Then it makes no sense to me to pay a third person to find out how much money A has. Finding out that there is more money available doesn't actually get B more money, so paying a third party subtracts from whatever money B would be receiving.
Also, am I missing something? I don't see any risk associated with accepting the money, regardless of what the offer is or how much A received...
It's a known irrationality among players of the game.
To a cold logical mind, any offer should be accepted, because getting some money is better than getting none. Even a payout of $1 is better than what you get from rejecting.
When the experiment is done in the wild, however, most people Reject when the offer is under 20-25% of the pot. I'm haven't read up on the explanation for this recently, but I seem to recall it has to do with a brain mechanism that makes people try to punish others for their lack of altruism.
... What incentive does A even have to give B money? Just "you're my friend, here's 15 bucks?" If B didn't contribute to A getting money, A has no obligation to giving B ANY money, and if A takes B out to a $10 lunch, B should do the smart thing and be happy about his free lunch, because if he rejects that $10, he pays for his own lunch.
Maybe I'm just coldly logical, and also appreciative of even small acts of kindness.
Oooooh I missed that line in the original description of the game. I thought it was "A gets money as long as A at least offers B money, regardless of whether B accepts it." That was dumb of me
( ... )
If A is a rational actor and aware of how this game is normally played, the existence of the third option should encourage A to offer a larger share.
If B is a strictly rational actor whose only decision scope is profit from *this game* he/she should accept A's offer and forget the third option exists. The third option is "lose money to learn a piece of information that doesn't change a decision moments before I would have learned it anyway."
If B is a less rational, more normal actor who is concerned with "fairness" then I suspect they will pay a good chunk to make sure they're not getting "screwed out of their fair share" of $500.
I'd also posit that, in the less-rational-B case offering anything between $100 and $249, inclusive, is likely to go poorly, and that offering anything between $51 and $99 might go poorly too.
It would be an interesting game if A and B had to reach a consensus before A knew how much was in the envelope.
What if B's option C is "offer someone a percentage of their given take to find out the amount A received, then choose"?
It eliminates the "well, if I reject, now I'm out money" thing. Though it then brings into question whether or not the third party will be telling the truth.
I'd expect that in the modified option, $40-45 is A's best offer regardless of how much money you have received (maybe as low as $35). Higher, and B would assume you had the $500 and were being stingy, rather than the $100 and being generous.
So A is allowed to lie about receiving $100, but not about receiving $500? Then if A claims to have been given $500, B can be sure it is the truth. I can't see how this is useful to anyone though.
Also, if A claims to have been given $100 and I pay a third party and learn that it was a lie I have additional information that A is a liar (where if A couldn't make a claim about how much they were given I would be entirely uncertain about A's proclivity for lying).
Comments 16
Reply
Reply
Also, am I missing something? I don't see any risk associated with accepting the money, regardless of what the offer is or how much A received...
Reply
To a cold logical mind, any offer should be accepted, because getting some money is better than getting none. Even a payout of $1 is better than what you get from rejecting.
When the experiment is done in the wild, however, most people Reject when the offer is under 20-25% of the pot. I'm haven't read up on the explanation for this recently, but I seem to recall it has to do with a brain mechanism that makes people try to punish others for their lack of altruism.
Reply
Maybe I'm just coldly logical, and also appreciative of even small acts of kindness.
Reply
Reply
Reply
Reply
If B is a strictly rational actor whose only decision scope is profit from *this game* he/she should accept A's offer and forget the third option exists. The third option is "lose money to learn a piece of information that doesn't change a decision moments before I would have learned it anyway."
If B is a less rational, more normal actor who is concerned with "fairness" then I suspect they will pay a good chunk to make sure they're not getting "screwed out of their fair share" of $500.
I'd also posit that, in the less-rational-B case offering anything between $100 and $249, inclusive, is likely to go poorly, and that offering anything between $51 and $99 might go poorly too.
It would be an interesting game if A and B had to reach a consensus before A knew how much was in the envelope.
Reply
It eliminates the "well, if I reject, now I'm out money" thing. Though it then brings into question whether or not the third party will be telling the truth.
I'd expect that in the modified option, $40-45 is A's best offer regardless of how much money you have received (maybe as low as $35). Higher, and B would assume you had the $500 and were being stingy, rather than the $100 and being generous.
Reply
Reply
Reply
Leave a comment