Suppose you are reading your favorite blogs, when the idea strikes you, “Okay, I need to do X, but I can’t do it without an incentive. I shall order chicken wings, which are delicious, upon X’s completion.”
Dozens of minutes later, X is finished! But wait! You fell victim to the planning fallacy! Everywhere in the city that delivers chicken wings is closed now because X took longer than you thought it would.
In this case, it would be fairly senseless to wait until the next day to order the wings, as by then the reward would be completely disconnected from the action. Driving 35 minutes to get them would also be pretty senseless. I don’t know about driving 15 minutes.
This seems like a fairly difficult problem, but also one that simply will not occur very often, especially if you make your incentive something that’s unlikely to be difficult to obtain by the time you finish X.
That’s how I interpreted it as well, but I’m not sure the OP is distinguishing the signalling purpose of pre-commitment strategies from mechanisms of pre-commitment..
Reputations of pre-commitment are about signalling credible consequences in circumstances of asymmetric information. When bargaining with oneself, information is about as symmetric as it can get. It’s not like you mistrust your future self’s willingness to go through with getting chicken wings. Any obstacle to getting them is transparent to all parties (you), and shouldn’t impact your future expectation of being able to reward yourself unless you’re staggeringly incompetent at obtaining chicken wings. If that’s the case, you’ll probably factor this in when planning your incentive.
Mechanisms of pre-commitment are a more salient tool when bargaining with oneself over time (cf. dynamic inconsistency ), but only when your goals are inconsistent over time. Post-X you presumably wants chicken wings as much as pre-X you, but is more informed about the cost of obtaining them. There is presumably a level of expense pre-X you would sensibly commit to for the specified reward. If some sort of catastrophe occurred as soon as you’d finished X, pre-X you wouldn’t expect post-X you to crawl through the dust with your one remaining limb muttering “must...get...chicken...wings...”
The issue seems to boil down to “are you staggeringly incompetent at rewarding yourself? If not, don’t worry.”
Are you entering into a sub function of the original x/y assessment here? As in if X is done, Y, but Y is a function in itself of assessing the optimal reward for X?
If it’s still important to add a reward of Y (in addition to the personal value of having completed X), you probably need to substitute with something novel and maintain the understanding that it is a reward for X (even if not the originally scoped one).
Can you provide some specific examples?
Let me make one.
Suppose you are reading your favorite blogs, when the idea strikes you, “Okay, I need to do X, but I can’t do it without an incentive. I shall order chicken wings, which are delicious, upon X’s completion.”
Dozens of minutes later, X is finished! But wait! You fell victim to the planning fallacy! Everywhere in the city that delivers chicken wings is closed now because X took longer than you thought it would.
In this case, it would be fairly senseless to wait until the next day to order the wings, as by then the reward would be completely disconnected from the action. Driving 35 minutes to get them would also be pretty senseless. I don’t know about driving 15 minutes.
This seems like a fairly difficult problem, but also one that simply will not occur very often, especially if you make your incentive something that’s unlikely to be difficult to obtain by the time you finish X.
That’s how I interpreted it as well, but I’m not sure the OP is distinguishing the signalling purpose of pre-commitment strategies from mechanisms of pre-commitment..
Reputations of pre-commitment are about signalling credible consequences in circumstances of asymmetric information. When bargaining with oneself, information is about as symmetric as it can get. It’s not like you mistrust your future self’s willingness to go through with getting chicken wings. Any obstacle to getting them is transparent to all parties (you), and shouldn’t impact your future expectation of being able to reward yourself unless you’re staggeringly incompetent at obtaining chicken wings. If that’s the case, you’ll probably factor this in when planning your incentive.
Mechanisms of pre-commitment are a more salient tool when bargaining with oneself over time (cf. dynamic inconsistency ), but only when your goals are inconsistent over time. Post-X you presumably wants chicken wings as much as pre-X you, but is more informed about the cost of obtaining them. There is presumably a level of expense pre-X you would sensibly commit to for the specified reward. If some sort of catastrophe occurred as soon as you’d finished X, pre-X you wouldn’t expect post-X you to crawl through the dust with your one remaining limb muttering “must...get...chicken...wings...”
The issue seems to boil down to “are you staggeringly incompetent at rewarding yourself? If not, don’t worry.”
Are you entering into a sub function of the original x/y assessment here? As in if X is done, Y, but Y is a function in itself of assessing the optimal reward for X?
If it’s still important to add a reward of Y (in addition to the personal value of having completed X), you probably need to substitute with something novel and maintain the understanding that it is a reward for X (even if not the originally scoped one).