elective-stereophonic
elective-stereophonic
[NXT 2.0] A framework for thinking about incentives
Please login or register.

Login with username, password and session length
Advanced search  

News:

Latest Stable Nxt Client: Nxt 1.11.15 | Latest Experimental Nxt Client: Nxt 1.12.0e

Pages: [1] 2  All

Author Topic: [NXT 2.0] A framework for thinking about incentives  (Read 4143 times)

Damelon

  • Hero Member
  • *****
  • Karma: +792/-54
  • Offline Offline
  • Posts: 2314
    • View Profile
    • Nxt Inside
[NXT 2.0] A framework for thinking about incentives
« on: February 15, 2016, 01:27:28 pm »

Dave and I went to an EXCELLENT presentation about incentive structures in Bitcoin by George Papageorgiou last week.

He outlined the different groups that have a stake in the system from an academic point of view and we were very happy to be there.

To me, this presentation presents a framework to think and talk about how a system emerges and how it will evolve. WE have that luxury; Bitcoin, being the first, didn't.
I suggest we use the knowledge in our discussions about 2.0, as we seem to lack a shared language and frame of thought.

Regardless of this, it's a fascinating presentation and once the video is uploaded, we will of course share it.
In the meantime, please read it and mull it over. :)

http://www.bitcoinwednesday.com/wp-content/uploads/2016/02/Bitcoins-Incentive-Structure.pdf
Logged
Member of the Nxt Foundation | Donations: NXT-D6K7-MLY6-98FM-FLL5T
Join Nxt Slack! https://nxtchat.herokuapp.com/
Founder of Blockchain Workspace | Personal Site & Blog

LocoMB

  • Hero Member
  • *****
  • Karma: +101/-37
  • Offline Offline
  • Posts: 751
    • View Profile
Re: [NXT 2.0] A framework for thinking about incentives
« Reply #1 on: February 15, 2016, 03:59:37 pm »

Definitely!
Maybe this is the right place to ask:

If fNXT is the token that is used for forging (i.e. is the reqard token that forgers receive), and it is not be used for anything else, because anything else is what NXT is supposed to be used for:

How then is the route for forgers to monetize their fNXT?
Surely, there must be such a route, because forgers have to pay their bills somehow, and a reward that is not exchangeble is not a reward at all, and esp. can the forgers not pay their bills with such a non-exchangable thing!

This seems to be quite the bottleneck.

I am sorry for not reading hte whole intial thread- but maybe this can be compiled in a sort of FAQ re. this topic?
As I see it, there already is an fNXT, and it is called 'NXT'.
And BCNXTs original vision was exactly that: using NXT as a network participation reward token, and Monetar System for all else.

So why not: create an NXT MS coin, and use that henceforth as currency, and keep NXT as NXT to forge.

Ultimately, the objective seems to be to decouple the currently dual function of NXT as currency and NXT as forging reward.
I am fairly sure that this is possible without forced redistributoin and without re-evaluation.

Or better- if re-evaluation, then on even terms such that the relative stakes of all participants do not change.
Was this discussed already?




 
« Last Edit: February 15, 2016, 04:03:32 pm by LocoMB »
Logged
TOX
90E54E5B5213290EE616D425CADC473038CFABFA53C913271AA8559D1937DC4AF3A354A9E4E5

MrV777

  • Hero Member
  • *****
  • Karma: +115/-4
  • Offline Offline
  • Posts: 988
    • View Profile
Re: [NXT 2.0] A framework for thinking about incentives
« Reply #2 on: February 15, 2016, 04:05:12 pm »

I am sorry for not reading hte whole intial thread- but maybe this can be compiled in a sort of FAQ re. this topic?
As I see it, there already is an fNXT, and it is called 'NXT'.
And BCNXTs original vision was exactly that: using NXT as a network participation reward token, and Monetar System for all else.

So why not: create an NXT MS coin, and use that henceforth as currency, and keep NXT as NXT to forge.

