https://en.wikipedia.org/wiki/Broken_windows_theory

FWIW, that’s a very politically loaded term. I would recommend avoiding it unless you intend to bait people into an unrelated conversation.

I also think it’s not a very accurate analogy.

It’s unfortunate that “broken window theory” has become synonymous with the awful and racist police tactics introduced in NYC.

Broken window theory is real: imagine a random person in spotlessly clean city versus a city where there is trash everywhere. More people would litter in the latter city, I don’t think that is too controversial. Thus a run-down neighborhood with lots of broken windows will normalize behaviors that would be stigmatized elsewhere.

The unfortunate part was when NY adopted “broken windows policing”, which is to go into those neighborhoods and make lots of arrests for petty crimes, as if replacing all the broken windows might somehow reduce crime. Of course this is pretty dumb, since entrenched normalization of deviance (decades of bad/discriminatory policy) can’t simply be fixed by zealous application of punitive measures.

> Broken window theory is real: imagine a random person in spotlessly clean city versus a city where there is trash everywhere. More people would litter in the latter city, I don’t think that is too controversial. Thus a run-down neighborhood with lots of broken windows will normalize behaviors that would be stigmatized elsewhere.

Citation needed. There are plenty of counterintuitive things in sociology; the broken window theory sounds plausible but that in no way implies that it's true.

Not citations but some empirical anecdata: visit SF, then visit Singapore. Or, notice how a relatively clean room in the house stays clean versus a somewhat messy room that degenerates into chaos.

I think few people would dispute that places with more litter have more littering!

But surely that's because littering causes litter?

>a run-down neighborhood with lots of broken windows will normalize behaviors that would be stigmatized elsewhere.

Of course, but...

>... as if replacing all the broken windows might somehow reduce crime. Of course this is pretty dumb

Calling it dumb seems to contradict the first quote above... If broken windows normalize bad behaviors, less broken windows would seem to prevent normalizing those behaviors.

>entrenched normalization of deviance (decades of bad/discriminatory policy) can’t simply be fixed by zealous application of punitive measures.

Why not? Punishing deviance prevents the normalization of deviance.

Curious, what would you think of broken windows policing as applied to high-crime white neighborhoods?

Did you read TFA? Do we ensure airplane safety through aggressive punishment of all minor offenses? Why not? Perhaps any real solution involves something more thoughtful than attacking obviously visible symptoms of the problem.

No contradiction. You can have one system, one small part of that system being no broken windows, where windows are commonly broken. You can have another system with unbroken windows. Repairing the windows and punishing people for breaking them won't magically transform the first system into the second. In fact, it might actually make things worse, depending on how it is implemented. For example, if it's already common to break windows in an area for a wide range of reasons and you start fining broke people money they can't pay and throwing them in jail... you haven't saved the windows, just caused a lot of damage.

What's inaccurate about it? My understanding is "Things looking bad allows people to assume that worse stuff will be tolerated". The way it was enforced seems to be where the political baggage comes in, but the fundamental theory shows up all over the place.

If not identical, it at least seems closely adjacent.

I'm curious to hear your perspective; what's wrong with the analogy?

I understand the 'broken windows' theory to be more applicable to leaving things that need to be fixed unfixed which slowly normalizes the state of things being 'messy', the concepts are peripherally related but do not at all address the same thing.

> leaving things that need to be fixed unfixed which slowly normalizes the state of things being 'messy'

Which is exactly what the author did with the landing gear of his aircraft, no?

No, the issue there was normalizing a specific, dangerous workaround. Not sloppiness in general.

This was a really good reminder of engineering wisdom.

Over the year I find myself and my team (civil-structural engineers) doing even more and more riskier designs. Things we wouldn't have considered OK during our early years. We do get review and approvals from seniors, but maybe their acceptance is also "normalisation of deviance" as described in the article. Sharing it with my team.

Would it be a case of the norm was to over-engineer (thinking victorian times mostly as examples of structures standing and still in use) compared with today in which we have a finer insight into the maths, which allows us to engineer just-enough.

Or is it a case of market/management pressure pushing the limits that is the driver? With those pushing the limits getting a competitive edge - that draws others into competing in the same way.

Can you give (, perhaps anonymised) examples? This makes me wonder which of the new breeds of skyscraper are actually a lot less safe than perhaps older ones. I have read the same with cruise ships as we have seen with a disaster or two on those

When it was being described the checklist a pilot has to go through before taking off, it resembled, though quite crudely and without life or death implications, the lengthy checklist the developers at our company had to follow back when we deployed everything manually. We had synchronization steps to pull modifications, couple compilation steps, we had to check if configuration files were added in order to accommodate extra services, there was cache mechanisms that had to be cleared, after that we needed to check for database migrations, then restart the systems, run the little tests we had and manually check if nothing broke. It happened often to see code conflicts, 40 to 50 with varying difficulty every release. It was insane and never quite worked, we always got weary of release days and never wanted to go through them as we didn't trust the system (with good reason).

Automation and good practices brought back confidence, so I was wondering two things.

First most obvious, how come do pilots trust a system that has these layers of manual security checks with their lives. And second, can't it all be automated? I understand aircrafts must have independent modules and every level of integration implies extra risk, but at least in security critical tests, automation would come a great way in ensuring crew safety, wouldn't it? Is it too hard? Systems are not flexible to implement automatic security checklists or is it just a culture in the field to check it all manually?

I highly recommend “The Checklist Manifesto” by Atul Gawande. One of the points that Gawande discusses is that checklists cannot be too detailed. He discovered this in working with the WHO when their first surgical room checklist attempt was a complete failure because it had too many steps (chapter 5, the first try). In chapter 6, he then revisits flight checklists and writes:

“Some tone after that first miserable try, I did what I should have done to begin with. I went to the library and pulled out a few articles on how flight checklists are made. As great as construction-world checklists seemed to be, they were employed in projects that routinely take months to complete. In surgery, minutes matter. The problem of time seemed as serious limitation. But aviation had this challenge, too, and somehow pilots’ checklists met it.”


“It is common to misconceive how checklists function in complex lines of work. They are not comprehensive how-to guides, whether for building a skyscraper or getting a plane out of trouble. They are quick and simple tools aimed to buttress the skills of expert professionals. And by remaining swift and usable and resolutely modest, they are saving thousands of lives.”

> First most obvious, how come do pilots trust a system that has these layers of manual security checks with their lives. And second, can't it all be automated?

It's one thing to have a script check the return code of a test suite. Often times aircraft checklists include things like physically verifying that something is present and moving in the correct manner when commanded to do so. Obviously you can add sensors, but then how do you check that the sensor is working?

