search

bitcoinfees / bitcoin-feemodel

Bitcoin transaction fee modeling and estimation
https://bitcoinfees.github.io
MIT License
9 stars 2 forks source link

suggestion: it would be really useful to include mempool total value (utxo totals), and total value xfer rate #2

Open tphyahoo opened 8 years ago

tphyahoo commented 8 years ago

For mempool value (utxo total sat in mempool): I would try to include this first by adding a second y axis for mempool utxo totals. If that screws up chart visually, it could be added as additional chart instead.

This would reveal if mempool spikes are likely spam, or looks like real value is pushing up the mempool.

A related metric would be total value xfer rate, i.e. utxo total sat / kB. If the fee rate is spiking, but the total value xfer rate is crashing, this would be suspicious, or at any rate evidence that there are plenty of low value transactions that could wait their turn without making more block space available.

If both fee rate and total value xfer rate are spiking, that means there are high values of bitcoin trying to move around, that are getting close to being crowded out. Which would be evidence that big blockers are right, and the time to 2mb-fork is closer than the small blockers think.

kravets commented 8 years ago

0.12.0 version eliminates what's called "priority transaction" sub-pool, i.e. upto 50,000 bytes in each block that was allocated to transactions with large amount of days destroyed, etc ... after the size of this "priority area" is changed to 0 by default, no transaction with the fee under the dust relay minimum will propagate at all. At this point, one can argue that there is no more spam left bc if a transaction is paying more than dust relay minimum, then it's not spam (according to Bitcoin's working definition).

tphyahoo commented 8 years ago

My definition of spam is suspect uneconomic transactions: transactions created specifically to manipulate statistics and public perception. It's impossible to prove definitely where tx statistics manipulation is taking place, but my purpose with this suggestion is to make it as difficulat as possible, by providing public metrics that are increasingly difficult/expensive to convincingly fake.

bitcoinfees commented 8 years ago

Thanks for the suggestions.

It's indeed useful, but outside the scope of this project, which is simply to estimate the required tx fee needed for a given wait time.

I believe tradeblock does show the data that you are looking for (under "recent mempool").