Unseeded Auction Thread

[Nagalim]

Am I completely off the track now?

You’re like a train coming in to station. 90% of the auction software is simply analyzing the block chain and finding that distribution list. 5% is the participant software (including price feeds), 1% is updating the website automatically, and 4% is registration (not yet implemented).

Multisig is totally viable, though it would be beyond my skill range. The auction transaction (I.e. derivation of the distribution list) is 100% verifiable using the block chain if we have people register their nsr-nbt link by signing some kind of ledger with their private keys (this does not need to happen on the block chain as there is no concept of order or time).

What I’m getting at is that we could take this beyond multisig and write it into the core protocol. Instead of sending funds to an auction address a participant would do a special burn event. At auction close, all funds on the ‘distribution list’ are created fresh on the block chain at the linked address. At that point the seeding is native (the entire auction becomes a custodial grant event) and must be controlled via voting on the blockchain (which is only 1 additional bit per block).

[masterOfDisaster]

Hoot hoot!

Sorry - couldn’t resist. Sometimes I play up

[cryptog]

Any update?

[Nagalim]

Nope. I’m working on logging, but haven’t done anything in a while. My work on the ALP software is really really helping me understand how to code this better.

[willy]

@masterOfDisaster mentioned this so often in the last week, that I’m now very keen on seeing a new auction.

Well done mOD

[zoro]

he knows for sure how to promote things!
i am also waiting…

[masterOfDisaster]

I’m glad that I could promote this idea, because I consider it the missing link between NBT and NSR.
At the moment there are only workarounds available to bring NBT to the market or remove them, typically manually held auctions that lack continuity and transparency and without rules that enforce such an auction - they are done, when Nu feels like it should…

Seeded auctions could pave the ground for a tier 6 liquidity management that is embedded in a framework with defined guidelines.

Guidelines like:

• tier 4 multi signature custodians seed granted NBT in auctions if defined conditions are met (like network wide sell side at below x% of buy side (at average) in the last y days) and distribute the z% of the proceeds as dividends (after the obtained NSR have been traded for PPC)
• tier 4 multi signature custodians seed granted NSR in auctions if defined conditions are met (like network wide buy side at below x% of sell side (at average) in the last y days) and burn the received NBT

The interval of seeded auctions can be adjusted to enable Nu to react more quickly to changed market conditions at the cost of additional administrative overhead.

Tier 6 can then used to refill tier 4 - same story: framework, conditions, rules, etc.

You might notice that I omitted tier 5 - shame on me

[Nagalim]

Working on a recycle function to return unregistered funds between auction closes. That way participants can know for sure what the current auction balance is. It’s important for the gaming aspects of it.

I just wish I wasn’t so busy all the time. The list goes: recycle, set up bigger nuauction server, alpha testing. I’ll skip the website upgrade for now.

[Nagalim]

The recycle function checks txns with 10-90 confirmations in between auction blocks for any txns that have inputs that are not registered. If there is an unregistered send address the auction server will send the full output back to the first unregistered address it finds in that txn. I’m doing my best to make sure this cannot be exploited (ddos is always a concern).

The end result should be that any output at an auction address with more than 20 confirmations or so should be reliable as a submission to the auction. That means anyone can verify at any point in time what the auction price was about 30 minutes ago. This of course creates an unavoidable haze about the very end of the auction that can only be verified in hindsight, but therein lies the gameablility of the auction.