I'm not 100% if I understand and can explain this, but the issue from my understanding is this:
If you make fNXT, you can prune everything from that blockchain except the vital information and be left with a very light weight main chain that everyone has to have.  If we keep NXT as the forging token, we have to keep all the current data that is in the blockchain already

The biggest issue though seems to be how to fairly setup 2 tokens without angering people  :)

TL;DR: fNXT is all about pruning and bloat elimination :)
« Last Edit: February 15, 2016, 04:08:00 pm by MrV777 »
Logged
NXT: NXT-BK2J-ZMY4-93UY-8EM9V
NXT nodes: 209.222.98.250, 216.155.128.10

Brangdon

  • Hero Member
  • *****
  • Karma: +229/-25
  • Offline Offline
  • Posts: 1389
  • Quality is addictive.
    • View Profile
Re: [NXT 2.0] A framework for thinking about incentives
« Reply #3 on: February 15, 2016, 05:07:54 pm »

If fNXT is the token that is used for forging (i.e. is the reqard token that forgers receive), and it is not be used for anything else, because anything else is what NXT is supposed to be used for:

How then is the route for forgers to monetize their fNXT?
By selling them to the people who bundle up side-chain transactions. NXT transactions need to be bundled up into ChildchainBlocks, and these need to be accepted into the fNXT chain. The people who create the ChildchainBlocks accept fees in NXT to do so, and then must pay fees in fNXT to the forgers. This creates demand for fNXT.

It also creates a kind of default market value. In a steady state and a free market, the value of all the NXT fees per minute will be slightly more than the value of the fNXT fees, because the NXT fees from all the transactions in a ChildchainBlock must be used to buy the fNXT fees to get the ChildchainBlock accepted.

Quote
Ultimately, the objective seems to be to decouple the currently dual function of NXT as currency and NXT as forging reward.
That's the means, not the goal. The goal is to make transactions almost completely prunable. However, the PoS security model means that transactions in the forging token cannot be prunable. Hence the split.

Quote
Or better- if re-evaluation, then on even terms such that the relative stakes of all participants do not change.
Was this discussed already?
Keeping the relative stakes the same is, I think, the default and most popular option, even if other options are fun to discuss. There are some people who strongly object to anything else, and I suspect those who like the other options don't feel nearly as strongly.
Logged

blackyblack1

  • Hero Member
  • *****
  • Karma: +165/-82
  • Offline Offline
  • Posts: 1764
    • View Profile
Re: [NXT 2.0] A framework for thinking about incentives
« Reply #4 on: February 15, 2016, 05:49:22 pm »

I don't know what this topic is about but here is a bit of calculation. I want to calculate a marketcap of the fNXT coin.
Let's suppose we will start with a single NXT sidechain with the same properties as we have today. I estimate daily forging rewards at 2500 NXT (I did ot with nxtportal info). Yearly rewards are 912000 NXT. Let's round up to 1M NXT. Let's define the marketcap at 100x yearly revenues (we will add speculative expectations here) so the marketcap is at 100M NXT. Looks like my initial estimate of the fNXT price being the 1/10 of the NXT was very close!
What do you think is it OK to let the coin with 100M NXT marketcap secure the 1B NXT supply? BTW we already have the same problem with NXT blockchain securing much higher marketcap of superBTC and potentially NXTUSD. So we will have fNXT with 900k USD marketcap securing the whole NXT ecosystem together with valuable assets. It looks dangerous.
Logged

durerus

  • Hero Member
  • *****
  • Karma: +106/-8
  • Offline Offline
  • Posts: 527
  • user-owner
    • View Profile
Re: [NXT 2.0] A framework for thinking about incentives
« Reply #5 on: February 15, 2016, 07:55:44 pm »

I don't know what this topic is about but here is a bit of calculation. I want to calculate a marketcap of the fNXT coin.
Let's suppose we will start with a single NXT sidechain with the same properties as we have today. I estimate daily forging rewards at 2500 NXT (I did ot with nxtportal info). Yearly rewards are 912000 NXT. Let's round up to 1M NXT. Let's define the marketcap at 100x yearly revenues (we will add speculative expectations here) so the marketcap is at 100M NXT. Looks like my initial estimate of the fNXT price being the 1/10 of the NXT was very close!
What do you think is it OK to let the coin with 100M NXT marketcap secure the 1B NXT supply? BTW we already have the same problem with NXT blockchain securing much higher marketcap of superBTC and potentially NXTUSD. So we will have fNXT with 900k USD marketcap securing the whole NXT ecosystem together with valuable assets. It looks dangerous.

