Schneier on Security
A blog covering security and security technology.
« Hiring Hackers |
| Ninth Workshop on Economics and Information Security »
June 10, 2010
The "Quake" Simulation and Risk Perception
EDITED TO ADD (6/10): Another article.
Posted on June 10, 2010 at 7:10 AM
• 40 Comments
To receive these entries once a month by e-mail, sign up for the Crypto-Gram Newsletter.
I have seen the enemy, and he is us.
That is a nice eye-opener if I've ever seen one.
It basically comes down to: No matter what you know, you lose!
Interesting, no doubt. But I wonder how far their behaviour is affected by knowing that it is 'just' a game? It would be interesting (but perhaps problematic ethically) to play the game with real incentives and disincentives. Maybe the reward of a few (real) dollars for money earned in interest but losing a grade if they get wiped out.
There's a not altogether different aspect to real life disaster management: governments are often making decisions on someone else's behalf. The potential gains (at least votes, maybe slush) are personal, but the potential losses may be spread out amongst 'others' - ie the general population.
ok seriously now:
why should I plan for a earthquake when the government will bail out my carelessness?
I've paid a hefty premium for a quake resistant house and I'm insured against quake and fire (I live in a quacky zone)
I'm paying a lot for nothing, as recent event gave me the demonstration that the government will pay for the reconstruction of those that are homeless
the problem is: which is right? why I should pay my own security plus (via taxes) for the carelessness of others? I want a refund on my insurances and planning then - where is equal justice if the government constantly unbalances the rights of the righteous?
I find this incredibly scary, because it means that the way decision makers for society are chosen today dooms us if any disasters of a large magnitude are actually going to happen. The only option to avoid this seems to be to put real risk experts into power. Unfortunately, there are only few of those around, and while I know several, not one of them would go into politics unless forced to. (Not a surprise.) So the way to do this would be to have a) a process that reliably identifies true risk experts and b) force the responsibility to make far-reaching decisions for society on them.
While I cannot see a) or b) happening any time soon, this really seems to be the only option for survival as a society, unless we are very, very lucky over a very long time. Somehow I cannot see that happening either.
Come to think of it, global warming is one such foretold disaster. The stock market crash was another one, but we were hit only mildly, a larger crash is sure to come.
So...the lesson is that we should spend every penny we have RIGHT NOW on terror-prevention, right? Why keep the money in the bank, we know an attack will be coming, severe or mild...
"If everyone around you has a house of straw, having a straw house yourself seems somehow safer."
Okay. This describes almost every organization's relation to their IT security needs that I've eve seen. Including the explicit questions "What are our competitors doing on XX issue?"
In real life, wouldn't it be better to keep assets as liquid as possible rather than sink my money into a house that I may or may not be able to continue living in after a major quake? No amount of renovation is likely to protect me from every eventuality and a nest egg should help finance my forced relocation.
The scary thing is that this probably can be extrapolated to climate change, too.
Is it possible that once the players "know" how to win, they simply chose not to because going right for the win would be more boring? Trying to get lucky and make more money while also winning seems like a more interesting challenge than simply winning right away. What sort of incentive is offered to actually win?
It's gonna be awesome when Wharton gets sued by id software. :^)
I could get into a nasty car accident on my way home from work. Does that mean I should spend all of my money on better restraints and a roll cage and other enhancements to the car? Heck, forget that, I'll just go purchase a tank. That'll help me win right?
On a side note, playing a game is less about following the rules and achieving the end objective than it is about learning the rules, and making your own to completely own everyone else. I'm sure if some of these people were aware that REAL LIFE earth quakes were going to be coming in a series, there would be a mad rush to bolster defenses.
Yeah, certainly sounds less interesting to just build a strong house and sit there.
I bet they would behave differently if you told them they would fail the course (or whatever) if their house fell down.
"The only option to avoid this seems to be to put real risk experts into power. Unfortunately, there are only few of those around, and while I know several, not one of them would go into politics unless forced to."
Maybe the have figured it isn't worth the risk?
Hey, stop groaning!