At the end of the day, most of the manual checklist items are still for a human to look at or listen to a thing and make a judgement call.

> First most obvious, how come do pilots trust a system that has these layers of manual security checks with their lives. And second, can't it all be automated?

The trust comes from understanding the system and associated risk that contribute to conditions where the system is out of normal operating range. Pilots learn that doing simple checks on parts of the system that are likliest to contribute to abnormal operations is life saving in the long run. E.g. may seem simple that control locks werent taken off but it happens and it's why we check the controls prior to take off. It's why we put the engines to fill throttle to make sure oil pressure is as expected.

Anytime there is an accident many pilots will read the reports and add an additional check or reemphasize a current check that they need to continue to use.

As far as automation. It could fix it. But it increases system complexity so there will be this valley where for a while the automated system isn't much better than normal checks, infact might be worse due to over reliance. Also logic in software or hardware can lie to you in ways that mechanical realities of moving control surfaces, checking oil level, checking tire pressure can't be beat (at least with a cheap solution). This is why the FAA for critical components requires multiple systems doing the same thing and concensus algorithms.

Recently I bought a lemon for a motorcycle. The ECU started giving me error codes twenty five minutes into my first ride. Engine started cutting out. I couldn't start it. On the other hand in a Cessna I could lose all electrical power (assuming magnetos are still operating) and still operate the plane just fine, although without radio comms. The Cessna is from 1972, while the motorcycle from 2017. Granted I could've checked wires or inspected more carefully but non the less, the 2017 motorcycle is a more complex system.

The motorcycle engine is LESS reliable than a well maintained Cessna. But I suspect the motorcycle was not well maintained, perhaps not well designed, and you should not have been driving it.

Short of a teardown I don't see how the buyer of a second hand motorcycle would have a chance at spotting such a fault without driving the bike. And even a teardown might not show the fault, in fact could make it worse.

That's equally true for an aeroplane engine - the difference is that no pilot would skip performing maintenance at the required intervals and there are maintenance logs etc. in place to enforce this. What makes the Cessna engine more reliable than the motorbike engine in practice isn't a simpler design but a better maintenance and operation culture.

Sure. But the accepted norms for buying second hand aircraft versus second hand motorcycles are quite different as well, which is why it makes no sense to say that the buyer should not have driven the motorcycle. The accepted norm for cars and motorcycles is that you pay the seller and you drive them home, not that you inspect the detailed maintenance log or do an on-site teardown. Likely the seller was aware that this bike had problems and simply did not tell the buyer, something that is close to fraud, and on top of that this placed the buyer in a potentially dangerous situation, especially on a new bike.

When I buy cars I do a pretty thorough inspection and an extended test drive. Even so I once ended up with a car on which the clutch pedal would not return once every few hundred strokes or so, something that by chance did not happen even once in a pretty long test drive.

There is a middle ground, and I think people should demand more in that respect. Yes, cutting engine power is more hazardous in an aircraft than in a motorcycle. So the aircraft should produce a clear warning when takeoff power is applied but control surface locks are sensed as present. Or when takeoff power is applied and the rudder/ailerons/elevators haven't been fully deflected since startup (indicating controls haven't been checked). Or when the altimeter reading has a gross disagreement with the GPS altitude, etc. We should expect more from our automation without letting it get in a position where it can compromise safety.

I know this is a bit of a tangent, but the motorcycle didn't happen to be a Royal Enfield did it? They seem to have lots of wiring harness abrasion issues.

Moto Guzzi V7. Might end up being simple ECU reset or fix. I’m just glad I was tooling around in a neighborhood getting comfortable, and letting the failure percolate. Otherwise would’ve been much worse had I gone straight to the highway.

I flew small airplanes for fun, nothing close to a fighter jet but I'll try to answer.

> can't it be all automated

No. Some of it certainly could, notably gear related like engine checks as in your car. And yes current systems are not very flexible, the strong bias towards "tried and trusted" does slow down the arrival of fancy electronics.

But checklists also include confirming you have weather information, making sure your surroundings are clear of danger, that your fly plan is defined... This is more about checking the human component and can't be automated.

> how come do pilots trust a system that has these layers of manual security checks with their lives

This is the best we came up with so far. Manual checks also give you a sense of control.

> First most obvious, how come do pilots trust a system that has these layers of manual security checks with their lives. And second, can't it all be automated?

Conveniently the word "deviance" has multiple meanings, two of which apply here.

The article uses deviance in the sense of "deviation from the behavioral norm" (where that norm is a external process). I completely agree with your assessment that these processes can become cumbersome and a barrier to doing work (at most big companies the process becomes the point and dominates actually getting work done). Automation can go a long way to improving that, both in reducing cynicism and letting people spend their time on the stuff that actually moves the company forward.

However the manual checks in aviation also serve to find unexpected deviation in the shape or performance of the device. The reason the first officer walks around the aircraft is not for exercise but because she may see something "out of the ordinary" for which no test exists -- something that should be straight looks subtly out of kilter, indicating that, say, a bolt on the wheel has come loose and under load the wheel tilts slightly. This kind if check is harder to find in software which uses many fewer humans intuitions, which is why we have code stylebook and code reviews.

Pool of hydraulic fluid near the landing gear wheels is one example of such an out of the ordinary thing for which no test exists. Presumably the hydraulics would still work for take-off, a wheel assembly might end up not being stowed properly or it might refuse to descend on landing.

The example in the article is an interesting one because to an outsider it clearly looks bonkers and yet, to an experienced insider it might look workable.

>can't it all be automated?

Yes, but two things. Airplane technology advances happen slowly but surely. Small changes can have big consequences when flying. So rightfully any changes are thought about for a long time until they're deployed in practice. And so far airplane safety is miles and miles above any other mode of transportation, so it would appear that this cautious behavior is working.

But even if we had fully automated "turn key" planes checklists do something for the pilots. They force pilots to understand their instrument panels, where the controls are and how the plane largely operates. Things like the start up procedure can vary really radically between planes even with the same types of engine. In an emergency, say an electrical failure, this can be really crucial if certain systems need to be restarted/turned off -- it's best that pilot doesn't spend time during an emergency understanding where everything is. Even if they've flown the same plane for years, they're only human and you forget what you don't use.

Also, you have to check that the checklist machine is working. Automated safety systems failing has caused crashes.

An airplane is a mechanical thing and in the end you have to check that things really work physically. I am sure they have self-check mechanisms that check that the software works right and the sensors produce the right signals but you still have to check that the rudder really moves.