By raising fees fNXT-forgers can always suck out value from child chains until security reaches a reasonable level.
Logged

LocoMB

  • Hero Member
  • *****
  • Karma: +101/-37
  • Offline Offline
  • Posts: 751
    • View Profile
Re: [NXT 2.0] A framework for thinking about incentives
« Reply #6 on: February 15, 2016, 08:11:53 pm »

 
By raising fees fNXT-forgers can always suck out value from child chains until security reaches a reasonable level.

sounds like a good dynamical argument...
Logged
TOX
90E54E5B5213290EE616D425CADC473038CFABFA53C913271AA8559D1937DC4AF3A354A9E4E5

durerus

  • Hero Member
  • *****
  • Karma: +106/-8
  • Offline Offline
  • Posts: 527
  • user-owner
    • View Profile
Re: [NXT 2.0] A framework for thinking about incentives
« Reply #7 on: February 15, 2016, 08:19:52 pm »

 
By raising fees fNXT-forgers can always suck out value from child chains until security reaches a reasonable level.

sounds like a good dynamical argument...

Sure, it's a measure in a case of emergency. But the market can anticipate. Child chains are useless without security. Everybody knows that. And forgers have central bank like power...
Logged

lurker10

  • Hero Member
  • *****
  • Karma: +168/-33
  • Offline Offline
  • Posts: 1334
    • View Profile
Re: [NXT 2.0] A framework for thinking about incentives
« Reply #8 on: February 15, 2016, 08:27:56 pm »

And forgers have central bank like power...

To a certain extent. Users of a sidechain can take a snapshot and migrate to start a new life on an independent network with their own forging if the fNXT forgers get abusive.
Logged
Run a node - win a prize! "Lucky node" project jar: NXT-8F28-EDVE-LPPX-HY4E7

durerus

  • Hero Member
  • *****
  • Karma: +106/-8
  • Offline Offline
  • Posts: 527
  • user-owner
    • View Profile
Re: [NXT 2.0] A framework for thinking about incentives
« Reply #9 on: February 15, 2016, 08:32:08 pm »

And forgers have central bank like power...

To a certain extent. Users of a sidechain can take a snapshot and migrate to start a new life on an independent network with their own forging if the fNXT forgers get abusive.
And again forgers will anticipate and won't overcharge. It's all about finding the equilibrium.
Logged

blackyblack1

  • Hero Member
  • *****
  • Karma: +165/-82
  • Offline Offline
  • Posts: 1764
    • View Profile
Re: [NXT 2.0] A framework for thinking about incentives
« Reply #10 on: February 15, 2016, 08:59:08 pm »

And forgers have central bank like power...

To a certain extent. Users of a sidechain can take a snapshot and migrate to start a new life on an independent network with their own forging if the fNXT forgers get abusive.
And again forgers will anticipate and won't overcharge. It's all about finding the equilibrium.
There are 4 directions of price movement.
1. Increase fees and decrease txs.
2. Decrease fees and increase txs.
3. Optimal fees (maximum NXT usage with maximum fees).
4. Minimum fees (a result of market based rates due to tragedy of commons).

The fees will finally find and equilibrium based on best proportion of txs for a given fee. But this equilibrium will not include proper network security in the equation since network security does not influence market. I'm not sure if it is a really bad thing since right now we have network security much lower than the sum of all assets on the AE. But the disbalance will be much higher after the fork.
Logged

lurker10

  • Hero Member
  • *****
  • Karma: +168/-33
  • Offline Offline
  • Posts: 1334
    • View Profile
Re: [NXT 2.0] A framework for thinking about incentives
« Reply #11 on: February 15, 2016, 09:05:56 pm »

The fees will finally find and equilibrium based on best proportion of txs for a given fee. But this equilibrium will not include proper network security in the equation since network security does not influence market. I'm not sure if it is a really bad thing since right now we have network security much lower than the sum of all assets on the AE. But the disbalance will be much higher after the fork.