@Wes P: "I could get into a nasty car accident on my way home from work. Does that mean I should spend all of my money on better restraints"
What would you do if you knew that you would *definitely* get into a car accident?
Because that would be the analogy to the experiment, as the players were told that there would be earthquakes: "Players are told at the start of the game that ... three to five quakes will hit during the course of the game."
I think it is best to buy a tent, when the quake appears it is more likely that you survive the quake. If the game only purpose is to earn money and survive, it is the best approach.
However, in the real life you have to defend of much more likely events, and also you want to spend the money in goods, and protect them from burglary.
@Paeniteo: I'd walk. And if I got told that 5 potentially very destructive quakes were imminent, I'd pack me and my family up and get the heck out of there as fast as possible, not pump all the money I had to be the last person on the block with my house standing.
But, I guess, my issue with this is is that disasters are never certain. Yea, theres certainly things we could prepare ourselves for more of: terrorist attacks, earthquakes, firestorms, nuclear meltdowns, alien invasions and the list goes on. But there's only so much money and resources with which to do that. And what if it doesn't happen... you're now stuck with a useless "solution", no money, and a FUD driven life.
I have a hard time viewing this as anything other than a game where 500 students have done nothing but try to win bigger than each other.
"If everyone around you has a house of straw, having a straw house yourself seems somehow safer."
If everyone in the game had a house of straw, they'd probably be pretty safe in any of the quakes ...
The San Francisco Bay Area is almost certain to have a catastrophic earthquake - The Big One - in our lifetime but average residents there act as though every day without a quake makes it less likely to happen. Those of us who acknowledged the risk thought we had prepared sufficiently and that the overpass was not likely to land on us. This was true enough for almost everyone in 1989. Reminds me of attitudes about risk in the Shuttle Orbiter program. Every successful flight was not proof of its safety, but simply another tick in the countdown to a catastrophic failure. Unfortunately there is no where else for the shuttle astronauts to be when things finally fail. As for the simulation, if disastrous quakes are certain to occur and the primary goal is avoiding harm, seems to me the book solution ought to be selling your house to another player and using the proceeds to buy a house off the board. But the issue ought to be framed as dynamic risk; what do we get in exchange for living in a quake zone (or flying the shuttle)? There's more to it than maximizes our bank accounts.
I will *definitely* eventually be killed. Does this mean that I should invest all of my assets, time and money, towards minimizing this eventually?
Overall, it occurs to me that too much is being read into this game simulation. It assumes that a strategy that fails in a very constrained set of circumstances is a "flaw", when in fact it might be the best strategy for overall success for the majority of the human race, even when it is an individual (or mass) disaster.
First of all, with no real penalty, most people will play a game which will result in guaranteed eventual failure (i.e. Space Invaders, Pac Man, Tetris, etc.) in exchange for the fun of trying to "beat" the game. If one day the people running this situation told their students how to win, and that 30% of their grade was resting on them getting at least one success while playing, I'd bet almost everyone will pass. They would sit down and play the "boring" version of putting all their money in the house until they won at least once, turn that time in for credit, and go back to having fun. Thus, we have to take any results we get with a grain of salt.
The second problem is the "all or nothing" aspect. It sounds as if all the players are hit by the *same* earthquakes, and each failure is final. In the real world, most disasters are localized, and the damage is rarely absolute.
Suppose you rewrote the game so that every player had 10 houses. The goal of the game was to both have houses at the end, with higher scores going to those who have a higher number, *and* having the most money, again with higher scores going to those who have the most. Make earthquakes destroy houses, but so long as you have houses left, you can spend money to rebuild houses.
Depending upon constraints, such as return on investment, cost of rebuilding houses, frequency and severity of earthquakes, etc., it is quite possible that the best winning strategy for the long run is "let them fall down, then build new ones".
This reminds of system safety research done by folks like Prof. Nancy Leveson at MIT:
"A corollary of this propensity for systems and people to adapt over time is that safety defenses are likely to degenerate systematically through time, particularly when pressure toward cost-effectiveness and increased productivity is the dominant element in decision making." ("A New Accident Model for Engineering Safer Systems," http://sunnyday.mit.edu/accidents/...
