My understanding is that:

  • The penalty applies to the block subsidy and not the transaction fees.
  • Under constant demand blocks can increase continuously by pushing up the median block size.
  • There doesn't seem to be a push back mechanism to confine block size except for reduction of the median by block space exceeding demand.

This means that as long as there are enough transactions waiting, blocks could always get full which would not incur a penalty. However, whenever there are enough transaction fees waiting to add e.g. 0.25% of the block subsidy with 5% additional block space, it would be economically viable to exceed the median block size by 5%.

In a scenario of constant demand excess such as Bitcoin is currently entering, blocks could continuously grow e.g. by the aforementioned 5% per day. Even a consistent growth of 5% per day would amount to 40% per week. As the block space supply could then grow exponentially and essentially unboundedly, all kinds of use cases that require cheap block space would be attracted, in turn creating a practically unbounded demand.

Meanwhile, assuming approximately homogeneous fee levels over the body of waiting transactions, e.g. transactions filling 5% of the block space should not exceed 0.0025 of the block subsidy, otherwise a bigger block would be more profitable to the miner. Hence, short term optimizing miners could have fees trending toward a small fraction of the block subsidy (for the assumed 5% growth it would be 5,25%).

This growth would be further exacerbated by the decreasing block subsidy reducing the cost of block size increases.

With sufficient demand, one would end up with a boundless block size, and a minuscule block reward, a tragedy of the commons. What am I missing?

  • 4
    There's a min. do-not-relay fee, so you can't just keep filling them cheaply. I'm working on a study (charts at the end might be of interest) and will gladly post some of it as an answer here once it's completed. – JollyMort Feb 20 '17 at 14:45
  • What you are describing does indeed seem like an interesting and plausible edge case. However, it's difficult to provide an answer when the question doesn't have any stated bounds. For example, it would be helpful to state when the tragedy of the commons is likely to occur, given certain assumptions. Then someone could either take issue with the assumptions, or they might go with your assumptions but provide a different take on the timing and effect of the outcome you're describing. Fortunately for you, @JollyMort has spent a great deal of time on this recently. – scoobybejesus Feb 23 '17 at 0:36

With sufficient demand, one would end up with a boundless block size, and a minuscule block reward, a tragedy of the commons. What am I missing?

I would argue you are missing physics. At some point, the physical limits of the network bound the blocksize. I.e., if a miner decides to keep pushing the blocksize higher, eventually a blocksize will be reached that takes considerably longer to propagate through the network than a smaller block.

Indeed, there are two prerogatives for blockchain mining:

  1. Crafting a block with the highest reward possible
  2. Propagating that block to the entire network before a competing block propagates through the network

The evidence for #2 can be seen in the bitcoin network - mining pools opt to mine empty blocks even though they could receive more bitcoin if they mined blocks with transactions. (although this behavior seems to have stopped recently with the advent of centralized block relay networks)

So, as is often the case, the phrase "tragedy of the commons" is missing a key qualifier that the coiner of the term realized he should have added - "Tragedy of the unregulated commons". In our case, the regulation is physics, so the Monero network will always adapt at the rate of the technology that runs the software.

Edited to add: The dynamic fee algorithm also provides an incentive to keep the blocks small. As the block size increases, the minimum relay fee decreases. Thus, even though you are filling the blocks with more transactions, it not necessarily the case that the transaction fees are worth more.

  • Craig Wright’s Scaling Bitcoin presentation explained that propagation is an economic phenomenon and new blocks propagate to 99% of the hashrate (not all nodes) in Bitcoin within a couple of seconds. He also claims Moore’s law (exponential growth) outpaces transaction demand growth (unless it is also exponential, not logistic). As I told @ArticMine long ago, Monero has no fee market either without an oligarchy of mining. – Shelby Moore III Jul 31 '17 at 16:47
  • Also a fee market is not viable because research seems to model that proof-of-work becomes incentives incompatible once revenue from transaction fees becomes significant relative to the protocol block reward. Really it seems the mess is only sustainable with an oligarchy on mining, which itself is not sustainable. – Shelby Moore III Jul 31 '17 at 18:41
  • 1
    @ShelbyMooreIII welcome to Monero SE! Would you be willing to expand upon your comments with another answer to this question or a related question that you create (instead of just linking to external sources where you address part of the question indirectly)? – Smart Kid Aug 1 '17 at 3:35
  • Re: edit about minimum relay fee decreasing, I presume that is miner policy about relaying transactions to each other, not economically enforced by the longest chain protocol, thus miners are free to ignore it. In any case, as fee revenue becomes significant relative to tail reward, incentives go haywire, so any model has to incorporate those other strategies. Thus I conclude there is no fee market w/o an oligarchy of miners. @Smart Kid, maybe if I can find time to organize all my thoughts on this issue carefully enough. – Shelby Moore III Aug 1 '17 at 12:28
  • Craig Wright is a known fraud. Anyway, Moore's Law is over arstechnica.com/information-technology/2016/02/… so not sure any conclusion based on that can be valid. – hyc Aug 1 '17 at 13:39

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.