The problem is that this cannot be made rigourous without making it into CDT.
I will use the smoking example because it is essential that you do not know which set you are in, which does not hold in the driving example. EDT tells you to calculate the probability that you will die if you do smoke and the probability that you will die if you do not smoke. These numbers can be compared. It’s true that any evidence as to which gene you have would reverse this decision, but this is not part of any general process. We could adopt a general rule of dividing up the possibilities in this way with respect to every variable. However, one such variable is what choice you will make. Just as you can make set(people with gene) and set(people without gene), you could make set(people with gene who chose to smoke), set(people with gene who chose not to smoke), set(people without gene who chose to smoke), and set(people without gene who chose not to smoke). At this point, we have recovered CDT.
That implies there’s no actual difference between the theories other than the level of detail at which you analyze the evidence you have. I admit it’s entirely possible that EDT says you can’t look beyond the very most basic level of information, but that makes it sound like a straw man of a theory; any theory that obliges you to ignore evidence known to be relevant is hardly deserving of the name theory.
As I didn’t quite get to, that’s not my only objection. I need to finish writing this thing.
The problem is that this cannot be made rigourous without making it into CDT.
I will use the smoking example because it is essential that you do not know which set you are in, which does not hold in the driving example. EDT tells you to calculate the probability that you will die if you do smoke and the probability that you will die if you do not smoke. These numbers can be compared. It’s true that any evidence as to which gene you have would reverse this decision, but this is not part of any general process. We could adopt a general rule of dividing up the possibilities in this way with respect to every variable. However, one such variable is what choice you will make. Just as you can make set(people with gene) and set(people without gene), you could make set(people with gene who chose to smoke), set(people with gene who chose not to smoke), set(people without gene who chose to smoke), and set(people without gene who chose not to smoke). At this point, we have recovered CDT.
That implies there’s no actual difference between the theories other than the level of detail at which you analyze the evidence you have. I admit it’s entirely possible that EDT says you can’t look beyond the very most basic level of information, but that makes it sound like a straw man of a theory; any theory that obliges you to ignore evidence known to be relevant is hardly deserving of the name theory.
As I didn’t quite get to, that’s not my only objection. I need to finish writing this thing.