Random Killing on a Canadian Greyhound Bus

After a random and horrific knife decapitation on a Greyhound bus last week, does this surprise anyone:

A grisly slaying on a Greyhound bus has prompted calls for tighter security on Canadian bus lines, despite the company and Canada's transport agency calling the stabbing death a tragic but isolated incident.

Greyhound spokeswoman Abby Wambaugh said bus travel is the safest mode of transportation, even though bus stations do not have metal detectors and other security measures used at airports.

Despite editorials telling people not to overreact, it's easy to:

"Hearing about this incident really worries me," said Donna Ryder, 56, who was waiting Thursday at the bus depot in Toronto.

"I’m in a wheelchair and what would I be able to do to defend myself? Probably nothing. So that’s really scary."

Ryder, who was heading to Kitchener, Ont., said buses are essentially the only way she can get around the province, as her wheelchair won’t fit on Via Rail trains. As it is her main option for travel, a lack of security is troubling, she said.

"I guess we’re going to have to go the airline way, maybe have a search and baggage check, X-ray maybe," she said.

"Really, I don’t know what you can do about security anymore."

Of course, airplane security won't work on buses.

But -- more to the point -- this essay I wrote on overreacting to rare risks applies here:

People tend to base risk analysis more on personal story than on data, despite the old joke that "the plural of anecdote is not data." If a friend gets mugged in a foreign country, that story is more likely to affect how safe you feel traveling to that country than abstract crime statistics.

We give storytellers we have a relationship with more credibility than strangers, and stories that are close to us more weight than stories from foreign lands. In other words, proximity of relationship affects our risk assessment. And who is everyone's major storyteller these days? Television.

Which is why Canadians are talking about increasing security on long-haul busses, and not Americans.

EDITED TO ADD (8/4): Look at this headline: "Man beheads girlfriend on Santorini island." Do we need airport-style security measures for Greek islands, too?

EDITED TO ADD (8/5): A surprisingly refreshing editorial:

Here is our suggestion for what ought to be done to upgrade the security of bus transportation after the knife killing of Tim McLean by a fellow Greyhound bus passenger: nothing. Leave the system alone. Mr. McLean could have been murdered equally easily by a random psychopath in a movie theatre or a classroom or a wine bar or a shopping mall -- or on his front lawn, for that matter. Unless all of those venues, too, are to be included in the new post-Portage la Prairie security crackdown, singling out buses makes no sense.

Posted on August 4, 2008 at 6:19 AM • 90 Comments

Comments

antibozoAugust 4, 2008 6:37 AM

Schneier> People tend to base risk analysis more on personal story than on data, despite the old joke that "the plural of anecdote is not data." If a friend gets mugged in a foreign country, that story is more likely to affect how safe you feel traveling to that country than abstract crime statistics.

I think there's something facile about your argument. A friend's personal experience in a foreign country is more predictive of what will happen to me than the experience of an anonymous person which was used to compile data. A friend has interests similar to mine; I will count on the friend's recommendations for where to stay, what places to visit. This is a natural consequence of the person's status as a friend, both because of our affinity and because we might have more to talk about later if we have seen some of the same things. So the places we go are correlated, and our exposure to random crime is not statistically unrelated.

This doesn't mean "data" should be ignored. But "data" is less targeted toward my experience than the anecdotes of a friend.

JeroenAugust 4, 2008 6:51 AM

@antibozo:

You are proving Bruce's point :). While visiting locations X, Y and Z may be factors contributing to the likelihood of you getting mugged, the fact that your friend got mugged in no way affects your chances of getting mugged. They are stochastic events. Maybe your friend was the first tourist (assuming you are tourists) getting mugged in 10 years, or maybe 1 in 2 tourists taking that route get mugged.

The point is that a single event has no statistical, and hence predictive, value.

antibozoAugust 4, 2008 7:03 AM

Jeroen> You are proving Bruce's point. :)

And you are missing mine. :^)

Crime statistics are not statistics for a person doing the particular things my friend was doing. They are statistics for a much larger set of behavior. If I have crime statistics for the particular places and activities my friend engaged in, yes, those are clearly more reliable than my friend's anecdote. But typically that's not what we have to work with; we have instead aggregate data and no way of subsetting that to something more targeted.

The data are indeed stochastic. But my behavior and my friend's behavior are related. Raindrops falling into a pond have a certain probability of hitting a lily pad. But my friend and I are not independent raindrops. We are much more likely to fall on the same lily pad.

gregAugust 4, 2008 7:31 AM

@antibozo

Its still bad way to make decisions in terms of safety, no matter how many analogies you come up with. Your lack of independence with your friend is not going to cause the fact that nobody is at risk from decapitation on a bus, or any other "random" violence act. If its rare its rare, you and your friend don't change that....

bobAugust 4, 2008 7:39 AM

"...her wheelchair won’t fit on Via Rail trains..."

Really? In this day and age? Canada doesnt have an ADA? (I guess that would be "CDA")

clvrmnkyAugust 4, 2008 7:43 AM

@greg: a very small number. Far smaller than those killed in auto accidents, which is either the no. 1 or no. 2 killer in Canada.

We know bus acidents are rare because when a bus full of people are killed in an accident, it makes the news.

@antibozo: There is absolutely nothing predictive about this story. Regardless of how you collect information on remotes deaths, you cannot predict how mentally ill person will act out a deadly fantasy, and you cannot predict where you will be when this happens.

bobAugust 4, 2008 7:47 AM

People keep saying you cant have metal detectors for buses because the number of tiny stops makes it prohibitive.

No problem; build it into THE BUS.

antibozoAugust 4, 2008 7:48 AM

greg> Your lack of independence with your friend is not going to cause the fact that nobody is at risk from decapitation on a bus

I'm not saying Bruce's entire essay is wrong, or that we should believe everything we see on TV. I'm saying I think the argument facile in one particular area because it ignores the correlation in experience we have with our friends.

clvrmnky> There is absolutely nothing predictive about this story.