Many aircraft do have configuration warning systems, but they are not exhaustive. And, as a sibling commenter mentioned, many of them are physical inspection items. (Fuselage free of ice and snow, fuel appropriate for the mission, appropriate fuel loaded, pitot cover removed, etc)

I've always found this odd though - surely for pitot tubes a sensor could be installed which flags whether the cover is on? Sure, you open yourself up to a sensor failure scrubbing a mission but leaving the covers on is somewhat more disastrous.

OK, "pitot tube inlet free of mud daubers" (which incidentally gets looked at when you manually remove the cover).

Part of the walkaround/pre-flight process is me getting my head into the aviation game (and out of the "stressful workday" or "other life BS"). It's a few minutes of partial meditation as I walk around the aircraft, pushing, prodding, and inspecting.

(In my aircraft's case, the empty weight is around 2700 lbs and examples can be purchased for under $100K. There's a substantial premium on weight and [to a lesser extent] space for redundant or automated systems to give me a checklist that says "ensure green 'OK' light is lit".)

"Add a sensor" is a non-trivial exercise in the aviation world. If you're doing things correctly you need triple redundancy so you can determine when one sensor is returning bad data. Adding sensors for every little thing would add enormous complexity to what are already extremely complex machines.

You also open yourself to the possibility that the sensor gives a false negative, which is more disastrous.

Most aircraft aren't that sophisticated. When they are, like Boeing 787 and Airbus 320 series, much of the checklists are automated, but there still remain portions the pilot must perform themselves. Quite a lot of the checklist items have to do with establishing situational, contextual and sequential awareness.

Also, general aviation is a much bigger portion of aviation than scheduled passenger airline service: number of airplanes, pilots, airports, and jobs. Any automation is going to be narrowly applied outside of general aviation, it's just not yet a general purpose solution.

About 2 weeks ago a German Luftwaffe Global 5000 Businessjet missed the runway and nearly crashed in an emergency landing. The currently suspected cause: Faulty maintenance and incomplete preflight checks. http://www.airliners.de/in-schoenefeld-jet-flugbereitschaft/...

An interesting paper about systems on the edge of failure is "'Going solid': a model of system dynamics and consequences for patient safety" by Cook and Rasmussen: https://qualitysafety.bmj.com/content/14/2/130

There's a good overview in this talk: "Architectural Patterns of Resilient Distributed Systems" by Ines Sombra https://youtu.be/ohvPnJYUW1E?list=PL1Fqq0rxeNehyA4eWbuCwizkq...

Nice read. I often notice something similar during software development. It's simple stuff like "I have to do a full rebuild or it won't work" or "I have to build twice" etc pp. I cannot understand how others can simply ignore what could indicate a fundamental problem.

It's fucking terrifying.

I notice this too. However I've come to see that the building twice folks are cranking on writing code and getting praise from management while I'm finding some bug in yarn and not getting any actual work done..

> Before you try and change the world just have a look at the foundation from which you are starting it from.

So true. People who think they're doing the right thing can cause so much damage.

Podcast seems to be 404 on libsyn. I was worried it'd gone Spotify exclusive but the correct feed URL is https://feed.pippa.io/public/shows/5bb78b332630bb48399f39fe and the episode URL is https://feed.pippa.io/public/streams/5bb78b332630bb48399f39f... (as an alternative to the linked article).

I see this sort of complacent deviance all the time at my job. Thankfully we don't can't kill people with it. But insular organizations seem to thrive on doing things "their way" thinking it right, when in fact to anyone else it would be total lunacy. If people saw what we did I think people on HN would have a good laugh.

This reminds me a lot of this idea: https://www.lesswrong.com/posts/Kbm6QnJv9dgWsPHQP/schelling-...

"Normalization of deviance" would be the mechanism by which a slope becomes slippery; a schelling fence (in this case... a checklist?) is then a possible fix, even if it's just to trigger a re-evaluation of "normal"

I'm amazed that outside a wartime situation that flying a plane like that with a known equipment failure (of that nature especially) was allowed. And if it wasn't allowed, I'm surprised it didn't end up in a grounding of the pilot.

It wasn't just the pilot. It was also a few people who heard the idea and said that's ok. You can't fix systemic issues by punishing selected few. Aviation seems to be pretty good at that.

IMO the article is excellently written and very compelling. I suggest people read it for enjoyment, despite the dry title.

What do you do on a daily basis that makes "How I Almost Destroyed a £50 million War Plane" sound dry?

Good point. I was focusing on the "... Normalization of Deviance" part when I wrote that.

A lot of incidents seem to arise from normalization of deviance + self regulation (especially in the moment). And hubris.

The author's situation reminds me of the Tenerife airport disaster- absolutely macabre and sobering. [0]

[0]: https://en.wikipedia.org/wiki/Tenerife_airport_disaster

First heard this term in the danluu 'wtf wtf' blogpost -- so useful. Also 'fast jet performance' is such a cool podcast topic.

This one: https://danluu.com/wat/

Great articles there.

Everyone here seems to be caught up in waxing poetic about how the normalization of evidence is a bad thing and talking about preventing it but there's another side to that coin. If everybody is breaking the process 9x/10 it's not "normalization of deviance", it's because you've got a half baked process that sucks and the people on the ground know it and work around it. Obviously you could fill a library with all the material written on how to write organizational policy but it's not as black and white as "violating policy = bad"

Think of the global financial crisis of 2008 when the economy collapsed because the banks hadn’t been properly regulated because they convinced the authorities that they could do it themselves.

Also Boeing and the MCAS fiasco

Excellent read and a key concept.

I'd had a general idea about this, but it was crystalized by a vehicle dynamics engineer and racer who mentored me in both engineering and race driving. He said:

"You need to know the difference between skill and getting away with something"

I.e., just because you survive an event does not mean that it is reliable. If it is based on skill, you can replicate it and reliably survive. If it based on luck, well . . . take the bit of grace you were given by the Gods and learn your lesson to not do it again. Mostly, THINK.

"How I Almost Destroyed a £50M War Plane and the Normalisation of Deviance"


"How I Almost Destroyed a £50M War Plane [while nearly killing myself and my colleague] and the Normalisation of Deviance"

Yeah I thought that was interesting. Perhaps an artifact of how soldiers are trained to regard their own lives versus their equipment?

"Normalization of Deviance" sounds like the name of a ship from the Culture series.

The concept of Normalization of Deviance was mentioned in this thread in March about some comments made on Twitter by a software engineer about the 737 Max crashes.


Complacency creep.

Talking to a few Royal Australian Air Force folks last year, apparently they had a similar-ish culture of “she’ll be right” arrogance/confidence, and then they had a cluster of accidents that caused a foundation level rethink and rebuild of a safety/process culture into every flight and every high risk activity.

It sounds like accident #’s have collapsed since then.

Be careful, aviation fanboy writing, not a qualified pilot/tech/engineer.

Tried to read your article, but the way the email prompt loaded made it impossible to dismiss so I entered a false address.

I'm running the latest version of Chrome on Ubuntu 16.04 - hope that helps you diagnose this issue.

Here's a screenshot: https://imgur.com/kwXdIsa

What I find odd is that he chose to get his 0g flight by flying a parabola, in the style of the vomit comet, instead of just flying in a roll. For example, roll by 200 degrees and then raise the nose by about 20 degrees. (probably can't just roll 180 degrees without stalling)

In my experience this also applies to dating, it’s really easy to overlook things in the early stages of a relationship (or what might become a relationship) that turn out to be quite significant later.

This read reminded me of one of the famous TOP GUN movie quotes:

"A good pilot is compelled to always evaluate what's happened, so he can apply what he's learned. Up there, we gotta push it."

Top Gun...

Anybody reminded of Captain Flashheart while reading this?

When I think of "normalization of deviance" I focus more on mundane things like speeding.

I'm a cyclist. People often ask me why cyclists break the law so frequently. The studies I've looked at don't suggest cyclists break the law any more frequently than drivers do, but they do break different laws. (To my knowledge I follow every applicable traffic law.) When I ask the driver I'm speaking with if they speed I frequently get a justification of speeding as if it's perfectly safe because "everyone does it".

In terms of problems people should focus on, cyclists tend to break laws in ways that mostly harm only themselves (e.g., running a red light), while drivers break laws in ways that tend to harm others in addition to themselves (e.g., speeding, failure to yield, etc.). The fact that drivers vastly outnumber cyclists also makes the priorities even more clear.

I commute by bicycle from the burbs to the city, and whenever I break a rule it's almost always to get out of the way to help cars out, or to keep myself safer in some way. Sometimes I'll get ready for a hook turn by using the pedestrian crossing lights and footpath instead of blocking the turning lane for instance. It probably looks like I'm jumping the red because I'm impatient, but I'm just saving everyone some time and making my day a bit safer.

Another example, I tend to pull over the stop line at stop lights, so that cars don't accidentally creep into me. It happens even in designated "cycle boxes", which cars are supposed to stop behind to increase visibility of cyclists.

Sometimes you can get into a situation where you're in more exposure than you originally bargained for as well, when new cars show up or the situation changes, and I'll sometimes abandon the original plan and bail off the road when it's safe. Often breaking some laws when doing so, but I'm going to take responsibility for my own safety when I need to.

What I dislike seeing is cyclists just flying through reds, darting over crosswalks when pedestrians are still crossing, or weaving in out of moving cars so they can feel like an NYC bike messenger for a minute. It's not without consequence to others, you just don't notice that people had to jump out of the way or slow down so they didn't hit you. That makes it bad sportsmanship in my eyes.

> Another example, I tend to pull over the stop line at stop lights, so that cars don't accidentally creep into me. It happens even in designated "cycle boxes", which cars are supposed to stop behind to increase visibility of cyclists.

On top of that, I have encountered lights where if you don't do this, you will hold up traffic indefinitely because the light only changes when something bigger (heavier? more metallic?) than a bike is at the stop line

> cyclists tend to break laws in ways that mostly harm only themselves (e.g., running a red light)

Running a red light could easily harm others - a driver could swerve to avoid you, crashing in the process; if a driver hits you because you ran a light, that could also cause a chain of accidents; and if a driver killed you because you ran a red light, that could seriously harm their wellbeing.

> if a driver killed you because you ran a red light, that could seriously harm their wellbeing

While there are people who would be seriously bothered if they inadvertently contributed to a cyclist's death, I think the number of such people is smaller than the number of people who claim they would be bothered.

One only needs to look at their revealed preferences to show this. Given the choice between waiting a minute or two (estimated 95th percentile; the median is likely around 15 seconds) to safely pass a cyclist, many drivers instead choose to pass a cyclist very closely and dangerously. If they actually would be bothered if they killed a cyclist, why do they do this? My life is not worth a minute of someone's time.

Of course, the bad drivers could claim that they didn't know passing a cyclist closely was dangerous, but in my experience few drivers claim that. I've spoken to many drivers when stopped at stop lights, trying to understand their perspective. Often they imply that I deserve to be passed dangerous because I broke some rules of the road. So they can't claim ignorance.

(The "rules" I broke are not actually rules. For example, I've had people tell me that I was going far below the speed limit when I was going 17-18 mph in a 15 mph zone. I intentionally picked that road because of the low speed limit, yet some people want to drive 40+ mph there and apparently that makes me a jerk who deserves to be run over.)

You cannot use driver behavior in the absence of an accident to infer the emotional impact of killing someone...

Why? These drivers often show no remorse for their actions, as I stated. I don't doubt I'd be a victim of a hit-and-run if any of those drivers actually did cause a collision.

I've seen arguments that people are "different" when driving, so perhaps they're a jerk when behind the wheel but a normal person otherwise. I haven't seen any clear evidence for this but it would indicate that the emotional impact would be different.

They show no remorse mostly because you didn't end up hurt.

You should know how bad people are at estimating risk. Combine that with some short-lasting road rage and you can easily find someone that will do dangerous things all the time but be distraught if that ends in death.

> They show no remorse mostly because you didn't end up hurt.

I disagree with the emphasized part of your statement. Drivers seem worried about the emotional consequences of killing someone. But apparently the emotional consequences of nearly being killed don't count? Or at least they haven't been mentioned here. Nearly being killed is not harmless.

> Combine that with some short-lasting road rage and you can easily find someone that will do dangerous things all the time but be distraught if that ends in death.

I can accept that many drivers are road raging and that their behavior otherwise would be different. But I can't accept that's all of them, as that seems inconsistent with my experience in over a decade of cycling.

They don't see it as 'nearly killed', so yeah no major emotional consequences.

They don't have to be correct to be understandable.

> I don't doubt I'd be a victim of a hit-and-run if any of those drivers actually did cause a collision.

The only way I can read this is that you believe, of the people that cause a collision with a bicycle (accidentally or otherwise), all (or almost all) of them would just leave you to die on the side of the road.

That is a staggeringly low opinion of humanity.

The low opinion is not of humanity, rather, certain drivers who are particularly dangerous.

I didn't have all drivers who'd be in a collision with a cyclist in mind, just a large fraction of those who deliberately pass cyclists closely. My wording was too strong.

Hit and run crashes account for roughly 12% of all crashes in the US, so they may be more common than you believe: https://aaafoundation.org/hit-and-run-crashes-prevalence-con...

>That is a staggeringly low opinion of humanity.

When people make sweeping assumptions about the nature of other people it says something about them too.

