Quote: (Originally Posted by Underwaterbear)
Okay, quick retort...
Yeah I agree, as far as a classical interpretation goes if for example you're in the insurance business or the builder of CCRs where there are many trials. But the means by which the probability of an event is calculated make no sense when an outcome is binary (as it is when there is only one trial).
Sure it does.
We are all going to die of SOMETHING. For example, you might be hit by lightning. Do you never go outdoors, thereby reducing your risk? Probably not. Why? Because the risk of the bad thing happening (Zzzzaaaap!) is small in the first place, so reducing it further by staying inside has low marginal utility.
One must
also consider the marginal utility of the contemplated activity. That utility value varies from person-to-person and event-to-event.
Many people judge the risk of a bad outcome from smoking to be acceptable to them - even though the actual risk is nearly 50% over a period of a few decades! Those aren't real good odds! But if you call each cigarette a "trial", then it doesn't look as bad.
Look at the person who chooses to engage in IV drug use. The risk (of death by HIV, if you don't OD first) is quite high - perhaps 50%? Yet the
perceived benefit is high enough that people do it.
Or, for that matter, commercial fishing. The risk of death is quite high, yet the reward (a paycheck) is quite material too. Some people consider the balance acceptable, others not.
When it comes to recreational activities nobody other than the participant can accurately calibrate the utility value, since there is no objective measurement possible.
Quote:
Whilst it is strictly correct that expectation theory (risk x consequence or any other formulation) can be used as an index or measure of risk I believe it leads to poor decisions because it understates the weight of the consequence. This is part of why we percieve risk so poorly (my point 1).
The two are completely decoupled in any honest assessment, and the utility value must also be considered; there is actually a triplet of data to weigh.
Since the penalty for a bad trial is typically death in this endeavor, you can call the consequence "infinite". However, this does not necessarily mean that the risk is unacceptable, because that consequence will visit
all of us some day and the personal utility value from diving may be
extremely high.
Therefore, the real question is
what percentage of trials will lead to the bad consequence? For virtually everyone, if the risk approaches that of winning the lotto, people will find it acceptable.
Some will find the odds of a bad trial at 1:1000 acceptable, because they find the utility value extremely high. Others will calibrate the number of acceptable risk to utility at 1:100,000, and still others will want the 1:17,000,000 or so of the lotto....
Quote:
My contention is that knowing the rate of occurence doesn't matter in a trial of one. Something either happens or it does not. If the consequence is undesirable, being responsible is about eliminating the event as far as possible or placing contigencies in place to mitigate the consequence.
By definition a risk scenario in which the consequence is death has an absolute "bad consequence".
This position would lead to a call for an absolute ban on diving.... and yet, we're not debating that... I don't think......