I didn't say there was.

D0RAugust 4, 2008 7:49 AM

And how many people have been beheaded by a crazy bloke in a Canada bus in the last ten years?

JoshuaAugust 4, 2008 7:54 AM

On the bright side, most likely everybody is going to forget about this in a week and none of the knee-jerk security theater measures will ever be put in place.

pound sandAugust 4, 2008 7:57 AM

800,000 people die in a tragic human caused event in Rwanda, in 93, I didn't know any of them. I recognize this as an abstract tradgedy. 3,000 people die in twin towers in New York, they speak the same language as I do, but they wouldn't likely speak to me because of class snobbery. How is that different from a million cambodians who live under a 5 year rain of bombs from the US Airforce? I feel tragedy when I know and care about someone. If a woman is abducted like Dru Sjodin and the media publish photos where she is beautiful and smiling, more people care than if it happens to someone without such photos. If you know someone, you care, or the media can make you care, I suppose, but not as much as if you are aquainted. Is it worse to be beheaded? many killings involve prolonged torture, whether done on bushs orders or done by a good republican like dennis rader, the btk of witchata.

xAugust 4, 2008 8:03 AM

Meanwhile, tons of people in Canada are enjoying marijuana, without being thrown in jail to be sexually or otherwise violently assaulted, while the U.S. drowns in prescription drugs and alcohol.

gregAugust 4, 2008 8:08 AM

@antibozo

Feeling that some things are relevant will not make it so. To sum up, the correlation in experience we have with our friends is almost always not useful to make correct or informed decisions.

We my use it as information thats important, but that does not make it so.

cinicAugust 4, 2008 8:13 AM

@bob

Do you work for a metal detector company?

Tell me how i can get my groceries home, or buy new knifes and forks or any of the normal things folks are suppose to do on public transport if no metal is allowed onboad.

metal detectors in airports are a pain.

Metal detectors on public transport is unworkable.

antibozoAugust 4, 2008 8:14 AM

greg> To sum up, the correlation in experience we have with our friends is almost always not useful to make correct or informed decisions.

I agree that statistics have value, but that statement is just plain false.

I have a friend who builds systems using the same gear I use and runs similar technology on top of them. This is not a coincidence; it is because we are friends, and talk about what we're doing to one another, that we use the same gear and technology. Naturally, his experience is far more valuable to me than the results of a survey of 10,000 sysadmins.

Clive RobinsonAugust 4, 2008 8:17 AM

@ bob,

"No problem; build it into THE BUS."

That's not the real problem.

The real problems are manpower space and time and their attendent costs,

On the assumption that everybody caries something that will trip the metal detector (shopping, coins, comb, belt buckle, under wired bra, etc etc etc) how do you deal with it?

You have a bus driver (hopefully) in a protective area what exactly is he going to do?

So you would need atleast another person to run the "frisk".

So Jo Public has to empty out their pockets etc whilst the rest of the que stands and (hopefully) waits for their 3 minute turn as oposed to the old 5-30 seconds to show their pass or pay.

Then there is the space for this extra person, a table or some such for people to put their offending items etc, so that's a big chunk of the passenger carrying space gone.

I could go on but I think you get the general idea, it is basicaly not cost effective as the requirments for manpower space and excessive time delays would make most people use an alternative.

Security checks only work at airports for three basic reasons,

1) There is little practical alternative to flying.
2) The majority of flights are point to point.
3) Passenger checking is a function of the terminus not the transportation.

Try putting passenger screaning on the plane door and making the plane do ten or twenty stoips every five minutes and watch the ensuing mayhem...

RoxanneAugust 4, 2008 8:28 AM

Most people in America just don't take the bus in the first place. It's regarded as a useful service only for the young and the poor.

If he had strangled his seatmate with his belt instead, would we demand that men with belts not be allowed to ride the bus?

My real question in this story is why none of the OTHER passengers jumped on the guy with the knife, but I don't have all the details, by a long shot.

ArkhAugust 4, 2008 8:49 AM

"I’m in a wheelchair and what would I be able to do to defend myself? Probably nothing. So that’s really scary."

He could have a gun. Which could even be used to defend himself at airport, in school, in a dark alley, wherever and whenever he might be attacked.

EntitlementAugust 4, 2008 9:17 AM

Wow, reading stories like this and the "scared" responses leaves me amazed. I expect the hysterical responses are used for "press appeal" to sensationalize these stories. I doubt they would publish the accounts of those that would say, "... that's the risk of living in a free society. I think we are still safe enough and we don't need any additional security measures"

The feeling one gets is that there is a growing sense of entitlement to always be "safe", no matter what one is doing.

Sorry, there are risks that come with living in a free society, and one of those risks is that some crazy person can randomly injure/kill you or someone you know. Can we all be made "safer" against these rare and random acts? Sure, but at what cost? The cost is the loss of basic freedoms for everyone and the eventual deterioration of free society.

RichAugust 4, 2008 9:21 AM

This isn't the first time a major incident has occurred on a bus. Take the 2005 London bus bombings, for example.

TfL's response to that incident was the following policy:

http://www.tfl.gov.uk/corporate/media/newscentre/...

excerpt:
"One of the three Tube posters advises passengers what extra security measures have been taken on the Underground. Under the headline "Working to keep your Tube safe", it says:

* More police patrols and search teams on trains and stations
* Random police checks at stations
* Additional checks on trains
* Extra staff patrols
* Police dogs on stations and trains

Another Tube poster and the bus poster seek continued passenger support, asking that unattended items and suspicious activity is reported to staff or the Police and that passengers hold on to their belongings and don't leave litter behind."

I think they realized then that metal detectors and airport-style security checkpoints weren't feasible for metro busses (for all the reasons previous commenters have mentioned), but it didn't prevent them from incorporating some additional security measures. What I don't have information on is how effective these measures were in preventing similar incidents.

