CounterpartyXCP / counterwallet

Counterparty web wallet
https://counterwallet.io
147 stars 162 forks source link

Default 'optimum' fee too high #824

Open Jpja opened 6 years ago

Jpja commented 6 years ago

https://github.com/CounterpartyXCP/counterwallet/blob/eb26f2e2cdb525160ff508a3312374dd0c2ec192/src/js/components/wallet.js#L664

The default fee is set to 'optimal'. It tends to overshoot significantly, hence it's more of a high priority fee than an optimum.

Even with 'low_priority' I've usually had my transactions confirmed in the first block.

At xchain right now (it uses the same estimates, I believe ) the 'optimal' fee is $4.05 per kB and low is $0.18.

Since 1) users are not informed of the fee they end up paying, 2) sometimes a custom fee is overwritten and set to default, 3) not all actions have a custom fee option and 4) fees can sometimes be several dollars, I'd prefer to err on the side of lower default fee.

The fix is simply changing line 664 to something like this: data['fee_per_kb'] = (fee_per_kb['low_priority'] * 3 + fee_per_kb['optimal']) / 4;

rubensayshi commented 6 years ago

completely disagree, I'd prefer to err on the on side of higher default fee than neccesary, complaints of txs stuck, trades failed etc are much worse than complaints of high fees...

counterblock has an option to use the blocktrail API, if not it uses bitcoind with nblocks=3 for optimal which should be fine, with 0.15 probably even slightly lower than blocktrail API.

if anything we should just fix the issues where the user doesn't get to choose low_priority imo

Jpja commented 6 years ago

Two separate considerations here


From real usage I've noticed that the optimal fee (nblocks=3) is completely off. I compared the fee ratio I'm charged with mempool - https://blockchain.info/charts/mempool-state-by-fee-level

chiguireitor commented 6 years ago

I'm also concerned with fees, but as Ruben says, if there's options of choosing the kind of fee on every transaction, that would rock.

deweller commented 6 years ago

I'm good with JP's suggestion here. I think a 75% weighted low_priority fee is a good compromise here until we can upgrade to bitcoin 0.15.0.

In my experience, the counterwallet fees from bitcoind are overkill.