I often listen to music or use my phone while walking through town. I have heard that this is dangerous and could cause me to get killed, so I’ll attempt a back of the envelope calculation to work out if I should stop.
Your post is very interesting. I do like your analysis and your beliefs are similar to mine. I would add the following cautions based on my recollections of US Pedestrian Traffic Fatalities data:
1. Don't walk Drunk. One third of US pedestrian fatalities occur with intoxicated pedestrians.
2. Don't J-walk. One half of US pedestrian deaths occur in roadways where pedestrians are not designated to be. I am aware that in many of these cases it is because no pedestrian facilities are provided.
3. Be more careful on Arterial roads. I am less sure of this number but close to 40% of US pedestrian fatalities occur on arterial streets. I would add that crossing a street also requires additonal vigulance.
4. Be more careful after dark. Pedestrians don't have visibility requirements so if it is dark there is a higher chance of fatalities per mile walked.
Based on these recollections, when in the daylight and on a sidewalk there is very little added risk with headphones and walking texting/emailing.
But as an former avid runner, I would not advise running in the street unless you have full control of your sense and take full responsbility for being alive after you enter the domain of the car.
Limit your device use to cases where you can retain situational awareness. That means stopping whatever you are doing on the phone whenever you see something that you should have noticed earlier (e.g. another pedestrian) but didn't, and whenever you anticipate that you might need to react quickly to something (e.g. when you cross any street).
For music/audio, I would restrict it to one ear.
If your goal in walking is to relax, or your goal for audio is to zone out or attend to the audio, don't do it on the sidewalk or anywhere else other people have a reasonable expectation that you are situationally aware.
You may have akrasia but the preference to act as if you didn't have akrasia.A strong time preference may simply be based on impulsiveness, while you would prefer to be less impulsive.
Yes. So you can try to compensate for them into account by focusing on egosyntonic emotional responses.
Risk aversion may simply be based on how thinking about risk makes you feel, while the full scope of potential gains are ignored.
Risk aversion is, by definition, how thinking about risk makes you feel. What is your point?
I find it absurd that, if someone decides on reflection that he would rather correct for these, someone else declares he shouldn't.
If you are trying to build an explicit model of your preferences and you miss important features then you are doing it wrong. Using that model for decision making will make you less happy on average than you would have been if you used a better model.
If you contribute to x-risk reduction during your life then the magnitude of negative side of this calculation increases substantially while the positive side does not.
If the numbers are largely made up, it really is a fallacy.But if they reflect best guesses on reality, it is useful for reasoning under uncertainty.
The account tacitly on offer (by said "rationalists") is that our "best guesses on reality" (on some things) are "largely made up."
Their real fallacy lies in their failure to recognize that when information is sufficiently unreliable, it is worse than nothing: it becomes an impediment to global intuition and sense of proportion.
Try to estimate probabilities. Try to get an intuitive grasp of these probabilities (this is difficult with very small or very large probabilities, but you could cultivate the habit by thinking of various events that have similar probabilities).
Estimate expected utilities of your choices from your emotional response to imagined outcomes (once you have an intuitive grasp of the probabilities, your emotional response should take them into account). When in doubt, err on the safe side.
I considered the sarcasm appropriate because the topic is evidently so trivial. For example, would you consider it appropriate for Overcoming Bias to run a similar column weighing (with math!) the relative risks vs rewards of (among others) 1: tying one's shoes prior to walking or simply leaving them untied to save time; (2) zipping one's fly upon leaving the bathroom vs saving time by leaving it unzipped; (3) putting on a seatbelt when driving vs not bothering, so as to save time; (4) washing one's hands before eating vs not-washing, and then working out the odds of catching a hand-carried disease vs the time and water consumed in washing, etc. Oh heck, ya know, come to think of it, I guess that I must admit that I actually would NOT be surprised to see columns on some of those actual subjects at this blog! So... never mind. Sigh.
Your advice is "watch where you are going". Instituting a policy of not using one's phone is a practical, actionable way towards implementing your advice. So why the sarcasm?
Ok, I was wrong about the enhanced risk of death from driving drunk: I can't find proper stats, but Steven Levitt says (http://www.freakonomics.com... it's 13x, and the risk of walking drunk is 8x that of walking sober.
Which would make your estimate of 3x within an order of magnitude, though still a substantial underestimate.
The point about relative risk from drunk driving and phone walking stands, though, I think.
Your post is very interesting. I do like your analysis and your beliefs are similar to mine. I would add the following cautions based on my recollections of US Pedestrian Traffic Fatalities data:
1. Don't walk Drunk. One third of US pedestrian fatalities occur with intoxicated pedestrians.
2. Don't J-walk. One half of US pedestrian deaths occur in roadways where pedestrians are not designated to be. I am aware that in many of these cases it is because no pedestrian facilities are provided.
3. Be more careful on Arterial roads. I am less sure of this number but close to 40% of US pedestrian fatalities occur on arterial streets. I would add that crossing a street also requires additonal vigulance.
4. Be more careful after dark. Pedestrians don't have visibility requirements so if it is dark there is a higher chance of fatalities per mile walked.