ZachAugust 4, 2008 9:27 AM

It seems to me that Ms. Ryder has presented a good argument for allowing law-abiding citizens to carry concealed handguns, especially on a bus. It would allow her to defend herself in exactly that type of situation.

I'm not suggesting that someone with a gun would have stopped the killer from killing, but it certainly would have stopped the horrible desecration the killer performed, as well as preventing him from moving on to other victims.

A bus is often on the road in desolated areas for hours at a time. Can we really expect to have police presence within 10-15 minutes of all our busses? I know there are cross-country busses that find themselves driving at night where there is likely NOBODY for 20 or 30 miles, let alone police. Given that we can't keep weapons out of prisons, do we really want a bus where only the bad guys have weapons?

Gun GnutAugust 4, 2008 9:35 AM

>"Hearing about this incident really worries me,"
>said Donna Ryder, 56, who was waiting
>Thursday at the bus depot in Toronto.
"I’m in a wheelchair and what would I be able to
>do to defend myself? Probably nothing.
>So that’s really scary."

And after the metal detectors are installed, she still won't be able to defend herself.

As Arkh at August 4, 2008 8:49 AM said, get a gun if you want to defend yourself. Otherwise, you'll have to depend on others.

http://www.a-human-right.com/dependonothers.html

Michael AshAugust 4, 2008 9:48 AM

I'd like to know how Donna Ryder defends herself when she's not on the bus. Probably doesn't, and probably doesn't have a problem with that, even though I imagine she's much more likely to be attacked outside.

IngvarAugust 4, 2008 9:50 AM

@Zach: The fact that buses in countries all over the world are driving long hours in desolated areas, but very few incidents, is a sign to me that more guns would just cause the incident rate go up. A few trigger-happy guys wanting to thwart suspicious activity, people who try to take the trigger-happy down...

No, it's quite OK as it is. The risky part with that bus trip is if you cross the street just behind the bus when you reached your destination.

CalumAugust 4, 2008 9:58 AM

@Rich - It's worth remembering that London has 30-odd years of experience in dealing with bombings and bomb hoaxes due to the IRA and company. They have a pretty good idea what works and what is security theatre.

As for guns, I do not want anybody waving a handgun around on a crowded bus, no matter how capable they think they are. Adrenalin makes for some pretty wild shooting even amongst professionals. As for depending on others, I presume you also grow your own food and smelt your own metals? Civilisation is based on division of labour; there's no reason personal security should be an exception.

Michael AshAugust 4, 2008 10:19 AM

@Calum

"Civilisation is based on division of labour; there's no reason personal security should be an exception."

Actually there's a very good reason: division of labor in the field of personal security doesn't really work all that well.

You mention food and metals. Those work because the food and metals producers have a good incentive to do what they do, and because their product can be finely divided. I give them money in exchange for their product, and a single agricultural worker can produce enough food for dozens of people.

Now take security. I could hire a bodyguard, but I don't have enough money. I could go in with a few dozen other people and collectively hire a bodyguard (or go in with the whole city and hire a bunch of police), but a bodyguard's services don't divide very well. You can take an agricultural worker's output and divide it up among 30 people who take it where they need it. But you can't take a policeman's product and divide it 30 ways in the general case.

So in the case of personal protection, you can't rely completely on other people. You can rely on them a bit to protect you in some situations, and you can rely on security professionals to provide a deterrent which lowers the overall crime rate, but you can't rely on them to protect you in many circumstances. The choice comes down to either taking your chances or becoming capable of defending yourself.

SeanAugust 4, 2008 10:21 AM

Afraid of the shadows, jumping at every sound, the coward dies a thousand deaths.

Do we as a nation really want to embrace this as our base outlook on life? Many of the replies to columns that discuss happenings like this seem to think that we do.

Where would this nation be if all our "Pioneers" were afraid of the unknown and quite rare incident. Funny thing, they built the nation while facing certain death from guaranteed threats, not by freaking out at shadows, but pressing on despite them.

There is no such thing as absolute safety or security. Face your mortality and live your life.

ZachAugust 4, 2008 10:30 AM

@Ingvar:

I'm not suggesting more guns. I should be suggesting that we not deliberately reduce that number by banning them, but I didn't word it quite that way originally.

We currently don't do any screening, so you don't know if the person boarding has a weapon. If you say that weapons are not allowed on the bus, but don't do any screening, then only people who don't follow the rules will have weapons. Is that the situation you want?

Like it or not, criminals will carry weapons into places you don't want them to. If they start using them on the bus you're on, what do you want to happen, and what do you think will happen?

Personally, I want someone to stop the aggressor by any means necessary. I think that is more likely to happen when they are able to disable the aggressor without having to get within reach.

This doesn't mean I think everyone should be armed. Most people who get a CCW permit go through some training and classes, and even continue to practice and train for new situations after they've received the permit. There is a large list of behavior that will get your permit yanked in most places, including misdemeanor crimes (such as brandishing, which can usually be construed as any action that reveals the gun in public without the intent to use it for defense.) As a result, most people who carry are very cautious about their weapon, and will not use it for borderline situations that they may be able to resolve without force.

AkosAugust 4, 2008 11:09 AM

@Clive Robinson
"Try putting passenger screaning on the plane door"
That's exactly what they do at Kiev Borispol airport.
X-ray before boarding the plane: the problem is that you're not allowed to carry on liquids unless you bought them in the duty free shop, and they can only guess where you bought your on-board vodka, so they confiscate shampoo but look the other way if it's alcohol.

AkosAugust 4, 2008 11:16 AM

"I’m in a wheelchair and what would I be able to
do to defend myself? Probably nothing.
So that’s really scary."

Don't worry, after the metal detectors are installed, you won't be allowed to carry any metal object onto the bus. So start saving for a carbon-fibre wheelchair and hope that there is no such thing as carbon-fibre knife. Or is there?

