Initial allocation and fundraising

Previous topic - Next topic

psztorc

That sounds both overly complex and out of scope.

People won't be able to "use" this software for anything until they believe that it will continue to work reliably for a few consecutive months. For many people, that will not be until they personally see it working for a few consecutive years.
Nullius In Verba

Max

...so You must guarantee, that Truthcoin is CPU mineable ?
No ASICs, no other special hardware etc
so, no Scrypt hashing...

What will You do ?

psztorc

Quote from: Max on July 10, 2014, 09:45:16 PM
...so You must guarantee, that Truthcoin is CPU mineable ?
No ASICs, no other special hardware etc
so, no Scrypt hashing...

What will You do ?
http://forum.truthcoin.info/index.php/topic,20.0.html
Nullius In Verba

Max

Thanks for the link :)
I asked only because i worried about fairness of allocation in ASIC involved coin.
Maybe You can do something for fairness even with double  SHA....who knows.

psztorc

Quote from: Max on July 11, 2014, 08:28:21 AM
I asked only because i worried about fairness of allocation in ASIC involved coin.
Maybe You can do something for fairness even with double  SHA....who knows.
Look at slide 4 of my presentation draft: https://github.com/psztorc/Truthcoin/raw/master/docs/TruthcoinValuable.pptx

(http://forum.truthcoin.info/index.php/topic,109.0.html)

If you already own Bitcoin, you'll own this coin as well. Future block rewards will likely go to ASIC-miners, as they do today for Bitcoin.
Nullius In Verba

psztorc

I have just finished another small presentation which examines this question more directly: https://github.com/psztorc/Truthcoin/raw/master/docs/InitialCoinAllocations.pdf
Nullius In Verba

koeppelmann

#21
I like the idea of a snapshot of the Bitcoin distribution for the initial cash-coin distribution.
However, I would make a opt in process necessary afterwards. So the snapshot is done and afterwards every user has like 6 month to claim its cash coins. First, this would have a great marketing effect to get user to download and try out the client. And second, after the 6 month the uncertainty about how much cash-coins are in circulation is much smaller.

keystrike

How much are you looking to fundraise?

psztorc

Money actually isn't the bottleneck, dev talent is. (This thread was started by one of the rare breed of 'good blockchain developers', "toast" from the Bitshares dev team).

Whatever it takes to get 1-2 extremely high quality developers.

(So the answer depends on finding developers and asking them "What do you want?")
Nullius In Verba

zack

Dyffy is paying me $1000 a month (minus tax) to build their version of truthcoin.
I am not loyal. If anyone would pay me more, I would work on their ideas instead.

The python version of truthcoin I built is working pretty well. It does all the prediction market stuff truthcoin is supposed to do. We have tested it with 5 nodes at a time. We have run it continuously for 4 days without it crashing.
My head-start on creating this software is very big. It will be multiple months before we see any other implementation start to catch up to where I am.

My guess for the most valuable truthcoin 2 years from now: Majority of coins will be owned by a rich person who can afford legal fees and advertising fees (and possibly mining fees). They will hire me to maintain it. Maybe it will be Dyffy, but I doubt it.

If truthcoin is launched as a proof-of-work blockchain, I expect there to be a period of time about as long as a year when we will need to pay $100-$1000 daily into mining costs. I cannot afford this. Until a rich person shows up willing to pay, I am focusing my attention on proof of stake, which will costs pennies a day to keep operational.

keystrike

Who is Dyffy?

How did you come to your $100-$1000 daily mining cost estimate?

zack

Dyffy is a startup trying to launch truthcoin.

We need to invest enough money into mining to making it prohibitively expensive to attack the chain.
If we invest $100 per day, then it would cost $4 to sustain an attack for an hour.

There is no magic number that would make us safe. If we do POW, it will be a compromise between needing security and not being able to afford security.

POW is a very expensive consensus mechanism.
Bitcoin costs $1.2 million per day to keep running right now.
So it would cost around $48,000 per hour to sustain an attack against bitcoin.

Bitcoin is actually a lot safer than $48,000 because the market for ASICS is not as liquid as the market for general purpose CPU cycles. Since our blockchain is going to be mined by CPU at first, it's market will be almost perfectly liquid. An attack will actually cost as much as my calculations show.

Proof of stake, if it works, would cost dollars a day to operate, even if the blockchain was as big as bitcoin.

I started studying how to make FPGAs. They are much more powerful than CPU mining, but not as powerful as ASICS.
It is an in-between step before making ASICS.
One way to make POW more affordable would be if we produced and sold custom truthcoin mining hardware that only worked for our blockchain.

keystrike

#27
Thanks for the reply!

Is there a link for Dyffy or any posts they made somewhere? Who runs it?

I am somewhat skeptical of changes to the POW hash function mainly because I don't think they have been necessary in the altcoin world (although I do not believe in alts in general -- Truthcoin is of course different as it is not really an alt). I think merged mining is always better as then you get the benefits of the bitcoin blockchain. In one of his papers, Poelstra mentions a few things to watch out for such as ease-of-verification, progress-freeness, optimization-freeness, and simplicity (mainly to avoid centralization).

POW is expensive but using resources in the real world seems to be the only way to protect decentralized consensus networks. I think Truthcoin might be hard enough to implement so why not stick to POW and use merged mining? Trying to create dedicated hardware at this point would be additional capital cost and be further centralization.

Would it be possible to do POW with a private blockchain at first for testing purposes?