Based on these recollections, when in the daylight and on a sidewalk there is very little added risk with headphones and walking texting/emailing.
But as an former avid runner, I would not advise running in the street unless you have full control of your sense and take full responsbility for being alive after you enter the domain of the car.
Limit your device use to cases where you can retain situational awareness. That means stopping whatever you are doing on the phone whenever you see something that you should have noticed earlier (e.g. another pedestrian) but didn't, and whenever you anticipate that you might need to react quickly to something (e.g. when you cross any street).
For music/audio, I would restrict it to one ear.
If your goal in walking is to relax, or your goal for audio is to zone out or attend to the audio, don't do it on the sidewalk or anywhere else other people have a reasonable expectation that you are situationally aware.
VV:
"Risk aversion is, by definition, how thinking about risk makes you feel. What is your point?"
The point is exactly what I wrote in the very sentece you quoted:
Focussing emotionally on the risk, you may ignore the full scope of the potential benefit if the risk doesn't happen.
"If you are trying to build an explicit model of your preferences and you miss important features then you are doing it wrong."
Exactly.
This is why you have to correct for risk aversion and time preference if, on reflection, you find you would rather go with expected value.
I see no need to discuss this further.
> So I can convert 20 minutes of my time today into 60 minutes of extra time alive?
Just don't ask about the marginal return on the additional units of exercise...
You may have akrasia but the preference to act as if you didn't have akrasia.A strong time preference may simply be based on impulsiveness, while you would prefer to be less impulsive.
Yes. So you can try to compensate for them into account by focusing on egosyntonic emotional responses.
Risk aversion may simply be based on how thinking about risk makes you feel, while the full scope of potential gains are ignored.
Risk aversion is, by definition, how thinking about risk makes you feel. What is your point?
I find it absurd that, if someone decides on reflection that he would rather correct for these, someone else declares he shouldn't.
If you are trying to build an explicit model of your preferences and you miss important features then you are doing it wrong. Using that model for decision making will make you less happy on average than you would have been if you used a better model.
You may have akrasia but the preference to act as if you didn't have akrasia.
A strong time preference may simply be based on impulsiveness, while you would prefer to be less impulsive.
Risk aversion may simply be based on how thinking about risk makes you feel, while the full scope of potential gains are ignored.
I find it absurd that, if someone decides on reflection that he would rather correct for these, someone else declares he shouldn't.
Indeed, that's the fallacy of fake precision: http://www.fallacyfiles.org...
If you contribute to x-risk reduction during your life then the magnitude of negative side of this calculation increases substantially while the positive side does not.
If the numbers are largely made up, it really is a fallacy.But if they reflect best guesses on reality, it is useful for reasoning under uncertainty.
The account tacitly on offer (by said "rationalists") is that our "best guesses on reality" (on some things) are "largely made up."
Their real fallacy lies in their failure to recognize that when information is sufficiently unreliable, it is worse than nothing: it becomes an impediment to global intuition and sense of proportion.
Try to estimate probabilities. Try to get an intuitive grasp of these probabilities (this is difficult with very small or very large probabilities, but you could cultivate the habit by thinking of various events that have similar probabilities).
Estimate expected utilities of your choices from your emotional response to imagined outcomes (once you have an intuitive grasp of the probabilities, your emotional response should take them into account). When in doubt, err on the safe side.
If the numbers are largely made up, it really is a fallacy.
So they seem.
Discounting and risk aversion aren't obligations or norms, just facts about most people's actual choices.
If you try to model your preferences explicitly and you omit these while they exist in your preference system then you are making a mistake.
"Exercise for 20 minutes, and you gain two units of microlife."
So I can convert 20 minutes of my time today into 60 minutes of extra time alive? I think I've discovered the secret to immortality!
I considered the sarcasm appropriate because the topic is evidently so trivial. For example, would you consider it appropriate for Overcoming Bias to run a similar column weighing (with math!) the relative risks vs rewards of (among others) 1: tying one's shoes prior to walking or simply leaving them untied to save time; (2) zipping one's fly upon leaving the bathroom vs saving time by leaving it unzipped; (3) putting on a seatbelt when driving vs not bothering, so as to save time; (4) washing one's hands before eating vs not-washing, and then working out the odds of catching a hand-carried disease vs the time and water consumed in washing, etc. Oh heck, ya know, come to think of it, I guess that I must admit that I actually would NOT be surprised to see columns on some of those actual subjects at this blog! So... never mind. Sigh.
If the numbers are largely made up, it really is a fallacy.
But if they reflect best guesses on reality, it is useful for reasoning under uncertainty.
Discounting and risk aversion aren't obligations or norms, just facts about most people's actual choices.
Your advice is "watch where you are going". Instituting a policy of not using one's phone is a practical, actionable way towards implementing your advice. So why the sarcasm?
Ok, I was wrong about the enhanced risk of death from driving drunk: I can't find proper stats, but Steven Levitt says (http://www.freakonomics.com... it's 13x, and the risk of walking drunk is 8x that of walking sober.
Which would make your estimate of 3x within an order of magnitude, though still a substantial underestimate.
The point about relative risk from drunk driving and phone walking stands, though, I think.