She has observed more or less the same phenomenon as the Wharton researchers in her analysis of accidents such as Bhopal and Chernobyl: As groups of people optimized their activities based on visible short-term gains, the non-occurrence of moderately likely accidents acclimatizes them to risk levels, and they systematically migrate to higher risk in a way that actually makes the accident more or less inevitable.
Something to think about in the wake of the Deepwater Horizon disaster.
I'd like to know whether the players learn to do better if they play the game over and over.
I would think that if the game goes on long enough the best strategy is to invest some of the money and divide the proceeds between improvements and reinvestment. In the long run this will produce a stronger house than the super-cautious strategy.
But if you don't upgrade your defenses right away, how will you fend off the inevitable Zergling rush?
Too many assumptions take place before the game starts:
1. Current building technology is sufficient to stand up to a sever earthquake. (Doubtful)
2. The earthquake will be scheduled to occur when the player is at home. (50/50--They might be at work or any number of other places.)
3. The house will be habitable and useful following a major earthquake. (It may be still standing but hard to access. A roof over our heads is only part of survival.)
4. Winning the game as defined by the creators is what interests the players the most.
5. Earthquakes are the only risks worth addressing. (In truth, most of us have limited resources and must make decisions about the best way to use them. The players know that earthquakes are just one of many risks to be faced.)
I agree with others who posit that the simple fact that it is a game may influence the player actions and outcomes.
That is actually a very good point. When the game is more complex and more rewarding, it seems as though players can be far more effective at responding to risk. This experiment is flawed because as far as I can tell, it lacks any sort of driving incentive to win.
Where can I PLAY the game?
It's impossible to judge how rational players' strategy actually is without knowing a lot more about the game (actually, about what THEY knew about the game).
Depending on the parameters, it's entirely possible that the rational strategy is to save your money until you reach a certain critical treshold where you can afford "enough" upgrades to make it worth spending...for example, if you need more than your starting money to buy enough upgrades to survive any quake at all.
And if players aren't given solid quantitative data bout when the quakes can occur, how severe they are, how much the upgrades cost, and how much they help--then they're simply making guesses about what parameters the instructors may have picked when they designed the game. Would you have guessed that they deliberately made it so that the optimal strategy was to spend everything at once and twiddle your thumbs for the rest of the game? That's not how most games work.
If the players DO have detailed knowledge of the mechanics, and their strategy is demonstrably irrational using that knowledge, THEN we might have an interesting result. But the article is way too vague to know if that's the case.
I'd be interested in seeing the results of the game with a pool of older players, particularly homeowners. I wonder if it makes any substantial difference.
The Loma Prieta quake resulted in 60-some deaths in a metropolitan area with a population of over 6 million at the time. I would argue that this indicates pretty good preparation.
Unless I read the article wrong this just profiles the behavior of students at Wharton.
"Kunreuther, the Cecilia Yen Koo Professor, and Meyer have run the Quake simulation for the past four years, using students in Kunreuther’s Risk Analysis and Environmental Management class as the guinea pigs/gamers. By now, about 500 students have played the game, and every time, they play it essentially the same way. "
Is it really a surprise that this group always "destroys themselves" for profit? Aside from the fact that games induce far riskier behavior because the penalties are fake (ever die playing Grand Theft Auto, or been kicked off stage in Guitar Hero?) the students at Wharton are likely to be far more profit-driven than a general population.
I'm not saying it's obviously a get-rich-or-die-trying culture there, or they need to re-evaluate their admissions process, but it also is not a fair sample and should not be extrapolated too far. I would wager a more general population that represents people outside Wharton would give different results.
Other research has found diversity in thinking:
"To test their idea that mixed groups would benefit survival, Ein-Dor and his colleagues put students in groups of threes alone in a room with a concealed smoke machine, which was switched on to simulate a fire. Groups were quicker to notice the smoke and to react to it if they contained individuals who scored high for insecure attachment."
@ all those who are saying this behavior is not realistic.