If NXT are converted to fNXT through burning, the network security will be higher than it is now. We can assume these users who have burned their NXT to receive fNXT are dedicated to forge all the time, and most of the fNXT will be forging compared to 60% of stake not forging in NXT 1.x network.
Logged
Run a node - win a prize! "Lucky node" project jar: NXT-8F28-EDVE-LPPX-HY4E7

Brangdon

  • Hero Member
  • *****
  • Karma: +229/-25
  • Offline Offline
  • Posts: 1389
  • Quality is addictive.
    • View Profile
Re: [NXT 2.0] A framework for thinking about incentives
« Reply #12 on: February 15, 2016, 09:55:00 pm »

Let's define the marketcap at 100x yearly revenues (we will add speculative expectations here) so the marketcap is at 100M NXT.
So you are assuming fee revenue is the only thing that affects the market cap? For example, that they won't own and forge with fNXT in order to help secure their NXT and influence the network? Currently that is the main reason people forge. Why would that change?

When we have multiple child-coins, as I understand it they'll all be able to trade with fNXT but not directly with each other. If so, then fNXT will have a bit more utility and could become the default currency in which to park "resting" capital in the Nxt eco-system. It will be the easiest currency to convert quickly to other currencies, and it will be the only one that receives any interest in the form of fees.

If a company has its own child-chain, then it could make sense to keep a significant balance of fNXT in order to forge frequently in order to get free fNXT transactions. It can use those free transactions to provide free or subsidised transactions to its child-chain. I know this free-transactions benefit of forging exists today, but I think it will matter more when we have child-chains.
Logged

Jean-Luc

  • Core Dev
  • Hero Member
  • *****
  • Karma: +816/-81
  • Offline Offline
  • Posts: 1610
    • View Profile
Re: [NXT 2.0] A framework for thinking about incentives
« Reply #13 on: February 15, 2016, 10:38:44 pm »

Why are we comparing the marketcap of the NXT plus child chains to that of fNXT?

For an irrational attacker who wants to destroy the system regardless of whether he gains from it, only the absolute marketcap of fNXT matters, whether he can afford to buy and then sacrifice 51% of it.

For a rational attacker though, it matters how much he can gain by slowly obtaining 51% forging power and then exploiting it by double spending on the NXT and other child chains. But he has a limited time to do such double spending, and all assets and the NXT itself are highly illiquid. He would empty the asset order books, then the centralized exchanges open orders for NXT, and probably still not be able to recoup his 51% investment before the attack is detected and trading shut down.

And if we are talking about a future with a lot of transactions and liquidity, then that 2500 NXT forging fees per day also needs to be multiplied accordingly, and with it the fNXT marketcap estimate.
Logged
GPG key fingerprint: 263A 9EB0 29CF C77A 3D06  FD13 811D 6940 E1E4 240C
NXT-X4LF-9A4G-WN9Z-2R322

durerus

  • Hero Member
  • *****
  • Karma: +106/-8
  • Offline Offline
  • Posts: 527
  • user-owner
    • View Profile
Re: [NXT 2.0] A framework for thinking about incentives
« Reply #14 on: February 15, 2016, 10:46:21 pm »

...
4. Minimum fees (a result of market based rates due to tragedy of commons).

The fees will finally find an equilibrium based on best proportion of txs for a given fee. But this equilibrium will not include proper network security in the equation since network security does not influence market. I'm not sure if it is a really bad thing since right now we have network security much lower than the sum of all assets on the AE. But the disbalance will be much higher after the fork.

Marketcap calculated from last trades is not at all a sufficient measuring stick of the security of a POS-system. Market price of fNXT might be dumped down initially with low liquidity and due to almost nobody buying and some people dumping. But will the NXT community as a whole, whose participants hold NXT at dollar prices of below one cent, dump several hundreds of millions of main chain tokens at ridiculously low prices and knowingly sacrifice the security of the functions on the child chain they have been using for years? I don't think so. Zero marketcap of fNXT with no liquidity on the market and no serious centralization of fNXT visible on the blockchain (thanks to the TF we already have through POS) is not such a big problem as one may think on first sight. With a low market price and a low demand for fNXT transaction costs on the child chain can be very low, much lower than now. This can be seen as a competitive advantage over other cryptos!

To put it simple: When I see my main chain token dumped, I will not be afraid as long as I can see no centralization on the blockchain. And I will definitely grab some, because I know of the potential of the platform and that the platform cannot exist without the majority of main chain tokens used honestly.