BasilAugust 4, 2008 11:40 AM

Soon after this happened, the Canadian minister for public safety said in the newspapers that the government is not looking into any knife or bus security measures because of the randomness of the attack, and the fact that anyone can get a deadly knife and use it for their kitchen.

Davi OttenheimerAugust 4, 2008 11:43 AM

I haven't read much about the attacker's profile. I sense that will be the bigger question of future security systems. In other words, can passengers be screened for symptoms when buying a ticket?

Most comments, including yours and Milan's, seem to center around detection of a weapon, rather than detection of the probability of attack. Looking for weapons is old-skool and ineffective, yes. Comments on looking for someone predisposed to health-risks such as violence upon others?

Chris MoorhouseAugust 4, 2008 12:55 PM

@Davi

I wondered when someone would come up with that. So, should we require medical stability/competency certificates just to ride the bus, or should it be more of a spot-check situation, requiring everyone to carry certificates at all times, along with ID?

Can I choose the doctor that issues the certificate, as is the case now for checks for health care workers in Canada? If my regular doctor won't issue one for me, can I get one in a walk-in clinic? What if my doctor is away, and I need to travel for a family emergency?

Or should we institute a government service that employs psychiatrists to do nothing but attempt to assess the stability of a bunch of strangers on a daily basis, in an effort "not to restrict a citizen's freedom to travel".

HUGE can of worms.

baldheadedguyAugust 4, 2008 1:05 PM

I have to wonder if the lady in the wheelchair has considered that she would be no more likely to defend herself from a crazy person on the street, than she would be on a bus? I mean, really, you take a risk of encountering said crazy man when you go out your front door. From what I read, this was truly random, and for that to be the case, the man could not have been well.

Davi OttenheimerAugust 4, 2008 1:17 PM

@ Chris Moorhouse

I think that's the real future of this issue, and should be discussed openly instead of the dead horse of TSA scanning for dangerous objects.

A long time ago, well at least a few years back anyway, when the TSA first started their bone-headed screening strategy many people brought up the Israeli airline safety record and the fact that passenger interviews were done by trained psych and profile specialists before you could enter a plane.

Today? On-line booking might be a quagmire because it requires information sharing and causes huge data (store and transmit) privacy issues, but training drivers or even guards on buses to screen passengers could provide a number of advantages in terms of customer service beyond just being an annoying control point.

I haven't seen it discussed elsewhere on this issue. Maybe people have just given up on figuring out intelligent/useful security and gone for the reactionary theater as Bruce often bleats.

CarlAugust 4, 2008 1:18 PM

Do we know that the bus attack was random?
The linked article doesn't address motive.

TSAugust 4, 2008 1:30 PM

@Jeroen

You're missing the point. Antibozo has a greater chance of getting mugged, not because his friend got mugged, but because he is putting himself in the same situation and circumstances where his friend got mugged.

Remember, mugging, unlike beheadings on Canadian buses, are not particularly uncommon, and are not completely random. Muggers like to work in certain areas, at certain times, and like certain targets.

If antibozo presents a similar profile as his friend, and walks into the same situation as his friend, he is certainly at greater risk than someone who avoids the area and/or does not look like an easy target.

Think of it this way. AB's friend dies in a skydiving accident. Antibozo also runs a higher risk of plummeting to his death, because they are friends and do the same things. AB's friend's death doesn't have any influence on antibozo's own chances of death, but antibozo still has a higher risk of death than others who like to keep their feet planted firmly on the ground to begin with.

DavidAugust 4, 2008 2:01 PM

"Entitlement": Right on. Freedom isn't free: the price is eternal vigilance ... and insecurity. I'll take it, thanks.

Pat CahalanAugust 4, 2008 2:03 PM

@ Bruce

> "Man beheads girlfriend on Santorini island." Do we need
> airport-style security measures for Greek islands, too?

(noise)

Clearly you're racially stereotyping here, Bruce. This incident only proves that we need airport-style security measures for *islands* in general. Further study would be required before we could correlate any relevance between this incident and the national association of this particular island with Greece.

(/noise)

Gun GnutAugust 4, 2008 2:54 PM

>I haven't read much about the attacker's profile.
>I sense that will be the bigger question of future security systems.
>In other words, can passengers be screened for symptoms when buying a ticket?

The Brady Bill brought us the "no guns" list.

Then the Patriot Act brought us the "no fly" list.

Now we need a "no bus ticket" list.

I'm sure they can use the Brady and/or TSA screening system at the bus ticket counter.

PhilipAugust 4, 2008 4:32 PM

@TS

No, the point is that antibozo doesn't really know what the chance of his being mugged is because a sample size of 1 is all but meaningless for stuff like this. It has no chance of filtering out the effects of chance. Or, to put it another way, how can you tell if that event was signal or noise?

What do you gain from knowning that your friend got mugged the one time he visited somewhere? What can you conclude, even with the overlap in you and your friend's tastes? That there's a 100% chance that you'll be mugged if you do the same stuff he did? That's what the sample says (1 mugging in 1 visit for your friend), but how realistic is that? (And the bigger issue is this: how can you tell how realistic it is without additional data?)

annAugust 4, 2008 4:46 PM

Apropos the guns, thank goodness, here in Canada neither the criminal nor the passengers are likely to have guns - possible but still not that likely. Note that the madman used a knife NOT a gun and if anyone had had a gun, the probability of other victims besides the one who was killed would have been greatly increased.

Trying to stop the attack would only have gotten someone else hurt - by the time anyone had figured what had happened, the victim was well on his way to death. The crazy mutilated the body AFTER death.

antibozoAugust 4, 2008 5:15 PM

Philip> That there's a 100% chance that you'll be mugged if you do the same stuff he did?

Obviously this would not be the conclusion. But you're throwing straw men around at this point.

Look at it this way: suppose I have an aggregate statistic from a reputable source that, of 100 people who visited Washington, D.C, 5 were mugged. And suppose I have 100 friends who visited Washington, D.C., and of those, 10 were mugged. The statistic based on my friends is more predictive for me, because I form a more targeted cohort with that group than the anonymous 100 who produced the first statistic.