I see patients every day who will not invest small amounts of time and energy into simple behaviors that will significantly reduce their chances of having a fatal heart attack: moderate exercise, like walking, 60min per day when they are watching 4 hours of TV per day. From my perspective this simulation exactly models human decision making in health.
@ Petréa Mitchell
"The Loma Prieta quake resulted in 60-some deaths in a metropolitan area with a population of over 6 million at the time. I would argue that this indicates pretty good preparation."
Fair enough, but 42 of the 57 fatalities were on the Cypress Freeway in Oakland, a structure that Caltrans had been told would collapse in the event of a significant quake. The bureaucrats who placed the bet on leaving it open until retrofits could be done were not the people who paid the human cost for its failure.
"The bureaucrats who placed the bet on leaving it open until retrofits could be done were not the people who paid the human cost for its failure."
Closing a freeway in a busy metropolitan area is not without cost either. Maybe pedestrians will get hit by cars that are rerouted to surface streets. How much does it cost if 50,000 commuters have to spend an extra 15 minutes in traffic twice a day for a year? This is a real cost if you close a freeway. The earthquake is a remote possibility.
I am sure that the bridges and overpasses in the Bay Area will not withstand a magnitude 10 earthquake. Should we close all of them until they are retrofitted? No? Will you blame the politicians if a magnitude 10 earthquake wipes out the Bay Area?
Being alive is not without risk.
Stop watching TV. A magnitude 10 earthquake is not known to have ever occurred, and if one were to occur (which is essentially impossible, it would constitute a fault rupture thousands of miles long), it would border on an extinction-level event for humanity.
6.8+ earthquakes, on the other hand, are something that happen with disturbing frequency in California, and a freeway that can't withstand one has no business existing in California. It is not a question of if, it is when, and the "when" could be tomorrow just as easily as it could be ten years from now.
A good way to win the game would probably be to live in a tent and just rack up the interest payments. No way a collapsing tent will kill me. And after the last earthquake I could build an awesome house.
Maybe I'm more sensitive on the topic than I should be, but after all these years I still have stark memories of driving past the collapsed Cypress on the _emergency bypass_ while the rescue and recovery efforts were still under way. Are politicians responsible for protecting us from every negative occurance in our lives? Of course not; I'm still a Schneierista after all. But are our elected representatives responsible for telling us we are driving on roadways, overpasses, and bridges that will probably collapse under the effects of statistically probable events, so that we can decide whether to participate in the risk (or instruct them to reprioritize the budget, or vote for someone who will)?
"Fair enough, but 42 of the 57 fatalities were on the Cypress Freeway in Oakland, a structure that Caltrans had been told would collapse in the event of a significant quake."
So, one freeway that wasn't built up to code in the first place (I was in the Bay Area at the time too, and I remember the reports that the original plan required additional structural elements to support the upper deck) equals total unpreparedness?
But the Bay Bridge was damaged too, someone is going to say. But it was designed to do what it did-- if an earthquake (or whatever) was bad enough to start doing real damage, isolate the damage from the rest of the bridge by letting one segment fail. (IIRC, a different set of bolts broke first than the ones that were supposed to, but the end result was as intended.)
@ Petréa Mitchell
"So, one freeway that wasn't built up to code in the first place...equals total unpreparedness?"
I don't recall suggesting that preparedness was totally lacking. A lot of people, agencies, programs, and structures acted as planned.
I do recall suggesting the choices in the Quake game ought to be framed in the form of dynamic risk; what do we get in exchange for living in a quake zone?
Translated to the real world, most people, including your household and mine (and those of everyone I knew personally), survived the bet we made every day we didn't pack up and move away from the fault. My question is whether the 42 people who were mashed flat in their cars knew the risks and entered into the wager willingly. Be well, Petréa.
I'm glad you took my meaning. I also wonder if people are better at assessing risk in certain types of situations, i.e. player-vs-player as opposed to player-vs-"nature".
Also, I would expect better strategies to evolve after several iterations -- the article wasn't clear whether students got to play the game more than once. I suppose the counter-argument is that life is a game we only get to play once.
Schneier.com is a personal website. Opinions expressed are not necessarily those of BT.