mirror of
https://git.wownero.com/wownero/wownero.git
synced 2024-12-22 23:08:53 +00:00
blockchain: tweak fee as a function of median values
Use the lesser of the short and long terms medians, rather then the long term median alone From ArticMine: I found a bug in the new fee calculation formula with using only the long term median It actually needs to be the lesser of the long term median and the old (modified short term median) short term median with the last 10 blocks calculated as empty Yes the issue occurs if there is a large long term median and, the short term median then falls and tries to then rise again The fees are could be not high enough for example LTM and STM rise to say 2000000 bytes STM falls back to 300000 bytes Fees are now based on 2000000 bytes until LTM also falls So the STM is could prevented from rising back up STM short term median LTM long term median
This commit is contained in:
parent
b8643752c1
commit
d37d30f79a
@ -3334,7 +3334,8 @@ uint64_t Blockchain::get_dynamic_base_fee_estimate(uint64_t grace_blocks) const
|
||||
}
|
||||
|
||||
const bool use_long_term_median_in_fee = version >= HF_VERSION_LONG_TERM_BLOCK_WEIGHT;
|
||||
uint64_t fee = get_dynamic_base_fee(base_reward, use_long_term_median_in_fee ? m_long_term_effective_median_block_weight : median, version);
|
||||
const uint64_t use_median_value = use_long_term_median_in_fee ? std::min<uint64_t>(median, m_long_term_effective_median_block_weight) : median;
|
||||
const uint64_t fee = get_dynamic_base_fee(base_reward, use_median_value, version);
|
||||
const bool per_byte = version < HF_VERSION_PER_BYTE_FEE;
|
||||
MDEBUG("Estimating " << grace_blocks << "-block fee at " << print_money(fee) << "/" << (per_byte ? "byte" : "kB"));
|
||||
return fee;
|
||||
|
Loading…
Reference in New Issue
Block a user