To help others grasp the gravity of this mechanic, let me give an example:
Participant A, Participant B, and Participant C all have registered address pairs. Participant A submits 10 NSR and 1 NBT at the beginning of the auction, which coincides with the market price of $0.1/NSR. Participant B goes fishing and submits 10 NSR, skewing the auction price. Participant C submits 1 NBT to balance the price again. At the end of the auction, during the last few blocks during which no transaction can be fully confirmed as being registered or not, Participant B submits another 1 NBT. Now the auction results are:
Auction price = 0.15 NBT/NSR
Participant A sells NSR for above market price
Participant B sells NSR for above market price
Participant C buys NSR for above market price

As you can see, Participant C gets the shaft for having been fooled by Participant B. However, if Participant B hadn’t’ve submitted at the last minute, Participant C would have been more than happy. The gaming done by Participant B and C here come from a very fundamental maneuver that C was making: Participant C submitted at a price not aligned with the market price and hoped the participation would be enough to make up for it. If Participant C had instead ignored the price and simply looked at the current auction volume, he would see 20 NSR and 1 NBT. By submitting 1 NBT, Participant C was drastically changing the amount of NBT in the auction. Had he instead submitted 0.1 NBT or 0.01 NBT, the price would barely have moved and he would be just fine. Participant A here is extremely vital and will be played by stabilizers in unseeded auctions and Nu/custodians in seeded auctions.

To Reiterate:
The benefit of the recycle function is not to know what the auction price is at any given moment, but instead to know what the auction volume is for either currency (NBT or NSR). The price can be changed at the last moment but the volume cannot be decreased if the submission was from a registered address (i.e. was not recycled).

Recycling also helps avoid participant error, something @masterOfDisaster experienced during one of the last auction tests.

[Nagalim]

Can anyone tell me if “listunspent” could ever return an output that I just spent via a “sendmany” that may still be in the mempool? I want to say the answer is no, because if I just spent it it shouldn’t return in the unspent list, but I’m not sure. I was hoping maybe someone might just know the answer.

Edit: lol, I totally have to do the recycling by forging a raw transaction, otherwise it won’t recycle the correct outputs. W/e, not that big of a difference between a manysend and forging a raw transaction anyway.

[cryptog]

I would be interested in having @JordanLee 's opinion on Unseeded auction.

[zoro]

why? this is a community driven operation, not a technical aspect.
We shouldn’t rely solely on an “authority” for everything we do in NU.
No offence my friend

[cryptog]

I hear you.

Instead of community-driven I would say shareholders-driven. I am here also for making business.

[Nagalim]

JL has responded in the past saying he doesn’t understand. When I’m up and running again I’m hoping the community will start to understand and facilitate things such the the idea propagates up. Grassroots and so on.

[zoro]

if JL dosen’t understand, then the rest of us are hopeless
anyway, lets try it and perhaps we get the whole idea.

[mhps]

@Nagalim, can the unseeded auction be made into a game that educates participants how it works?

[Nagalim]

The biggest hurtle is that there is no ‘sendfromaddress’ rpc. If we fix that, seeded auctions can definately be a heavy nbt use case, complete with gaming and pegging and shareholder leverage and tons of stuff. It’s like an open market with added mystery and intrigue. (Also note that it locks up funds like TheDAO locks up eth)

[mhps]

The biggest hurtle is that there is no ‘sendfromaddress’ rpc.

can you elaborate or point to where it was discussed?

I am thinking splitting the “bet” nsr and nbt from participants according to the auction price from @jooize’s blind auction, instead of the average price, and the extra nbt or nsr will be given to the “winners” who get closest to the blind auction price…

[Nagalim]

Very interesting. So you force the blind auction price to be the unseeded auction price. Im trying to think of the implications if there is a higher volume in the dual sided auction than the blind auction. Very interesting idea.

sendfromaddress

[mhps]

the blind auction

if the the blind auction nsr/nbt price is higher than the average price of the unseeded auction, then there will be nsr left after the unseeded auction payback. They can be given as reward according to how close a bidder bidded to the bllind auction price as a incentive

reward = weight_i * sum_of_i(weight_i ) * total_award

, where weight_i is for the ith bidder, and strongly decreases as the bidding price is different than the target – the blind auction price.

weight_i = 1 / ( bid_price_i - blind_auction_price )^2

It’s also possible that Nu seed the auction with NSR that to be auctioned in the blind auction.

One thing to think about is that Nu wants the blind auction to compete to result in high prices

This game can be extended to bidding on other things.

I wonder if parking can be used in the game, too.