If Bruce had compared anonymous crime statistics with a story told by someone I meet randomly in the subway, I wouldn't quibble. But friends are not random variables; they are more predictive, per capita, than anonymous strangers, because my trajectory correlates more strongly with theirs. That's why we're friends.

JilaraAugust 4, 2008 5:23 PM

There is an easy way to resolve all this. Do people really NEED to travel? If you have a clear need to travel, perhaps because of business or a personal emergency, you can apply for a travel permit. After appropriate background checks, you will be escorted by an armed guard to your personal travel container, which looks very much like a cell, on the travel provider of choice (bus, train, plane, etc.) When you reach your destination, you will be released and escorted to a transfer point.

People travel too much nowadays, and have started to regard this freedom of relocation as a right, not a priviledge granted by the powers that be. Remember, in the old USSR, people didn't need to travel and maps were strictly illegal, and aside from a few malcontents that would be better in Siberia, people functioned just fine. Once upon a time, people were proud that they had never strayed more than ten miles from home in their entire lives. Nowadays, people seem to think they have a right to move freely about the planet. They want to go somewhere? Let them walk! With periodic checkpoints along the way, of course. Everyone will be safer, that way.

Gun GnutAugust 4, 2008 5:31 PM

>Use the fire extinguisher as a weapon.

Only professional firefighters should have fire extinguishers.

Besides, most people don't know how to use one. The criminal will just take the fire extinguisher away from you and use it against you.

Inn JolmeAugust 4, 2008 5:39 PM

"Of course, airplane security won't work on busses."
Not surprising, considering it doesn't work on airplanes...

PhilipAugust 4, 2008 5:50 PM

@antibozo:

Yeah, a statistic based on experiences of people who will be doing stuff more like you would is going to be better than a statistic based on everybody. I wasn't trying to make a straw man; when I read your first post I thought you were claiming one friend's experience as worth more than the results of data gathered on annymous people (the "a friend" vs "abstract crime statistics" thing from the post) -- and a sample size of one is never very predictive.

Eric CramptonAugust 4, 2008 6:06 PM

Very nice op-ed in Canada's National Post.

"the West's legitimate security experts are working full-time these days trying to eliminate expensive forms of dimwit "security theatre" that already exist in our transport system - a time-consuming kabuki that plays out, for instance, every time we take off our shoes or are forced to surrender our toothpaste before boarding a flight. Most of them would be happy just to have a short breather from moral panics."

DavidAugust 4, 2008 6:34 PM

Yeah, chopping off heads is rare, but apparently chopping off feet is less so, with many washing up on Canadian beaches and such.

Paul RenaultAugust 4, 2008 7:23 PM

Thanks, Eric Crampton, for posting that link.

"Unfortunately, the West's legitimate security experts are working full-time these days trying to eliminate expensive forms of dimwit "security theatre" that already exist in our transport system - a time-consuming kabuki that plays out, for instance, every time we take off our shoes or are forced to surrender our toothpaste before boarding a flight."

Jonathan Kay and the editorial board of the National Post should be commended!!

FrancesAugust 4, 2008 9:53 PM

@David - the feet are not chopped off, they just separate from drowned bodies.

As for the wheelchair and Via Rail. No, we don't have a Canadians with Disabilities Act. Unfortunately. And, as far as I know, Via can accomodate wheelchairs, the problem is getting on to trains. Via does not have lifts available everywhere but there was a court case fairly recently in which they were told to get with it and put them in place.

AnonymousAugust 4, 2008 10:41 PM

@antibozo

"Look at it this way: suppose I have an aggregate statistic from a reputable source that, of 100 people who visited Washington, D.C, 5 were mugged. And suppose I have 100 friends who visited Washington, D.C., and of those, 10 were mugged. The statistic based on my friends is more predictive for me, because I form a more targeted cohort with that group than the anonymous 100 who produced the first statistic."

If the the muggings are drawn from a binomial distribution, your samples are only barely distinguishable at the "usual" 0.05 level.

antibozoAugust 4, 2008 11:04 PM

Anonymous> If the the muggings are drawn from a binomial distribution, your samples are only barely distinguishable at the "usual" 0.05 level.

Are you determined to miss the point? Okay, let's say that of my 100 friends, 50 were mugged. Is that distinguishable enough for you?

It doesn't matter, because that isn't the point; it's only meant to illustrate. The point is that a cohort composed of my friends is more predictive for me than a random cohort, and that it takes a smaller sample of friends than of random observers to reach a given confidence.

Does that penetrate?

Michael AshAugust 4, 2008 11:14 PM

@ antibozo

I don't think anyone would say that you should ignore evidence pulled from a sample of 100 friends. But this is entirely different from reacting to the experience of ONE friend, which is pointless because a sample size of one tells you absolutely nothing beyond that the event is possible, no matter how closely correlated your habits are.

clvrmnkyAugust 4, 2008 11:17 PM

@antibozo: "A friend's personal experience in a foreign country is more predictive of what will happen to me than the experience of an anonymous person which was used to compile data."

Then what exactly are you trying to say here? If it isn't something about how news like this is or isn't valid information for personal safety, then what? The information you get from your friend is only more or less "predictive" for a certain set of situations and results.

Recently a fair number of Canadians have died while on vacation in Mexico -- or rather, such deaths have made it to the news and now appear to be a trend (while the data does not bear this out). Many Canadians have responded by changing their travel plans. This is both irrational and understandable. Here is an example where close-to-home news is not predictive. It just isn't true that Canadians are more likely to die while on vacation in Mexico; not any more than they might while on vacation anywhere far from home.

Using this particular tragedy as fuel for /any/ discussion regarding security or public safety is pointless. There is nothing we can do to guard against random attacks by mentally ill people living among us.