Too late to edit now, but on second thought many drivers who pass me dangerously might not believe that I deserve to be passed dangerously. I've heard many times from these drivers that they "had no choice but to pass me dangerously". But they did have a choice. And they seem entirely aware that what they did was dangerous, which is my point in that paragraph.

As a driver in a country that respects biking (The Netherlands), it's completely normal to, say, spend quite a bit of time driving very slowly behind a cyclist or two because there's oncoming traffic so no way to pass them without having the whole road available.

It's also perfectly normal to be extremely careful at crossroads because some cyclists consider red lights suggestions.

Of course, when cyclists clearly misbehave there'll be grumbling, but I find it interesting that even in those cases, my 'rage' pretty much never crosses the line of respecting cyclists at all costs. It's just part of how we're taught to behave as drivers (and, I suppose, the legal ramifications that generally favor cyclists, afaik).

As computerfriend pointed out, I said mostly. There are many possible ways a cyclist running a red light could go wrong, but I think the majority of them involve harm going primarily if not exclusively to the cyclist.

Not at a pedestrian crossing. I see this all the time in Cambridge. If you run a red light at a pedestrian crossing the most likely thing that could go wrong is that you hit a pedestrian. Hopefully they won't be badly injured, but they might be. Please don't do this.

> Please don't do this.

Was this advice intended for me? I don't run red lights or stop signs, and am particularly careful in pedestrian heavy areas.

Cyclists have killed pedestrians before, e.g., Chris Bucchere: https://www.sfchronicle.com/bayarea/matier-ross/article/Bicy...

> Was this advice intended for me?

No, all cyclists.

Your claim was that "cyclists tend to break laws in ways that mostly harm only themselves", and you cited running a red light, although I suspect running a stop-sign is much more commonplace. As I am sure you know, these are considered to be quite dangerous and serious traffic violations when performed by a car, to the extent that most drivers would not even run a red light if stopped at an intersection with no opposing traffic visible. By comparison, it is difficult for cyclists to speed on most roads, but there is no evidence to suggest that they would be any less inclined to do so than drivers were the speed attainable and were it to seem subjectively reasonable. These are "different laws" in the sense that one is a violation of a safety margin designed to allow prompt braking under all circumstances and the other is an unexpected, unpredictable violation of the rules governing patterns of traffic behaviour.

This makes me think that the "difference" you cite is less a matter of which laws the cyclist chooses to break, and rather more due to the vehicle with which the cyclist chooses to break them. Were cyclists likewise equipped with an exterior body and 1-2 tonnes of mass they would be responsible for a wholesale slaughter.

Unlike the selfish car driver, of course, the cyclist at least has the ethical honesty to knowingly commit to breaking specific laws, secure in the knowledge that they will only harm themselves, their loved ones, and the driver of the car that kills them.

> This makes me think that the "difference" you cite is less a matter of which laws the cyclist chooses to break, and rather more due to the vehicle with which the cyclist chooses to break them. Were cyclists likewise equipped with an exterior body and 1-2 tonnes of mass they would be responsible for a wholesale slaughter.

It's the net effect of both. From a consequentialist perspective it does not make much sense to focus so heavily on cyclists running red lights. Drivers speeding is much more dangerous and occurs at a much higher frequency, yet it is much more socially acceptable than cyclists running red lights.

> Drivers speeding is much more dangerous and occurs at a much higher frequency, yet it is much more socially acceptable than cyclists running red lights.

Unless you're on a road which was designed and marked with the expectation of mild speeding, which as far as I understand is most of them. In that case it puts you at the target level of danger, not an increased level of danger.

Everything cyclists do generally only harms themselves, everything cars do generally harm others. Not sure why this is so complicated to you. Cars are killing machines, which kill 40,000 people a year in the US, more than guns.

In all those examples the cyclist is still the one with the highest stakes by far. When a cyclist runs a red light it's vastly more probable that he get killed than that some innocent pedestrian gets killed by a swerving car (and the chances that the driver of a car gets killed are basically 0). While that doesn't absolve them of all responsibility it is morally much more justifiable than for example a car speeding.

They said "mostly" and, while your comment is undoubtedly true, it is a little bit disingenuous to suggest those things as likely when cyclists are the most vulnerable road users.

> disingenuous to suggest those things as likely when cyclists are the most vulnerable road users.

Why is it disingenuous? Cyclists can be jerks too

As a pedestriab I've been nearly hit by a cyclist way more often than nearly hit by a car. Running red lights on a bike in a city is very dangerous to people crossing the street.

Cool. Now, how many pedestrians do you think were killed by cars last year, and how many by cyclists? Or does dying not really worry you?

There are way more cars than cyclists

One does not excuse the other.

I live in an area where cycling is really only done by the poor and certain enviro-progressives and unfortunately it seems like many of these folks don't seem to give a shit about cars -- blowing through four-way stops, occupying the entire lane of a two-lane country road, and just generally being a nuisance. It's illegal to ride on sidewalks here but it really is the best thing some of these people can/should do (at cruising speeds, anyway)

> It's illegal to ride on sidewalks here but it really is the best thing some of these people can/should do (at cruising speeds, anyway)

Riding bikes on the sidewalk is known to be more dangerous than riding on the road: http://bicyclesafe.com/#crosswalk

Because the harm done by cyclists is a tiny fraction of that done by drivers. Jerk cyclists very occasionally kill pedestrians, whereas jerk drivers kill pedestrians all the time. So it's disingenuous to equate them.

Seriously, we should install governors and bike control software on all bikes because they are killing 102 (1) people per day!!

1 https://en.m.wikipedia.org/wiki/Motor_vehicle_fatality_rate_...

> People often ask me why cyclists break the law so frequently.

At least where I've ridden in the States, cyclists tend to be treated as something between cars and pedestrians. For example, a driver often squeezes by you on a 2-lane road with oncoming traffic. In these circumstances, it's hardly surprising that cyclists make up their own rules. I regularly break the laws for cars on a bike, because I am more maneuverable, and I'm the one who's toast if either of us makes a mistake.

When deviance is normal, you have to adapt.

I know someone who hasn’t driven since a cyclist ran a red light and crashed right into the middle of the passenger side front door (which had to be replaced for >$500 of which the cyclist paid $0). There was also a kid in the back seat who was pretty upset for a long time.

And the cyclist had a pretty bad day too.

Running a red light is dangerous and can have plenty of collateral damage.

My point wasn't that bad things can't happen when cyclists run red lights, and nowhere in my post did I justify running red lights. Again, when a cyclist runs a red light the potential harm is mostly on the cyclist.