@lurker10: That you don't forge doesn't mean that you will dump all your fNXT. As an investor in the NXT platform you will hold, because you speculate in future value. The split doesn't change that. As a speculator you just see that with 2.0 NXT gains potential without losing any of its functions. That's not at all a reason to dump your forging tokens at ridiculously low prices. It is a reason to buy at such prices. So IMO we don't need burning as a distribution model. 1:1 is the way to go.

P.S.: And because all that will be anticipated, ridiculously low prices of fNXT will never happen.

@Brangdon:
...
And the ability for trading to fNXT, to other child chain tokens, and to assets on other child chains if allowed.
« Last Edit: February 15, 2016, 10:49:18 pm by durerus »
Logged

lurker10

  • Hero Member
  • *****
  • Karma: +168/-33
  • Offline Offline
  • Posts: 1334
    • View Profile
Re: [NXT 2.0] A framework for thinking about incentives
« Reply #15 on: February 16, 2016, 08:27:37 am »

@lurker10: That you don't forge doesn't mean that you will dump all your fNXT. As an investor in the NXT platform you will hold, because you speculate in future value. The split doesn't change that. As a speculator you just see that with 2.0 NXT gains potential without losing any of its functions. That's not at all a reason to dump your forging tokens at ridiculously low prices. It is a reason to buy at such prices. So IMO we don't need burning as a distribution model. 1:1 is the way to go.

P.S.: And because all that will be anticipated, ridiculously low prices of fNXT will never happen.

I fail to understand reasoning behind burning as a distribution model being worse than 1:1. In my opinion, we can only make slightly educated guesses on how both models will function. Even with all the 2+ years of forging data the future is too vague to predict. I wouldn't call them arguments, it's just too unclear what will happen.

Here is my attempt to guess. Distribution by burning separates all users to three distinct classes: forgers, investors in NXT coins, investors in assets. It's a conscious decision to join one of the classes before the hardfork snapshot is taken. Users who make a decision to speculate on NXT coins or assets are unlikely to burn their NXT to get fNXT, why would they? The total supply of NXT will shrink and each of their NXT coin is guaranteed to become a little 'bigger' in respect to the total supply of NXT. Users burning NXT to get fNXT are taking a higher risk, users staying in assets are taking a higher risk. Higher risk is potential for a higher reward. Staying in NXT is a lower risk class. One educated guess I can make is most of the forgers will use their fNXT to forge. There will be speculators in fNXT but they will probably forge too a lot of the time. Forging is the only function of fNXT, why not forge? Most of the stake forging means increased security of the network.

1:1 model distribution offers less of conscious decision making and less of responsibility taking to the users compared to the burning model. It inflates the total supply to 2 bln, fudsters will love that.

I guess we need to estimate which model hurts less users and which model is most efficient and go with it.

The free market will decide the price. I don't expect ridiculously low prices of fNXT, there is no reason to expect it. You're right, there will be some sort of equilibrium reached, it's always reached if there is no intervention from outside and the rules of the game are clear to all.
Logged
Run a node - win a prize! "Lucky node" project jar: NXT-8F28-EDVE-LPPX-HY4E7

Brangdon

  • Hero Member
  • *****
  • Karma: +229/-25
  • Offline Offline
  • Posts: 1389
  • Quality is addictive.
    • View Profile
Re: [NXT 2.0] A framework for thinking about incentives
« Reply #16 on: February 16, 2016, 07:32:01 pm »

For an irrational attacker who wants to destroy the system regardless of whether he gains from it, only the absolute marketcap of fNXT matters, whether he can afford to buy and then sacrifice 51% of it.

For a rational attacker though, it matters how much he can gain by slowly obtaining 51% forging power and then exploiting it by double spending on the NXT and other child chains. But he has a limited time to do such double spending, and all assets and the NXT itself are highly illiquid.
It's been a basic principle of Nxt 1.x security that the forgers of the NXT chain must have a large stake in NXT. That principle is being dropped for Nxt 2.x. A forger could have no NXT at all. Hence they don't have to care if their actions damage the NXT chain.