Which, really, is all I got from Bruce's comments.

clvrmnkyAugust 4, 2008 11:37 PM

@Roxanne: "My real question in this story is why none of the OTHER passengers jumped on the guy with the knife, but I don't have all the details, by a long shot."

Because he was a crazy person with knife who suddenly and without warning attacked his neighbour.

As much as we like to believe that we are the hero of our own story, until you are confronted with a situation where everything you know is wrong you will never know how you will truly react. Few people are trained how to disarm a determined, violent person with a knife, and fewer still have the training to take themselves out of the everyday and attempt to do it in an extremely restricted environment. All while other people are also reacting in their specific way.

We've all been there: you are doing something mundane like waiting in line, only paying enough attention to maintain the illusion of consciousness as we go over something in our minds -- daydreaming. And then someone breaks the social contract by doing something sudden and unexpected (however benign). Sure we snap out of our reverie, but few of us can do more than simply stare or react. We are not trained professionals who are on the constant lookout for hinky behaviour. Only cops and psychopaths are that in the moment all the time. I'm not even sure about the cops.

Now, take that same feeling and magnify it to horrific levels.

From what we know about this story is that the attacker made it clear in seconds that he wanted everyone off the bus while he busied himself with his efforts. The other passengers obliged him because it was the most sensible thing to do given how insensible the entire situation became in mere seconds.

Remember, from what we know about this story, the attacker did not brandish a knife early on, or warn people, or otherwise draw attention to himself. Since people on long-haul public transportation studiously ignore each other, he would have had to acted seriously
Not Quite Right to show up on any personal safety radar.

It is foolish to consider what we would have done in the same situation, or assert we would have somehow behaved any different than most: get the hell off the bus.

Rob LewisAugust 5, 2008 12:24 AM

@clvrmnky, (11:17)

While there may have been a handful of incidents or murder victims out of a about a million vactioners in Mexico, the decision to vacation elsewhere might be due to accounts of corrupt and inept police and court systems there. There may not be any more risk of incident, but I would chose to vacation in some other banana republic where there are not regular media accounts of the troubles that begin after the incident as taken place.

Dimitris AndrakakisAugust 5, 2008 2:22 AM

> "Man beheads girlfriend on Santorini island."
> Do we need airport-style security measures for Greek islands, too?

Obviously no. We need them on girlfriends :-)

BTW, as one would imagine, the media over here (:= Greece) is throwing a "shock and horror" party on the issue.

AnonymousAugust 5, 2008 5:35 AM

@antibozo:

"Are you determined to miss the point? Okay, let's say that of my 100 friends, 50 were mugged. Is that distinguishable enough for you?"

If 50 of your friends were mugged, then chances are 40 or 60 strangers were mugged. Basically, this is a dangerous place to go, friend or not.

To highly the fallacy you are partaking even more:

Suppose one million people traveled somewhere, and 100 were mugged. Then we gather up all these 100 people in a convention and proceed to find similarities between them and you.

There will be lots and lots of them.

Question: do you travel to that place or not?

B.F. SkinnerAugust 5, 2008 10:40 AM

We are neurologically hardwired it seems to believe those people we know (the storyteller) to faceless government/scientific drones who produce endless unread statistics.

Good book on the subject just came out
The Science of Fear by Daniel Gardner.

Since this IS the way we are, as humans, the question is how do we program ourselves, our neighbors and our society?

Adam WightmanAugust 5, 2008 10:51 AM

The prevailance of fear amongst the western public will only result in the elimination of our free society - through our fear, we will allow the government to take our freedoms away to make us safe. However, that will not make us safe, but likely will make us much less safe. So, people get killed on the street, does that mean we should have checkpoints on the street, like Baghdad? If you fear mongers keep going like this, we will have passed all of our freedoms to some big protective government. Well, hate to tell you, you are a free autonomous person, that is your nature as a living individual, government is not supposed to take care of you like some nanny. If you think that they are supposed to, you are not even truly human. Government should exist to ensure private property, apprehend criminals and,where it is feasible, take care of sick people.

antibozoAugust 5, 2008 11:13 AM

Anonymous> If 50 of your friends were mugged, then chances are 40 or 60 strangers were mugged. Basically, this is a dangerous place to go, friend or not.

And yet, back when you were paying attention, we already established that only 5 strangers were mugged. Which means that my friends are statistically much more likely than an average person to be mugged in Washington, D.C., which means that *I* am statistically more likely to be mugged in Washington, D.C., since whatever high-risk behavior they must be doing in order to get mugged so frequently I'm more likely than the average person to be doing also, because these friends are pre-selected to behave similarly to me, by virtue of being my friends.

Anonymous> Suppose one million people traveled somewhere, and 100 were mugged. Then we gather up all these 100 people in a convention and proceed to find similarities between them and you.
Anonymous> There will be lots and lots of them.

Yes--in fact, there will be statistically more similarities between me and those people than between me and 100 randomly selected people. Guess what: mugging is not a random variable; it subsets people by behavior.

Anonymous> Question: do you travel to that place or not?

100 out of a million may be too small a statistic to alter my behavior.

If you still don't get it, oh well... you can lead a horse to water, as they say.

antibozoAugust 5, 2008 11:17 AM

antibozo> Yes--in fact, there will be statistically more similarities between me and those people than between me and 100 randomly selected people. Guess what: mugging is not a random variable; it subsets people by behavior.

Excuse me; haven't had coffee yet. There will be statistically more similarities between one of those people and the other 99, than between a randomly selected person and 99 of them.

So if, upon meeting them, I would consider them friends, then their experience is more statistically significant to me. But again, 100 out of a million may be too small to alter my behavior.

StatsAugust 5, 2008 11:31 AM

@antibozo

Dude, you need to learn some basic stats.

All the examples you give do not in anyway indicate that you are outside NORMAL randomness no matter HOW MUCH you and your friends are correlated.

In god we trust, as for the rest of you, show me the data.

That quote has one thing missing. Lots and lots of people still don't believe data over stories, and you are one of em.

Enjoy your stories.

antibozoAugust 5, 2008 11:45 AM

Stats> All the examples you give do not in anyway indicate that you are outside NORMAL randomness no matter HOW MUCH you and your friends are correlated.

So you don't believe that self-selected cohorts will exhibit statistical similarities that distinguish them from randomly selected cohorts? Impressive. I guess we don't need that whole random sampling thing after all.

kangarooAugust 5, 2008 4:25 PM

Stats> All the examples you give do not in anyway indicate that you are outside NORMAL randomness no matter HOW MUCH you and your friends are correlated.

See antibozo, you seem to think that we aren't a billion anonymous particles with velocities falling under a normal Gaussian distribution.

You are so crazy! You must study something unscientific, like hard-sphere packing effects, where the density of particles against a hard-wall is much more accurately determined by sampling the radial distribution function around a particular particle, than the mean radial distribution around all particles.

I mean you just gotta take stats 101 and learn that everything is independent and Gaussian! There is no such thing as multi-modal distributions, or Cauchy distributions with no mean or standard deviations, or dynamical non-equilibrium systems where the least likely trajectory for any particular particle is the mean trajectory, or bifurcational phenomena or ...

You some kinda anti-intellectual?

antibozoAugust 5, 2008 10:00 PM

kangaroo> You some kinda anti-intellectual?

I guess I must be... Thanks for the laugh! ;^)

jsAugust 6, 2008 4:05 AM

I think this sums it up best (from the National Post editorial): "Let's not try to protect ourselves from crazy people by trying to out-crazy them."

bobAugust 6, 2008 7:29 AM

@Clive Robinson: There isn't any queue, we're talking about podunkville-a (since its in Canada) somewhere with no more than 3 people getting on. Besides I'm not advocating metal detectors, just saying that buses traveling in brownian motion picking up random pax here and there in tiny quantities bear no similarity to airliners which take a big stack of people from one point to another point all at once and therefore saying that you could not have airport-style metal detectors at every stop is irrelevant.

And finally, you could easily put a security guy on every bus to run the machinery and do pat-downs and stuff and it would be consistent (if not cheaper) with the amount of money they are spending on airports.

@Ann: "neither the criminal nor the passengers are likely to have guns."

- True in the US as well. However, since firearm possession by ordinary people is basically unlawful; the criminal is far MORE likely to have one since he doesnt give a shit what you, anybody else or "the law" (you collectively) thinks.

@Ann: "...if anyone had had a gun, the probability of other victims besides the one who was killed would have been greatly increased."

- You make an awful lot of stew with no oysters at all. Backup that statement with some (relevant) facts. In Ohio (as in each of the 45 states) the anti-gunners made similar predictions when we passed concealed carry 5 years ago. Since then there has been -->>NOT ONE SINGLE

You also leave out the fact that if anyone else had a gun the probability that the VICTIM would NOT have been killed increases.

Canada does not have "gun-free" buses, they have buses where ONLY the whackos and thugs carry guns. The saving grace is that there arent very many of them (just like in the US).

Predicating your safety on the idea that irrational people can be made to behave rationally by passing a law telling them to be rational doesn't even qualify as theater.

AnonymousAugust 6, 2008 8:05 AM

@kangaroo

"I mean you just gotta take stats 101 and learn that everything is independent and Gaussian! "

I have no idea who raised this Gaussian stuff. Do you?

And conditioning, independence, and so forth, are crucial concepts in probability and statistics. You do indeed learn this in "stats 101", and it's just as relevant to "multi modal distributions, [...]" as anything else.

"You some kinda anti-intellectual?"

Alice went to the beach and got mugged.

Bob went to the beach and got mugged.

Charlie went to the beach and got mugged.

antibozo now stands up and says "Hey! Alice, Bob, and Charlie are my friends! Because they are the same as me, I'm gonna get mugged too!"

Then Dave, Eve, and Freda -- who antibozo doesn't know -- went to the beach and got mugged.

As near as I can understand antibozo, he now asserts that Dave, Eve, and Freda would actually be his friends -- if he only got to know them.

(At this point, what few readers are left can check the victimology article at Wikipedia).

antibozo also says that if a sufficiently large number of people go to the beach and do not get mugged, it no longer matters if he knew the mugging victims or not.

The latter situation -- almost no one gets mugged -- is trivially observed reality. Which makes his "friends of victims become victims" argument look prima facie untenable.

But as they say, you are free to believe whatever you want to believe. This doesn't make one an anti-intellectual though.

kangarooAugust 6, 2008 12:11 PM

Alice went to the beach and got mugged. She was wearing a red hat.

Bob went to the beach and got mugged. He was wearing a red hat.

Charlie went to the beach and got mugged. He too was wearing a red hat.

Then we find out that Dave, Eva and Freda, who got mugged at the beach, also wore a red hat.

Then we also find out that all of them were transgendered Australian aborigines born in 1965 who like to smoke pot and are world-famous ballroom dancers.

Of course, it's trivially true that almost no one gets mugged at the beach. It's also true that no one has collected data in the past on rate of muggings at that beach of transgendered Australian aborigines born in 1965 who like to smoke pot and are world-famous ballroom dancers, who also happened to be wearing red hats. And of course there's no possibility in a theory-free way to make a statistical analysis of that even if you collected the data, because it would be a post-hoc data collection.

So what are we left with? Not all problems are simply problems of statistical analysis - particularly problems involving extremely complicated phenomena. Sometimes some thread can be pulled out to reduce the problem to a statistical analysis, but sometimes you can't, and have to investigate critically as a one-shot.

As I said, some problems are like looking at particles in infinite volume under pressure. And some problems are like understanding a piece of literature. Peoples lives are often like the latter.

Lots of science has been done with the latter approach. Some events are naturally narrative and unrepeatable -- see the history of the development of geological ages in 19th century Europe (Murchison et. al.).