You can find plenty of real world examples of cyclists harming others from running stop signs or red lights, e.g., the Chris Bucchere case comes immediately to mind:


The number of these cases where bad cycling caused harm to others is much smaller than the number of cases where bad driving caused harm to others. That's my point.

You are technically correct, but it is the worst kind of correct. Because you are arguing a losing case with a valid point. That most of the time they will mostly only harm themselves does not actually make the law breaking excusable.

Contrast this to speeding. Speed limits by law should be set to the 85th percentile speed, based on the evidence backed notion that drivers will predominantly pick a speed which is safe based on the road conditions, and that it’s most safe when drivers are all proceeding at the same speed.

The vast majority of every day speeding that I witness, and that I participate in, is of the kind where the limit is set illegally low due to political pressure or other factors, and the predominant and safest speed is higher than the limit. This is true, for example, for many highways in MA which have limits of 55 mph.

That’s not to say that I never truly speed, which I shouldn’t do, and is entirely unjustified.

Nowhere am I arguing that cyclists running red lights is acceptable or excusable. My argument is that drivers speeding is often seen as acceptable, but cyclists running red lights is not, and I'm trying to highlight the asymmetry in the effects.

> Contrast this to speeding. Speed limits by law should be set to the 85th percentile speed, based on the evidence backed notion that drivers will predominantly pick a speed which is safe based on the road conditions, and that it’s most safe when drivers are all proceeding at the same speed.

The 85th percentile rule is meant for highways to my knowledge, not every road. Edit: Here's a criticism of the 85th percentile rule: https://www.strongtowns.org/journal/2018/1/25/speed-kills-so...

The point is that drivers speeding is responsible for 10k+ deaths a year, and it is largely accepted by society. But cyclists running stop signs or red lights is responsible for a negligible amount, and most people would say this is a much bigger problem than speeding

Equally, I know a cyclist who started running red lights because he was hospitalised by somebody whose defence was "but you're a cyclist and you stopped for the red light, how was I supposed to expect that?"

Not defending bad behaviour here, just observing how it can be incentivised.

Do you drive?

Drivers kill over 40,000 people every single year, more than gun deaths? And many, many more seriously injured.

Now, are you just pretending to care, or are you going to stop driving.

> cyclists tend to break laws in ways that mostly harm only themselves

Clearly you're not a pedestrian in Manhattan.

I mentioned Chris Bucchere in several other posts.

You and the poster above you also miss the point: Drivers cause much more harm than cyclists in total, yet drivers speeding is deemed much more acceptable than cyclists running red lights.

I don't doubt that cyclists can harm pedestrians or others, and I don't believe that cyclists breaking the law is acceptable. Just that when cyclists break the law the amount of potential damage is much more limited than when drivers do.

I think you ought to compare a driver running a red light vs a cyclist running a red light to make this point. I might even agree with you, but speeding is conceivably even safer than going the speed limit if everyone else around you is doing the same thing no matter what the law says.

Running traffic signals on the other hand is just plain stupid and self-centered no matter what you're piloting at the time.

I can compare drivers running red lights to cyclists running red lights based on my experience. I don't believe there's any hard data on this.

From what I can tell, drivers tend to run red lights either because the light just changed and they couldn't stop in time or because they're distracted. In contrast, cyclists who run red lights tend to wait for traffic to clear. I didn't realize the difference until I tried talking to cyclists who run red lights (not recommended by the way; they tend to be jerks). So in that sense the two are not comparable.

Plus, as bikes tend to weigh less and travel less fast than cars, the potential for harm for a cyclist running a red light is greatly reduced compared against a driver running a red light.

Here are some helmet cam videos that I took that show red light running drivers:



After the first video, I saw that the driver stopped nearby, so I went to speak with them. They admitted they were distracted. I have no idea what happened in the second video. If I were not paying attention in both cases I would not be writing this post now. In contrast, if a cyclist did the same then I'd probably be here.

That's not to say that all cyclists who run red lights look carefully. Here's a helmet cam video that I took which shows a cyclist brazenly running a red light in a way that I think is particularly dangerous:


And here's a more recent video of mine showing a Bird scooter rider running a red light:


(Apologies for the language.)

This last one was probably the most dangerous time someone ran a red light that I've personally seen. I was wearing a high-viz jacket at the time and find it hard to believe they would not see me if they looked. They might have been blinded by the sun.

As for speeding being safer, sure, that's true sometimes, but what's the net effect? Just focusing on the good isn't representative.

I also sometimes see people bring the time saved from speeding into consideration, but in my experience speeding doesn't save much time unless you're on the highway. If you're in a more densely populated area with traffic lights, your speed is more limited by the traffic lights than anything else. I know this as a cyclist because it's not uncommon for me to see the same people at each light during my commute. The effective traffic speed is doable for a cyclist. Speeding won't get you there faster.


> I think it should be a case closed Darwin award if I hit them and kill them if I have the right of way

I'll be sure to stay away from you.

> I used to know a paramedic who would make sure that the ride to the hospital for such a cyclist was super bumpy and unpleasant so as to make that experience memorable in the hopes of encouraging them not to repeat it.

That's sadistic. Most of the time it's not the cyclist's fault.

It's not the cyclist's fault when he/she runs a stop sign or a traffic light? Explain? It would certainly be my fault if I did that in the car.

Running a stop sign or traffic light should not be a death sentence.

Plus, there have been many times when I've been on the receiving end of dangerous "punishment passes" from drivers who believe I'm doing something illegal when I'm not. I've had people pass me aggressively because they think I'm going too slow, when the speed limit is 15 mph and I'm going 17-18 mph. I've have people pass me while yelling to get in the bike lane when there was no bike lane and I was in the sharrow lane on an official city bike route. Etc. Just because you believe someone is breaking the law does not give you the right to kill or otherwise harm them, particularly if you're wrong about the law.

Edit: I believe I misread you about the paramedic. I read it as the paramedic would give rough rides to all cyclists (and most of the time it's not the cyclist's fault that they were in a crash), but you meant they gave rough rides to only cyclists who ran red lights or stop signs. That's still bad and sadistic as far as I'm concerned.

You're right it shouldn't be a death sentence but neither should pointing a gun at your head with only one bullet in it and pulling the trigger. They're both different versions of the same game to me. I think people who run these traffic signals do not realize how much danger they are in and I don't think such people would particularly choose to play Russian roulette. And yet there they are running red lights and stop signs.

In the split second that my car is in an intersection and and such a person comes barreling through, the best I can do is slam on my brakes and hope for the best. I'm not even going to try to get out of the way because that could only make things worse by hitting another car or pedestrians who had nothing to do with the situation. If I had infinite time to respond I might be able to think this through, but we are talking less than a second of reaction time here. Really? This is worth it to you?

