The bitcoin network has returned to somewhat normal operations after considerable backlog earlier this month which sent fees to $5 or more per transaction.
That backlog has now seemingly cleared to what appears to be insignificant levels with some reporting transaction confirmations even for fees that only pay pennies, although 21.co is still recommending a transaction fee of $2.50.
This cyclic high congestion followed by a return to somewhat normal operations has been noted a number of times, suggesting there may be two factors.
The first one is speculative and theoretical. As fees increase, individuals become economical with transactions, avoid using the network, so reducing demand and thus reducing fees.
Then, once fees are reduced and especially if there is any fast price movement, they start transacting again, spending their bitcoins or moving them to exchanges.
The second factor is that it may well be the case bitcoin is designed, whether intentionally or otherwise, to not have a so-called fee market for two main reasons.
Firstly, blocks are generated at somewhat random times. They are meant to average 10 minutes between blocks, but at times there can be one hour without a block or there can be 15 blocks in one hour.
If each block clears the backlog that would be fine, but if it doesn’t, then suddenly fees spike if it takes 20 or 30 minutes to find a block, so pricing out those that paid the recommended fee at the time, getting them stuck.
The second technical aspect is difficulty which is somewhat related to the above, but also has its independent aspect. As hashpower is added before difficulty has adjusted, then blocks are found faster.
That’s what happened in this case and cleared the backlog. Hashpower moved from Bitcoin Cash to bitcoin after the latter became more profitable to mine. Blocks therefore started being found faster, so in effect increasing capacity and thus clearing the backlog.
But at some point difficulty adjusts. Which in effect means capacity is reduced. Thus the recommended fee is no longer sufficient, so bitcoiners get stuck.
These two factors mean that there can not be a predictable fee market. This can somewhat be alleviated by incorporating Replace by Fee (RBF) whereby you double spend your own transactions by paying a higher fee.
That doesn’t address the unpredictability, but it can at least get you unstuck. However, it may well be a case of turtles all the way up as the same unpredictability would probably apply to RBF.
We haven’t yet seen it in action. It’s not clear why Bitcoin Core developers have not implemented it as the default option because although there was public resistance, a network with full blocks needs RBF. Otherwise, people are left waiting for three days.
In any event, now that we have seen how a full blocks system operates in action, it appears that instead of fees ever growing, demand reacts to supply, and reduces as fees increase. Making it all cyclical.
That suggests the settlement system might not quite work as envisioned and the scalability debate may have been for nothing as an on-chain increase whether LN or not seems inevitable.