It's the problem with simple risk assessments --- sometimes peoples intuitions are wrong, but sometimes the statistical analyses that we use to tell them that they are wrong is itself meaningless.

==== Metaphor Alert! =========

Like the assessment of discount rates. It's an old story that people don't behave as economic models of discount rates predict that they should behave -- that "they're irrational" to take a $1.10 in ten minutes for a $1 now, but not take a $1.21 in 20 minutes. But it's the discount rate that's wrong -- the loss of information about the future doesn't fall off with a constant acceleration, but is more than exponential. So people's intuition is right; economic theory is wrong. Messy systems sometimes just don't fit the simple math.

Data may not be the plural of anecdote, but sometimes "data" is just an anecdote.

antibozoAugust 6, 2008 1:59 PM

Anonymous> As near as I can understand antibozo, he now asserts that Dave, Eve, and Freda would actually be his friends -- if he only got to know them

No. What I wrote was that *if*, having met Dave, Eve, and Freda, I would consider them friends, then I would assign more weight to their data than I would to data from George, Helen, and Imogen, whom I've never met.

If you still don't understand what I actually did write (rather than whatever deluded misintepretation you've concocted in your head), just go back and read it a few more times. I'm not going to waste any more time repeating it yet again for the benefit of someone who is determined not to comprehend it.

antimediaAugust 6, 2008 5:54 PM

@pound sand "800,000 people die in a tragic human caused event in Rwanda, in 93, I didn't know any of them. I recognize this as an abstract tradgedy. 3,000 people die in twin towers in New York, they speak the same language as I do, but they wouldn't likely speak to me because of class snobbery. How is that different from a million cambodians who live under a 5 year rain of bombs from the US Airforce?"

It's different in that the first two events are historical facts and the latter "event" is a figment of your imagination.

Some statsAugust 7, 2008 6:00 AM

@antibozo

I have concluded that, based on the evidence above, you are in fact a wanker.

You put folk down about not reading your posts when you clearly have not read theres either...

Think we should have taken note of the whoosh post.

ModeratorAugust 7, 2008 4:31 PM

Greg,

Calling someone a wanker is childish; changing names to do it is cowardly as well. Stop using sockpuppets if you don't want to be banned.

gregAugust 8, 2008 3:26 AM

@Moderator

This was not me.. and the computer I'm on is not single user and I'm not root.

If I would call someone that, i would not hide it and I would not do it from a machine with a static ip. I'm not stupid. Dam almost all my post here have been from 2 computers with static IPs. I know you know if I use a different name (usally something like lumberjack).

I think I know who it is but cannot prove anything, not that it really matters.

For what its worth, Sorry. And also why don't you delete it?

ps I would sign things.... but this hardly seems worth it.

antibozoAugust 8, 2008 4:03 AM

greg, Moderator,

Don't delete it on my account. I got a good laugh out of it. :^) Cheers...

gregAugust 8, 2008 4:41 AM

@antibozo

dude that was fast...Do you notifications or something?

I would prefer it deleted... My permanent internet record is bad enough with my own bad days without others adding to it.

On second thought, its a pretty minor infraction on the list. ;)


sarahAugust 9, 2008 8:01 PM

I have a hard time understanding why no one intervened on the Canadian bus to help stop the killer. Maybe I have not read enough about it, but I can't believe so many people would stand by without doing anything.

EvieAugust 9, 2008 10:21 PM

Let me respond, please, to some of the comments made by "Sarah" and others.

You might want to refer to news clips put out by Canadian media, ie CTV has some on YouTube and other places.

To start off, let me remind you that everyone onboard was at a particular disadvantage: it was late at night, it was dark, many passengers were sleeping and at first didn't know what was happening. However, people DID attempt to disable Tim's attacker. But he was a Big man with an even bigger - Rambo style - knife. As they moved towards him, he took Tim's head clean off.
Let me ask you, if you were face to face with an armed man strong enough to decapitate someone, and he had done so right in front of you, how "brave" do you think you would suddenly be? People typically overestimate their ability to courageous in dangerous situations.

Regardless, let me close by saying that people were in fact brave that night. They got everyone else off the bus safely, and then prevented Tim's attacker from leaving the scene by blocking the door and disabling the bus. These actions probably saved more lives. That is worth something, don't you think?

AnonymousAugust 12, 2008 2:01 AM

Being 5ft3 and 125 lbs doesnt help matters much either. Statistically psychopaths are known to prey on smaller people. They perceive them as weak and incapable of putting up much resistance. This was premeditated murder on Li's part hence his choosing the smallest and quietest man on the bus. Li obviously suffers from the inability to stand up for himself according to his history. So he chose a very small man to take out his lifetime of frustrations on. They should execute him publicly and brutally for all his kind to see. The Romans new how to execute ruthless criminals properly in order to make a point. Li should be crucified and left to rot in the hot sun...

EvieAugust 13, 2008 10:31 PM

Hey, Anonymous...wouldn't happen, my friend - Canada hasn't had the death penalty since July 1976; they won't bring it back for Li or anyone else. Although I share your frustration in many respects, I side with the government on the issue of capital punishment. An eye for an eye makes us all blind and ultimately solves nothing.

Anyway, when I think of Li spending each day, hour by hour in Kingston Pen, for however long he lives, I think that's just fine. My tax dollars can pay for that. Let him live with what he did, and let Tim rest in peace.

Your stats re serial killers are a grim bedtime story. Incidentally I am not much bigger than that myself...I am feeling like I should either a) start working out at the gym or b) get a big dog. Maybe both.

Leave a comment

Allowed HTML: <a href="URL"> • <em> <cite> <i> • <strong> <b> • <sub> <sup> • <ul> <ol> <li> • <blockquote> <pre>

Photo of Bruce Schneier by Per Ervland.

Schneier on Security is a personal website. Opinions expressed are not necessarily those of Co3 Systems, Inc..