As much as I am in favor of cyclists and I am in favor of bike paths and every other variant thereof I do not get the incentive to save 30 to 60 seconds dodging a traffic signal versus the cost of what happens if it all goes wrong and I don't think I ever will.

Downvote this at will. This thread had me on the verge of deleting my Hacker News account but then I found out you can't do that. So I am posting this final response to say that this is what I think and if someone won't hire me because I think this way I probably don't want to work for them.

As for that paramedic, he didn't start that way. There used to be this particularly obnoxious cyclist in his town that made a habit of flying through intersections and taking his hands off the handlebars and giving the fingers to all the cars he cut in front of. One day he got hit and broke his leg cursing the driver that hit him the whole time. Said paramedic was the one who took him to the hospital knowing exactly who he was and what he had been up to. I still think the guy was a Darwin award in the making and he probably died later on in life if he continued doing crap like this. If that makes me a despicable human being then I am a despicable human being.

Perhaps we've misunderstood each other. The impression I got from your post was that you believe harming or killing cyclists is acceptable if you perceive that they are breaking the law (regardless of whether the cyclist is actually breaking the law; see the examples I gave previously). Many drivers seem to believe that. If that's not what you believe then I apologize for misunderstanding you.

Add on top of that the belief of many drivers that their own law breaking is acceptable, speeding in particular. Speeding is a contributor to roughly 10,000 traffic fatalities per year while running red lights is a contributor in only 700-800. Speeding is clearly a bigger problem, yet cyclists in particular running red lights gets more attention than speeding. (Likely the vast majority of traffic fatalities from running red lights don't involve cyclists at all.) One of the points of my first post in this discussion was that speeding should also be seen as unacceptable, but many of the replies I received focused on how cyclists running red lights is unacceptable and that speeding is fine. In other words, many people missed the point.

> I used to know a paramedic who would make sure that the ride to the hospital for such a cyclist was super bumpy and unpleasant

That behaviour is incompatible with a professional registration. (At least, in England it would be.)

Despite what was widely reported in the media, I did not run a single red light (one or multiple). During the preliminary hearing, video evidence proved that the eyewitness who made the claim that I ran "several red lights and stops signs" was lying. Furthermore, the prosecution's expert witness accidentally proved that my light was yellow. The DA subsequently dropped the red light violation from from my charges. All the details here: http://bikelash.net

I prefer what you may have written up before you got yourself a lawyer. and especially the comments from fellow cyclists on your behavior that day...


You can prefer whatever you want but you can’t argue with the video that captured the entire accident. It showed that my light was yellow and that pedestrians started crossing as much as 13.9 seconds before the WALK indicator illuminated.

People are responding angrily to your comment about running a red light. That’s understandable as it can be extremely dangerous for everyone involved.

There is however a good rule I heard was in use somewhere: Cyclists can regard stop signs as yield signs, and red lights as stop signs.

To reiterate a point I've made several times here: I'm not advocating for running red lights. Quite the contrary. I think all cyclists should stop at stop signs and wait for red lights.

My broader point is that cyclists running red lights gets a lot of attention, but ultimately drivers speeding causes much more harm, yet that is accepted by society at large.

Motorists run red lights differently to cyclists, because you are at the front of the lights you just don't see it. However, watch any lights in the UK and you will be shocked at how many motorists waltz through after the lights have gone red, waaay after the amber light has gone out.

As a cyclist you filter to the front so you just don't see this as often as it happens. Statistically there are more cars on the roads than bicycles so the majority or red light jumping is almost certainly by cars. On a bike it is just a bit more blatant with turns on left (UK) made that motorists would never do if there was a red light stopping them.

Jumping pedestrian crossings - 'pelican crossings' - is another thing. Cyclists get this wrong but there is also a lot of common sense, if people have crossed already, why stop? Plus if the traffic was cyclists and buses then these crossings would not be needed.

Funny the flack you have got here.

" harm only themselves " would apply to lots of violations when you're not driving a moving, multi-ton brick-on-wheels

And? If there's an asymmetry in the risk, that means you should account for the asymmetry.

Over the years I've noticed some people seem to think that merely categorizing an argument is enough to refute it, but you need to explain why arguments in a particular category are bad.

(Yes, I am aware this post may be perceived as an example of the bad argument I just mentioned, though I did justify it.)

I think matters are complicated in this regard, I have heard that cyclists count as pedestrians (and not as vehicular traffic) in some jurisdictions/countries, so in such cases they can even and should ride on the opposite side of the road from vehicular traffic.

I thought it was safer if cycling on roads, to ride forward on to the one coming traffic, rather than have traffic come up from behind you.


Wrong way cycling is much much more dangerous than right way from what I understand.

Running red lights is considerably more unsafe than speeding.

Not true. In the US, running a red light contributes to 700-800 traffic fatalities per year [1]. In contrast, speeding contributes to about 10,000 fatalities per year [2].

[1] https://www.autoaccident.com/statistics-on-intersection-acci...

[2] https://www.nhtsa.gov/risky-driving/speeding

You might want to look at something other than fatalities, sure, but this simple analysis suggests that speeding is more than an order of magnitude worse.

Edit: And if we look at the total number of people killed in traffic by cyclists each year, I don't even know where to start, because it seems to occur so infrequently that no one keeps statistics on it. I'd be very surprised if it were more than 5 per year.

That shows that speeding it's a greater source of unsafety, but I wouldn't be surprised if running a red light is less than 7/100ths as common as speeding, in which case that statistic would support running a red light being the more dangerous act.

Fair points. I'm thinking about the net effects rather than per act.

You are misinterpreting the grandparent: Their claim is comparing an event of running a red light vs speeding once. Your statistics have the incidence of both behaviors already baked in. So with your numbers, you could say that all the speeding your average driver does in a year is riskier than all the red light running they do in a year.

It's precisely the difference in the incidence of the behavior that matters here though. I don't have specific numbers, but I think it's a very safe claim that most drivers speed far more often than they run red lights. Many a commuter will be speeding a large percentage of the time they are not in a traffic jam. On the other hand, I don't see of any people in the US that just run any and all red lights as a matter of course. And why does it happen that way? Because we realize that running a red light in most circumstances is crazy dangerous. People tend to try to avoid extremely dangerous behavior if they can help it, so you will find few fatalities per year, but that's because almost nobody does it. How many Americans die each year to Russian roulette? Probably fewer than either red lights or speeding, although I'd say playing Russian roulette is far more dangerous than either.

Given the numbers you found, running a red light would be just as dangerous as speeding if people ran a red light 13 times less often than they speed (for whichever time definition we choose for a unit of speeding). I suspect that people speed far more often than that, compared to running red lights.