While NXT is the only child-chain, and all the value of the system is tied up in it, it probably won't be an issue. In future we may have multiple child-chains, with some of them sponsored by companies or countries, and with other companies or countries wanting to damage their rivals by damaging their child-chains. They'll have incentive, and no costs, because they'll have no stake in the chain they are attacking. They can produce ChildChainblocks, and forge them, for free. This really bothers me.
Logged

blackyblack1

  • Hero Member
  • *****
  • Karma: +165/-82
  • Offline Offline
  • Posts: 1764
    • View Profile
Re: [NXT 2.0] A framework for thinking about incentives
« Reply #17 on: February 16, 2016, 07:54:13 pm »

Let's define the marketcap at 100x yearly revenues (we will add speculative expectations here) so the marketcap is at 100M NXT.
So you are assuming fee revenue is the only thing that affects the market cap? For example, that they won't own and forge with fNXT in order to help secure their NXT and influence the network? Currently that is the main reason people forge. Why would that change?
This is true. We actually have enthusiast forgers here but they are loosing interest over time. We can estimate the share of altruist forgers. 400M NXT coins are forging hence only 40% of users are interested in forging. I estimate forging being profitable (able to pay for the hosting) starting from 1M NXT stake. 75% of the forging coins are on the big accounts starting from 4M NXT. So maximum 10% of coins are forging not for profit. So we can add 10% margin to our calculations.

When we have multiple child-coins, as I understand it they'll all be able to trade with fNXT but not directly with each other. If so, then fNXT will have a bit more utility and could become the default currency in which to park "resting" capital in the Nxt eco-system. It will be the easiest currency to convert quickly to other currencies, and it will be the only one that receives any interest in the form of fees.
NXT assets will be visible on all chains so the best way to transfer value between chains is buying superBTC on one chain and selling on another. fNXT will not become a parking coin nor storage of the value not will have a high liquidity.
Logged

Brangdon

  • Hero Member
  • *****
  • Karma: +229/-25
  • Offline Offline
  • Posts: 1389
  • Quality is addictive.
    • View Profile
Re: [NXT 2.0] A framework for thinking about incentives
« Reply #18 on: February 16, 2016, 08:19:04 pm »

Let's define the marketcap at 100x yearly revenues (we will add speculative expectations here) so the marketcap is at 100M NXT.
So you are assuming fee revenue is the only thing that affects the market cap? For example, that they won't own and forge with fNXT in order to help secure their NXT and influence the network? Currently that is the main reason people forge. Why would that change?
This is true. We actually have enthusiast forgers here but they are loosing interest over time. We can estimate the share of altruist forgers. 400M NXT coins are forging hence only 40% of users are interested in forging. I estimate forging being profitable (able to pay for the hosting) starting from 1M NXT stake. 75% of the forging coins are on the big accounts starting from 4M NXT. So maximum 10% of coins are forging not for profit. So we can add 10% margin to our calculations.
That's a minimum, not a maximum. The 75% big accounts may or may not be forging for profit. The bigger accounts benefit more from revenue, but they also benefit more from securing their stake. The 25% who are forging at a loss are definitely doing it for reasons other than revenue.

I suspect many of the 60% who don't forge, do so to let the other 40% get more revenue. In other words, they are altruistic, but in a different way. They are giving up their forging revenue to encourage others.

Quote
When we have multiple child-coins, as I understand it they'll all be able to trade with fNXT but not directly with each other. If so, then fNXT will have a bit more utility and could become the default currency in which to park "resting" capital in the Nxt eco-system. It will be the easiest currency to convert quickly to other currencies, and it will be the only one that receives any interest in the form of fees.
NXT assets will be visible on all chains so the best way to transfer value between chains is buying superBTC on one chain and selling on another. fNXT will not become a parking coin nor storage of the value not will have a high liquidity.
"If allowed". If the market cap of fNXT being low is a problem, we may need to adjust the design to keep it high.
« Last Edit: February 22, 2016, 07:14:31 pm by Brangdon »
Logged

sadface

  • Sr. Member
  • ****
  • Karma: +16/-2
  • Offline Offline
  • Posts: 273
    • View Profile
Re: [NXT 2.0] A framework for thinking about incentives
« Reply #19 on: February 16, 2016, 08:22:02 pm »

NXT assets will be visible on all chains so the best way to transfer value between chains is buying superBTC on one chain and selling on another. fNXT will not become a parking coin nor storage of the value not will have a high liquidity.

where did you get that from?

e: i mean the part where nxt assets are visible on all chains.
Logged
Pages: [1] 2  All
 

elective-stereophonic
elective-stereophonic
assembly
assembly