At 17:35, you say: "Either at random or some a priori fixed rule, *it's not going to matter*". In case the yes/no-announcements are public and sequential, it seems like it would matter. Say two bidders Bugs and Daffy value a gold nugget at 5 and 3 respectively, with epsilon 1. Further, assume Bugs always goes first. Daffy has this strategy: If Bugs decides to bid past 3, Daffy will bid as if sincerely valuing at infinity. If Bugs stops at 3 (says no at 4), Daffy stops at 3 as well. Clearly if the tie break is a coinflip, Bugs is better off stopping at 3, where he gets an expected utilty of (5-3)*0.5 = 1, and 0 if he bids sincerely (or anything else).
Thanks sir for all these great lectures you provide!
always enjoying the way you teach.
Thanks, I really enjoyed CS364A, hope this module is even more interesting.
1:00
At 17:35, you say: "Either at random or some a priori fixed rule, *it's not going to matter*".
In case the yes/no-announcements are public and sequential, it seems like it would matter.
Say two bidders Bugs and Daffy value a gold nugget at 5 and 3 respectively, with epsilon 1. Further, assume Bugs always goes first.
Daffy has this strategy: If Bugs decides to bid past 3, Daffy will bid as if sincerely valuing at infinity. If Bugs stops at 3 (says no at 4), Daffy stops at 3 as well.
Clearly if the tie break is a coinflip, Bugs is better off stopping at 3, where he gets an expected utilty of (5-3)*0.5 = 1, and 0 if he bids sincerely (or anything else).