We are in a similar boat regarding people killed in traffic by cyclists. How many cyclist-hours do we see in US roads, vs cars? There's also probably very few people killed in the US by being killed by T-35 tanks running red lights, but that's not because T-35s are safe and we should just use them and then ignore traffic signals. Running someone over with a bicycle while running a red light is still probably far less dangerous than acting the same way in a car, or an old T-35, but we can't tell that just by looking at fatalities alone.

I agree with your basic point; see firethief's comment. My larger point is this: If we want to minimize fatalities then we should focus on what causes the most fatalities. Speeding is a contributor to far more fatalities than cyclists running red lights, yet speeding is socially acceptable and cyclists running red lights is not. Neither should be socially acceptable.

About 3 trillion miles are driven, and about 9 billion miles cycled in the US each year. The fatality rate per mile for cycling about 5x the risk of driving, but hour for hour it is about the same because of the difference in speed.

The original point was that speeding is dangerous to others and normalized by society, but a cyclist running a red light is not dangerous to others, but it is not normalized. Speeding drivers kill 10k people a year. Cyclists running red lights kill less than 1 a year on average.

Both completely depend on circumstances. So much so that it's pointless to compare them without specifics.

No, it's not.

Running a red light is pretty dangerous in almost any context. Running red lights consistently will 100% yield accidents.

Speeding is normative, it's something that everyone does, every single time they drive.

For speeding up to 20K above limit, you often don't get a ticket. Doing 120 on a 100 km/h road is normative. If it was 'very dangerous' the police could easily enforce a 100 cap.

Running red lights consistently will probably kill you, and if a cop sees you doing it, you will 100% get pulled over.


Lots of traffic laws on the books are irrelevant to actual safe use of the road.

-No attachments like dashcams on the windshield in NJ. Fine if you attach it to the dash instead and it goes to the same spot. Arbitrary and useless.

-Technically illegal to turn left across a solid yellow in many states. Universally ignored.

-Speed limits often too low, sometimes too high.

-Can't go through yellow lights in some states

That's an uncharitable reading of my post. By applicable I meant laws that apply where I am, e.g., I am in Texas right now so I don't follow the laws of another state.

Speeding doesn't cause harm. Having an accident causes harm. And traffic engineers say that speed limits are set in the expectation that the average speed on the highway will be 5 to 10 percent higher than the posted limit, which means that the posted speed limit itself is not supposed to be a hard limit such that exceeding it is "deviance".

If the posted limit wasn't supposed to be a limit then why do you get a ticket for exceeding it?

As bitdivision points out, mostly you don't if you're going within the range the traffic engineers gave (less than 10 percent over the posted limit).

But what that really means is that enforcement is at the discretion of the police--which in turn means that, for example, many localities have speed traps where they all of a sudden enforce the posted limit strictly in order to increase their revenue. Or police have a quota of a certain number of tickets per month, and if it's towards the end of the month and they're having trouble meeting quota, they pick some locations and decide to strictly enforce the speed limits there until they have enough tickets for the month.

All of this just reinforces my point that the posted speed limit is not set on the basis of preventing accidents, but for other reasons.

Of you're going <10% over the limit you normally won't get a ticket in my experience.

Artificial arbitrary limits. There's no question that many jurisdictions decrease speed limits to below what the road sustains, primarily for taxation without representation (speeding tickets).

The road's natural limit is what the road should be. Of course, there's souped up muscle cars and the like that can run 150mph all day.. But those kinds of cars are outliers. I'm talking about the 2 lane road that used to be 45 near our house, that got shoved to 35. Or 55->30->55 within a half mile for bumfucksville... but that cop is sitting right there on the 30.

> The road's natural limit is what the road should be.

I agree with the principle, but I don't think that in practice this principle can be satisfied by having a posted speed limit that you can be ticketed for exceeding. The "natural limit" of a road is not a single number. It depends on many factors, some of which (e.g., weather, driver skill/fatigue) can't even be known when a single number is being determined for a posted limit.

If posted limits were advisory only--in other words, you could not be ticketed simply for exceeding a posted limit, but if you were in an accident, the fact that you were exceeding a posted limit could be considered as a factor in determining fault--that would be different.

I don't think we should be ticketed for speeding.

Conversely, I do think that we should be ticketed for "hazardous driving" (weaving in and out of cars, cutting people off, and other things that directly lead to wrecks), failure to stop at signs and lights, failure to yield when other cars present.

There also have been many times in which I have driven significantly under the posted limit because the conditions severely prevented it. Driving at the posted limit would also have been hazardous.

Conversely, driving 80 on the interstate while staying in a group of others going 80 +/-3 is not hazardous. Yet, that would get us ticketed in a jurisdiction where I am not a voter. Hence, my complaint of 'taxation without representation'. It also nearly guarantees my agreement of guilt, as they're counting on me not able to make it back for a court date.

(Sigh, gotta love the downvoters. Actual discussion is just 'too' hard.)

> I don't think we should be ticketed for speeding.

I don't either, but current speeding laws allow cops to do so.

> I do think that we should be ticketed for "hazardous driving" (weaving in and out of cars, cutting people off, and other things that directly lead to wrecks), failure to stop at signs and lights, failure to yield when other cars present.*

I don't think we should be ticketed for any of these things, if they do not cause an accident. But if you are in an accident and it is found that you did any of these things, they should be valid factors to consider in assigning responsibility.

> There also have been many times in which I have driven significantly under the posted limit because the conditions severely prevented it.

Yes, I agree. In other words, you, exercising your judgment, have a much better knowledge of what is actually a safe speed than the lawmakers and bureaucrats that determined what went on the posted speed limit sign.

No no no. Allowing people to drive like crazy until they get into an accident is a formula for normalizing very unsafe behavior.

> Allowing people to drive like crazy until they get into an accident is a formula for normalizing very unsafe behavior.

No, it isn't, because if they get in an accident you hold them responsible, and when their insurance company finds out the accident was the result of driving like crazy, they either drastically raise their rates or terminate their policy altogether. Plus it would be perfectly reasonable for the state to revoke their driver's license if they caused an accident due to driving like crazy.

The problem with the system we have today is that (a) we punish people who have not caused harm, and (b) when people do cause harm, we are hesitant to hold them responsible to the extent I described above. Lose your insurance and your license because of one accident? Wow, that seems really harsh. Do they really deserve that? My response is, if they really were driving like crazy, then yes, they do deserve that, because they should have known better. But our current system doesn't seem to like taking that position.

People drive 45 down my residential street and it's dangerous for pedestrians, that deserves a ticket imo

Have there been any accidents?

Climate change is another example of this probably. Except the whole planet instead of a plane.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact