3 ms·

I thought a Vicrey-Clarke-Groves auction could be undermined by bidder collusion and in particular in some circumstances by a single bidder making multiple bids

I thought a Vicrey-Clarke-Groves auction could be undermined by bidder collusion and in particular in some circumstances by a single bidder making multiple bids under different names. Isn’t that exactly what scalping groups would do?

qsort 6y ago
It can be undermined by collusion, but not if a sufficiently large portion of the bidders is actually honest like, presumably, is the case here. I'm not sure how a single bidder making multiple bids under different names would impact the math, but intuitively I'd say under the same assumption as above the system should hold: if I make N bets out of M with N << M, the effect should be small enough that the guarantees still (broadly) apply. I could be wrong here, feel free to correct me if you know better.
- paulcole 6y ago
  > if a sufficiently large portion of the bidders is actually honest like, presumably, is the case here. I’m not sure why you think this is the case here? If you create a system that can be gamed by a large portion of the bidders being dishonest, a large portion of the bidders will become dishonest.
  - qsort 6y ago
    I suppose, but how realistic that really is on a scale this large? Wouldn't it be just a giant prisoner's dilemma? Unless almost all parties collude, it's still optimal to bid truthfully.
    - paulcole 6y ago
      If a system for selling tickets has a weak point, it will be found and exploited by scalpers. That's kind of the whole point of the original article. Every "solution" thrown out by well-meaning third parties is either exploitable by scalpers, makes the experience significantly worse for legitimate buyers, or just shifts which legitimate buyers are able to make a purchase (i.e. I didn't get to buy a ticket but you did. This solution would let me buy a ticket but wouldn't let you buy one).

3 ms·

I thought a Vicrey-Clarke-Groves auction could be undermined by bidder collusion and in particular in some circumstances by a single bidder making multiple bids

by paulcole 6y ago

qsort 6y ago
It can be undermined by collusion, but not if a sufficiently large portion of the bidders is actually honest like, presumably, is the case here. I'm not sure how a single bidder making multiple bids under different names would impact the math, but intuitively I'd say under the same assumption as above the system should hold: if I make N bets out of M with N << M, the effect should be small enough that the guarantees still (broadly) apply. I could be wrong here, feel free to correct me if you know better.
- paulcole 6y ago
  > if a sufficiently large portion of the bidders is actually honest like, presumably, is the case here. I’m not sure why you think this is the case here? If you create a system that can be gamed by a large portion of the bidders being dishonest, a large portion of the bidders will become dishonest.
  - qsort 6y ago
    I suppose, but how realistic that really is on a scale this large? Wouldn't it be just a giant prisoner's dilemma? Unless almost all parties collude, it's still optimal to bid truthfully.
    - paulcole 6y ago
      If a system for selling tickets has a weak point, it will be found and exploited by scalpers. That's kind of the whole point of the original article. Every "solution" thrown out by well-meaning third parties is either exploitable by scalpers, makes the experience significantly worse for legitimate buyers, or just shifts which legitimate buyers are able to make a purchase (i.e. I didn't get to buy a ticket but you did. This solution would let me buy a ticket but wouldn't let you buy one).