Edit: Merged mining : http://dot-bit.org/Merged_Mining

Merged mining is dangerous unless all miners jump in as it would make it easier for one to do a 51% attack. (Like what Luke-Jr. did to CoiledCoin.)

turpin

#28
Rather than either shares based on Bitcoin snapshot or shares to miners, I suggest a rewards of Truthcoin in proportion to total coin (Bitcoin + Truthcoin) time participation (which is proportional to Bitcoins bet on the wager).  Having consistent rules independent of time or block number would avoid accusations of preming or unfair early adopter rules.  Opportunities for participation will increase as the gambling increases, so although early adoptiong will have advantages, early adoption before gamblers become aware of the network will have only marginal advantage.  I imagine initial Truthcoin shares would be sent to the Truthcoin address corresponding to the participating Bitcoin address.  Voting/participating both your Bitcoin and Truthcoin shares would be easy because they are the same address on different networks, thus have the same keys and the same signature, making Truthcoin share payouts easy to compute.  Bitcoin payouts of commission earning to Truthcoin shareholders would be accomplished simply by sending Bitcoin to the Bitcoin address
identical to the Truthcoin address.  Mining rewards likewise would be sent to the Bitcoin address that the Truthcoin miner indicates - making Truthcoin among the first mergemining operation to pay miners in Bitcoin.

So we get the recursive formulas.  If
coin@n = (Bitcoin@n +Truthcoin stake @n), and
participation@n = (+/-)wagers@n, and
Truthcoin@(n+1) = Truthcoin@n + (coin@n)*(participation@n)
Then,
Truthcoin@(n+1) =  Truthcoin@n + (B@n +Truthcoin stake @n)*((+/-)wagers@n)

Thus for the special case of constant wagers (wagers@n = w for all n = 0,N, otherwise wagers@n=0 for n<0) for each block, with no share transfers at a particular address so Bitcoin held @ n = constant B:
Truthcoin @ 1 = B*w
Truthcoin @ 2 = B*w + (B + Truthcoin@1)*w
                      =  B*w + (Bitcoin +  Bitcoin*w)*w = 2*B*w + B*w^2
Truthcoin @ N = N*B*w + (N-1)Bitcoin*w^2 + ... + 2*B*w^(N-1) + B*w^N
                      = B*Sum(N*w^(N+1-n))
Since I can't do infinite series as well as I could in college, I will now resort to proof by Wolfram:
http://www.wolframalpha.com/input/?i=B*sum+(N*w%5E(N%2B1-n)),+n%3D1+to+N
Truthcoin @ N = B*N*w*(w^(N-1))/(w-1)
(lim n->infinity)(w/(w-1)) = 1
Thus for large w, Truthcoin @ N ~ B*N*(w-1)^N, so "exponential" growth
Also
For small wagers, w = 0.5, Truthcoin @ N ~ N. (Exercise for the reader - I just verified this with a calculator at a few values of N.)

So basically participation has diminishing returns of shares with time unless wagers become at least 0.5 Bitcoin per block, which is where share growth linearly increases with time. The exact threshold can be changed by throwing in some adjustments, but accusations of premining shares become provably false because participation becomes increasingly valuable not only with persistence (large N) but even more so with large wagers.  You can't really premine or even get an unwarranted early adopter advantage as long as wagers are still increasing and your Bitcoin value is constant.

So what is the incentive for early adoption then?  Because the assumption of constant Bitcoin is false.  Early adopters will get dividends in Bitcoin, which increases their Bitcoin share, giving them a further advantage over late adopters.  This only works though if they have significant savings, that is they only spend a fraction of their income.  If they plan to spend all their earnings, they will asymptotically approach the interest share of late adopters with the same Bitcoin holding.  Also, Bitcoin is supposedly deflationary against fiat currency so people who are late adopters to both Bitcoin and Truthcoin are at a double disadvantage.

I believe this type of allocation will easily make the Truthcoin and Bitcoin networks self-reinforcing and make betrayal dis-incentivized. The time-invariance of rules is very important because it makes altcoins or alt-side-chains uncompetitive against an established network unless they have real added economic value by specializing. If rules have some explicit time dependency or "premining"/"preallocation" then a community of sufficiently wealthy investors might be motivate to make an independent network rather than sink capital into acquiring interest in an existing network. Early announcement doesn't eliminate that possibility, it only reduces it in the short term.

Additionally, if there are transaction fees for converting Bitcoin*participation->Truthcoin, it will encourage people to hold their Bitcoins in a single output.  This will unburden the Bitcoin blockchain of unnessary transactions and untrimmable junk from multiple address/multiple output wallets.  Outputs that are economically combinable into one output will get combined by Truthcoin shareholders seeking more Truthcoin, rather than waiting indefinitely long for "free" transaction eligibility. And this doesn't just happen once, but every block that pay Truthcoin dividends will incentives a transaction to combine multiple Bitcoin outputs into single Bitcoin outputs. Truthcoin will add a time value and transaction fees to Bitcoin beyond whatever time value and transaction fees it already has. This will benefit miners greatly - even if they irrationally don't participate in merge mining.

zack

I don't understand turnpin's proposal at all.
Are you talking about allocating the initial coins?
Are you proposing to give the coins to miners? bitcoin holders? It sounds like you want to give them to gamblers.

It is possible for me to bet against myself, and neutralize my risk. It makes it look like I am gambling a lot.
The money is inaccessible for the time it is locked up to gamble. So this is identical to paying someone for making a security deposit. It runs into the same inefficiencies as a bond-based POS distribution http://www.truthcoin.info/blog/pow-cheapest/#the-coinbase-rot-paradox-less-is-more
search for "Tendermint"