Byteball spends transaction fees to the witnesses (and perhaps the payer portion is effectively burned as it is passed along?) instead of employing proof-of-work (but I am not yet clear if this is used as the metric of the chain length in any way in consensus algorithm).. These fees per section 1. Introduction: Exchange rate on page 3 are tied to the system wide exchange value of adding bytes to the database.
Furthermore, I am sorry to say that my analysis is that once the exchange price has risen high enough, the transaction fee will exceed the value of the transaction and so the transaction can't be economic.
This appears to break the coin.
The transaction fee is equal to transaction size. How can it be related to the exchange price?
Because the bytes being spent in the transaction have an exchange value. As the exchange value rises, then the quantity of bytes that will be transacted will decrease. When it decreases below the size of the transaction, the transaction is no longer economic.
Well, while theoretically possible, the valuations need to be quite high for this to happen. The size of a simple transaction is about 500 bytes, let's round it up to 1000 bytes. For this fee to be equal to $1.00, the market cap will have to be $10
12, or $1tn.
Okay good point, but the counter point is that we shouldn't assume that all lower valued transactions will be necessarily simple transactions. Let's think multi-sig, smart contracts, etc, then we might see problems at around $10 - $100 billion market. And if those are micropayments (which smart contract gas is for example), then we are getting into potential trouble at $100 million market cap.
Typical multi-sig and smart contracts are not going to grow the size of transaction by orders of magnitude.
~2000 bytes per Ethereum transaction and that is diluted since not all of those are smart contracts. I figure 10,000 bytes per smart contract might be the average, which is an order-of-magnitude higher than your example. And we are concerned here not with the average, but the outlier transaction that becomes uneconomic to do. So that might be 100,000 bytes.
And I don't see much use of Byteball in microtransactions, in particular because we don't address the blockchain bloat (or rather ledger bloat) problem.
Okay but the gas payment of a smart contract is typically a microtransaction. Also many smart contracts are of microtransaction in value, i.e. there are many services that require many, many smart contract executions each of microvalue but in aggregate of great value.
Also there is probably also a tradeoff in that the lower the value of the fees for witnesses at low market caps, then perhaps the more incentive witnesses have to game and attack the system. That is getting into what I alluded to by "nothing-at-stake".
That would be true if the ability to earn fees were the only witness' stake. But it isn't. As I said, the stake is outside the system, it is the reputation, the trust, the business that a witness has in the real world and would damage if it were to misbehave.
But their reputation on the coin itself is only as valuable as the fees they can earn with that reputation. Their external reputation may still have the same value even if they profit by attacking, destroying, and shorting on the exchanges some two-bit altcoin. They don't have any huge stake in the coin itself, neither in the form of mining hardware investment nor in the form of opportunity cost of huge future fees forsaken. I am not going to waste my time arguing with someone who isn't well versed/experienced with economics. You asked me to present the flaws I see, so I have presented. I see no benefit to arguing. Either we can agree on rational truth, or not. I will submit if you provide a truthful and compelling argument. Otherwise it means I think you are incorrect.
Separately, per my prior post a few moments ago, my secret chain branch can employ entirely fake witnesses and Sybil addresses and how can you objectively determine the difference?
