I'll make a note to send you a pm at the next beta stage. 
Count me in, too!
Topic: Vickrey system in exchanges? (Read 1291 times)
Count me in, too!
I'll make a note to send you a pm at the next beta stage.
Sounds great.. I might try it.
I see. Sounds like a pretty good system, but the advantage of the "second-highest" system is that in a normal system, the buyers have to worry about what other people are bidding, so they don't bid too high. In a Vickrey system, they are never charged more than they had to pay to win.
That's very interesting. Is the final price the high bid, or the second highest?
I found out there's quite a bit of academic work on auction theory. A lot of it is out of my league, but it's convinced me that I'm on the right track.
I found out there's quite a bit of academic work on auction theory. A lot of it is out of my league, but it's convinced me that I'm on the right track.
Highest bid, but its never more than what the buyers were willing to pay for.
That's very interesting. Is the final price the high bid, or the second highest?
I found out there's quite a bit of academic work on auction theory. A lot of it is out of my league, but it's convinced me that I'm on the right track.
I found out there's quite a bit of academic work on auction theory. A lot of it is out of my league, but it's convinced me that I'm on the right track.
I can say that a version of that kind of auction works here in the Philippines, when selling wholesale amounts of fish after bringing them to port. After the fishing vessel docks and loads up tub after tub of fresh catch, their people at the market would start taking bids from prospective buyers. Called the "Whispering Auction", the brokers would take bids from the buyers by having them "whisper" their price per tub.
This generally helped with price discovery, raising the profit margins of the fishing vessels and prevent overpricing.
This generally helped with price discovery, raising the profit margins of the fishing vessels and prevent overpricing.
This question is related to an online game I'm designing. Items will be bought and sold in a bid-ask system inspired by bitcoin exchanges. I figured this would be the best place to ask this question, because there are people here with experience in using these sorts of systems.
To summarize, the system I'm talking about is one where users can either place a "bid" order to buy an item, or an "ask" order to sell it. If the highest bid is priced higher than the lowest ask, the transaction takes place, and the price is that of whichever order was placed first.
This works fine for a high-volume commodity like bitcoin, but in my game design the system will need to deal with low volumes, including unique items that may only be sold once. In addition, I want players to be able to buy and sell "derivatives" of a portfolio of commodities, so it is extra important that a good, automatic price discovery mechanism is used so that the portfolio can be sold for a fair price at a pre-designated time.
What I'm worried about is the problem of "auction sniping". I notice in many games and some real online auctions, the dominant strategy is to wait until the last minute, then bid a tiny amount more than the highest bid. This is bad for the seller in at least two ways:
1. Because the auction is decided entirely at the last minute, the seller's opportunity to attract bidders is severely limited. You can run the auction for a whole week, but good luck getting any bids until five minutes before close. Only people who happen to be online at that time will bid.
2. Buyers are competing on timing rather than price, meaning that the final price will often be below fair market value.
To that end, I'm considering modifying the exchange rules based on the concept of a Vickrey auction. The differences would be:
-All bids and asks are hidden.
-If you place a bid, and then an ask is placed that's below your bid, you pay whichever is higher: the price of the ask, or the price of the second-highest bid.
-If you place an ask, and then a bid is placed that's above your ask, you gain whichever is lower: the price of the bid, or the price of the second-highest ask.
The advantages of this system are:
-No possibility of "auction sniping".
-Even though bids/asks are hidden, there is no risk of bidding too high or asking too low. You never pay more than you needed to, you are never paid less than you could have been.
The disadvantage is that bids and asks would be hidden, which may be offputting to some people.
What do you think? Would this work?
(A complication I just thought of: If I bid $15 for 5 units, Bob bids $10 for 1 unit, and Charlie bids $5 for 4 units, do I pay $10 each for all five, or do I pay $10 the first one, and $5 for the next four? I'll have to think about that. EDIT: Recording this for my own reference, since I doubt anyone will read this. I think I've concluded that it is most fair to all parties that the $15 bidder pays $10 for all five units. The other alternatives I've considered would leave him paying less than $10, which would be confusing and frustrating to Bob.)
To summarize, the system I'm talking about is one where users can either place a "bid" order to buy an item, or an "ask" order to sell it. If the highest bid is priced higher than the lowest ask, the transaction takes place, and the price is that of whichever order was placed first.
This works fine for a high-volume commodity like bitcoin, but in my game design the system will need to deal with low volumes, including unique items that may only be sold once. In addition, I want players to be able to buy and sell "derivatives" of a portfolio of commodities, so it is extra important that a good, automatic price discovery mechanism is used so that the portfolio can be sold for a fair price at a pre-designated time.
What I'm worried about is the problem of "auction sniping". I notice in many games and some real online auctions, the dominant strategy is to wait until the last minute, then bid a tiny amount more than the highest bid. This is bad for the seller in at least two ways:
1. Because the auction is decided entirely at the last minute, the seller's opportunity to attract bidders is severely limited. You can run the auction for a whole week, but good luck getting any bids until five minutes before close. Only people who happen to be online at that time will bid.
2. Buyers are competing on timing rather than price, meaning that the final price will often be below fair market value.
To that end, I'm considering modifying the exchange rules based on the concept of a Vickrey auction. The differences would be:
-All bids and asks are hidden.
-If you place a bid, and then an ask is placed that's below your bid, you pay whichever is higher: the price of the ask, or the price of the second-highest bid.
-If you place an ask, and then a bid is placed that's above your ask, you gain whichever is lower: the price of the bid, or the price of the second-highest ask.
The advantages of this system are:
-No possibility of "auction sniping".
-Even though bids/asks are hidden, there is no risk of bidding too high or asking too low. You never pay more than you needed to, you are never paid less than you could have been.
The disadvantage is that bids and asks would be hidden, which may be offputting to some people.
What do you think? Would this work?
(A complication I just thought of: If I bid $15 for 5 units, Bob bids $10 for 1 unit, and Charlie bids $5 for 4 units, do I pay $10 each for all five, or do I pay $10 the first one, and $5 for the next four? I'll have to think about that. EDIT: Recording this for my own reference, since I doubt anyone will read this. I think I've concluded that it is most fair to all parties that the $15 bidder pays $10 for all five units. The other alternatives I've considered would leave him paying less than $10, which would be confusing and frustrating to Bob.)
Jump to: