sipa changed the topic of #bitcoin-wizards to: This channel is for discussing theoretical ideas with regard to cryptocurrencies, not about short-term Bitcoin development | http://bitcoin.ninja/ | This channel is logged. | For logs and more information, visit http://bitcoin.ninja
tromp has quit [Read error: Connection reset by peer]
tromp has joined #bitcoin-wizards
bsm117532 has quit [Remote host closed the connection]
rmwb has quit [Remote host closed the connection]
danrobinson has joined #bitcoin-wizards
rmwb has joined #bitcoin-wizards
goatturneer has joined #bitcoin-wizards
beatrootfarmer has quit [Ping timeout: 246 seconds]
rilos_ has joined #bitcoin-wizards
rilos has quit [Ping timeout: 246 seconds]
thrmo has quit [Ping timeout: 246 seconds]
thrmo has joined #bitcoin-wizards
tromp has quit [Remote host closed the connection]
tromp has joined #bitcoin-wizards
bsm117532 has joined #bitcoin-wizards
bsm117532 has quit [Ping timeout: 260 seconds]
tromp has quit [Read error: Connection reset by peer]
rilos_ has quit [Ping timeout: 260 seconds]
rilos has joined #bitcoin-wizards
bsm117532 has joined #bitcoin-wizards
bsm117532 has quit [Client Quit]
bsm117532 has joined #bitcoin-wizards
Chris_Stewart_5 has joined #bitcoin-wizards
Ylbam has quit [Quit: Connection closed for inactivity]
pro has quit [Quit: Leaving]
jtimon has quit [Ping timeout: 240 seconds]
tromp has joined #bitcoin-wizards
thrmo_ has joined #bitcoin-wizards
danrobinson has quit [Quit: danrobinson]
tromp has quit [Remote host closed the connection]
thrmo has quit [Ping timeout: 240 seconds]
danrobinson has joined #bitcoin-wizards
danrobinson has quit [Client Quit]
danrobinson has joined #bitcoin-wizards
Chris_Stewart_5 has quit [Quit: WeeChat 1.4]
rilos has quit [Ping timeout: 260 seconds]
rilos has joined #bitcoin-wizards
tromp has joined #bitcoin-wizards
beatrootfarmer has joined #bitcoin-wizards
goatturner has joined #bitcoin-wizards
goatturneer has quit [Ping timeout: 245 seconds]
MaxSan has joined #bitcoin-wizards
dabura667 has joined #bitcoin-wizards
beatrootfarmer has quit [Ping timeout: 260 seconds]
MaxSan has quit [Ping timeout: 240 seconds]
goatturneer has joined #bitcoin-wizards
MaxSan has joined #bitcoin-wizards
goatturner has quit [Ping timeout: 260 seconds]
danrobinson has quit [Quit: danrobinson]
tromp has quit [Remote host closed the connection]
MaxSan has quit [Quit: Leaving.]
goatturner has joined #bitcoin-wizards
goatturneer has quit [Ping timeout: 260 seconds]
tromp has joined #bitcoin-wizards
danrobinson has joined #bitcoin-wizards
rmwb has quit [Remote host closed the connection]
MaxSan has joined #bitcoin-wizards
thrmo has joined #bitcoin-wizards
thrmo_ has quit [Ping timeout: 268 seconds]
legogris has quit [Remote host closed the connection]
danrobinson has quit [Quit: danrobinson]
rmwb has joined #bitcoin-wizards
rmwb has quit [Remote host closed the connection]
legogris has joined #bitcoin-wizards
rmwb has joined #bitcoin-wizards
tromp has quit [Ping timeout: 268 seconds]
goatturneer has joined #bitcoin-wizards
goatturner has quit [Ping timeout: 245 seconds]
thrmo has quit [Ping timeout: 240 seconds]
beatrootfarmer has joined #bitcoin-wizards
[7] has quit [Disconnected by services]
TheSeven has joined #bitcoin-wizards
goatturneer has quit [Ping timeout: 246 seconds]
rmwb has quit [Remote host closed the connection]
rmwb has joined #bitcoin-wizards
dnaleor has quit [Ping timeout: 258 seconds]
goatturneer has joined #bitcoin-wizards
beatrootfarmer has quit [Ping timeout: 260 seconds]
<midnightmagic>
dang he left before I could thank him
nona has joined #bitcoin-wizards
nona has quit [Read error: Connection reset by peer]
rilos has quit [Ping timeout: 240 seconds]
rilos has joined #bitcoin-wizards
Pr0t3us has joined #bitcoin-wizards
<adam3us>
what do people think about fee estimation, is it biased to go up in dollar terms other things being equal, if USDBTC increases? the fee estimation is automated market referring backwards to recent satoshis/byte rates. samson has made this point a few times and now I am wondering if this even is behind some fee increases
<adam3us>
the only way it falls is if individuals override wallet defaults arrived at via fee market esimation, and maybe that's a small enough effect to not have much hysteresis.
coredump_ has quit [Ping timeout: 246 seconds]
BashCo has quit [Ping timeout: 245 seconds]
<kanzure>
adam3us: you want some anti-hysteria fee estimation protection?
<adam3us>
seems like fee estimation could usefully take into account a BTCUSD price feed. I am supposing at least the core one would not because what would be the price feed.
paveljanik has quit [Quit: Leaving]
<sipa>
adam3us: what does BTCUSD have to do with it?
<adam3us>
sipa so samson's theory was that automated fee estimation, looking only at historic sat/byte would automatically grow in USD terms as BTCUSD increased. at $300 BTCUSD a 50c fee feels high but acceptable, at $3000 BTCUSD a $5 fee seems to high for some use cases.
rilos_ has joined #bitcoin-wizards
<sipa>
adam3us: so? free market
<sipa>
if the fee is too high, fewer people will pay it
rilos has quit [Ping timeout: 245 seconds]
<adam3us>
sipa but it's a machine market. if the human feedback loop is too weak, and people grumble but dont know how or service doesnt provide option etc for most users, then the hysteresis maybe isnt enough to not artificially push the fees up, just via unlinked to USD automation.
<sipa>
but BTCUSD doesn't even enter the picture
<sipa>
i don't see how you could even use it
rmwb has quit []
<adam3us>
sipa i know. how you would use it is overlay historic USD price on the sat/byte so it's c/kB or what have you.
<adam3us>
sipa if samon is right and the hysteresis human feedback loop is weak, this would result in more stable fees without shooting up each time the price increases, and waiting for pain point for users to slowly fumble around, or services adjust their fixed fee pricing etc. some services dont care so much because user pays even. even big services with millions of users.
<adam3us>
s/samon/samson/
harrymm has quit [Ping timeout: 246 seconds]
BashCo has joined #bitcoin-wizards
harrow has quit [Quit: Leaving]
ivan has quit [Quit: lp0 on fire]
<waxwing>
btcusd does not enter into the picture, but btc/purchasing power does. usd is a decent proxy (or eur, whatever you prefer).
harrow has joined #bitcoin-wizards
ivan has joined #bitcoin-wizards
rilos has joined #bitcoin-wizards
rilos_ has quit [Ping timeout: 240 seconds]
tromp has joined #bitcoin-wizards
sn0wmonster has left #bitcoin-wizards [#bitcoin-wizards]
Belkaar has quit [Ping timeout: 268 seconds]
Belkaar has joined #bitcoin-wizards
Belkaar has joined #bitcoin-wizards
Belkaar has quit [Changing host]
tromp has quit [Ping timeout: 246 seconds]
btcdrak has quit []
harrymm has joined #bitcoin-wizards
rilos has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<adam3us>
yes but the question is specific. there is a direct mechanical driver between recent fees in satoshis/byte and current estimate. but there is only a weak, reactive and mostly indirectly (via often non-cost exposed) intermediaries on the hysteresis to normalise in purchasing power terms. therefore i suspect samson is right and this is artificially driving fees higher than otherwise
<sipa>
if that's the case, the fee market is not working
tromp has joined #bitcoin-wizards
dabura667 has quit [Remote host closed the connection]
nona has joined #bitcoin-wizards
nona has quit [Client Quit]
tromp has quit [Ping timeout: 240 seconds]
laurentmt has joined #bitcoin-wizards
laurentmt has quit [Client Quit]
vicenteH has joined #bitcoin-wizards
<adam3us>
sipa minimally in a rising BTCUSD market, with a mechanical driver, and a weaker reactive on the downside - a few people who now how or even have direct control (not set by their fee agnostic service provider who is passing the fee on to them, their only means to lower fee is to complain or switch service provider which may have a high switching cost), then it will be like petrol prices
<adam3us>
ie if there the price of crude oil goes up, pump prices instantly go up, when crude oil falls there's a hysteresis while people start nudging prices down by shopping around, if they are price sensitive, it takes a while for competition to push prices down.
onabreak has quit [Ping timeout: 260 seconds]
thrmo has joined #bitcoin-wizards
AaronvanW has joined #bitcoin-wizards
jtimon has joined #bitcoin-wizards
thrmo has quit [Ping timeout: 268 seconds]
pro has joined #bitcoin-wizards
onabreak has joined #bitcoin-wizards
coredump_ has joined #bitcoin-wizards
Pr0t3us has quit [Quit: Leaving]
tromp has joined #bitcoin-wizards
tromp has quit [Ping timeout: 260 seconds]
Guyver2 has joined #bitcoin-wizards
<stevenroose>
I don't know how fee estimation currently works, but it's not working very well. It's not working well with price boosts and not working well with tx/mempool bursts
<stevenroose>
A day of overflowing mempools and blocks full of high fee txs (whether or not artificially created by miners) causes everyone to overpay for a whole week
gazab has quit [Quit: WeeChat 1.7]
gazab has joined #bitcoin-wizards
coredump_ has quit [Ping timeout: 246 seconds]
<kanzure>
are there good numbers about "overpay for a week"?
Dyaheon has quit [Ping timeout: 260 seconds]
Dyaheon has joined #bitcoin-wizards
<stevenroose>
kanzure, I don't have any
<mryandao>
maybe its worthwhile visiting the idea of including a TTL for bitcoin transactions?
<mryandao>
so I can just commit however much txfee, and by TTL if it has not yet been confirmed, at least I know its been dropped from mempool and I can try again with a new fee
priidu has joined #bitcoin-wizards
<stevenroose>
even though that's not how a ttl should work, it is not possible using scripting
<stevenroose>
recently bumped into that. it's possible to set a date from which a tx can be accepted, but not after which it cannot be accepted
<stevenroose>
so you could not use a checklocktime-like thing to make your tx invalid after a certain block
<stevenroose>
(even though that's ofc not how you would like to do a TTL :D)
<mryandao>
does op_csv work with unix-timestamps?
<bsm117532>
I'm a bit afraid of a run on the bank: fee event where market participants decide fees are too high and that causing them to transfer their assets to another coin/class. We're approaching the point were fees are exceeding reasonable transaction fees and spreads on other assets like stocks.
<kanzure>
"reasonable transaction fees" ?
<bsm117532>
On a dollar basis, when tx fees exceed that of other asset classes...
<kanzure>
nah how about something that is an internal defintion only, like looking at small outputs that are becoming unspendable or something
<bsm117532>
Speculators don't give a shit about that. ;-)
tromp has joined #bitcoin-wizards
tromp has quit [Ping timeout: 260 seconds]
<mryandao>
does bitcoin give a shit about speculators?
tromp has joined #bitcoin-wizards
<bsm117532>
The price certainly does. And as a matter of system engineering, banks are engineered to handle excess load (e.g. via FDIC insurance) but we have nothing to buffer fees in a sell-off event. They'll spike like crazy. (regardless of the cause of the sell-off event)
<bsm117532>
And the fact that fees are unrelated to the magnitude of value transfer, only the highest BTC holders will be able to get their funds transferred. Others will have to wait for the fee to fall.
<instagibbs>
bsm117532, you can't exit if you can't move your coins ;)
<Taek>
instagibbs: playing the long game
<Taek>
adam3us: I've been thinking about fee markets and hysteresis a lot the past few weeks, and there are I think a few things worth noting
<Taek>
0-1 confirmations takes a fee of 300 sats, yet there are plenty of transactions above 540 sats. 200 sats gets you in within 8 blocks
<Taek>
that's a huge spread
<Taek>
Miners can pretty easily screw with this just by refusing to mine low value transactions every once in a while
<Taek>
The way fee estimation currently works, if you want to get into the next block, you'll way over-estimate if a miner is maliciously trying to push up fees
<Taek>
second thing, estimation doesn't consider the current tpool at all
<Taek>
and I think that's super valuable information, because ultimately it's the 1mb of highest fees that's going to make it into the next block, modulo secret deals and shenanigans
<Taek>
So I'm going to say we've got a 'loose' fee market - people are often way overpaying to get into blocks, where if they were willing to take a 25% risk of missing their 1 block target, they'd be able to cut fees by 1/3 or more
<Taek>
And I'd also posit that a loose fee market results in much worse inflation, because people see that it takes 500 sats to get into a block (that's what bitcoinfees.21.co is recommending), and so that's where they set their mental bar for what is reasonable
<kanzure>
Taek: you could ask users to draw a probability density curve for how many confirmations they want how quickly :P
<instagibbs>
Taek, I'm kidding of course. Aggressive RBF will become the norm in the future.
<Taek>
I mean, users are going to draw curves that simply aren't realistic. You can't guarantee yourself getting into the next block no matter what you pay - if the next block the miner just decides 'empty' is the good thing to do, there's nothing you can do about it
<kanzure>
oh you want the wallet to do this by default..?
<Taek>
yeah I think so
<Taek>
most users just put the fee that their wallet tells them
<Taek>
and the wallet operators generally don't like receiving emails about failed confirmations, so there's strong bias to overpay
<Taek>
the strong overpay bias of course results in a fast upward sprial, where I think a lot of it is probably not needed
<Taek>
Hysteresis: The current system looks backwards at previous blocks with a half life of 350 blocks? Something like that
<Taek>
A halflife of several days is going to mean that there's no response at all to things like reduced pressure at night
<instagibbs>
Taek, 3 time horizons to choose from
<instagibbs>
with 3 different halflifes
<Taek>
ah ok
<instagibbs>
(for 0.15)
<Taek>
user chooses? I still think that's really not the way to go
<Taek>
because ultimately, there's exactly one fee value that will get you in for fastest and cheapest
<instagibbs>
getting off-topic, but no theres a raw API and a "smart" one
<Taek>
ah ok
<Taek>
not off topic I think, how users interact with the system is absolutely going to drive the fee rates
<Taek>
We haven't had a chance to delpoy it yet, but with Sia we're experimenting with a transaction pool based fee system
<Taek>
basically you adjust the recommended fee based on what's in your transaction pool
<Taek>
the blocks are used as a fallback to make sure you aren't grossly off-target
<instagibbs>
seems sane for picking a range, RBF often will be needed due to bursty blocks
<Taek>
intending to aggressively use rbf is probably a good strategy to keep fees low
<Taek>
if you know the mempool is empty, you can try a low fee. If you see it starting to fill up, you can go ahead and rbf it up a bit
<Taek>
What you really need is to ask users what fee they are comfortable paying, and then let the wallet play whatever games it can to get a fee that's as low as possible
<Taek>
and then the wallet knows when it's hit a fee that it can't increase anymore, and the user has been priced out
<Taek>
also, something you may not be aware, but there are definitely users in the wild who set fees with the intention of waiting 3 days or more for their first confirmation
<mryandao>
why not have a user commit to a fee, and introduce a "limited lifetime" for 0-conf tx?
<Taek>
I know a guy who typically uses 20 sats as his fee, and he usually gets in within a week
<mryandao>
then when said validity time expires, a user can know that all mempools will have discard said tx and they'll try again with a different fee
<instagibbs>
mryandao, bramc has proposed a mechanism like that, but it's not reorg safe
<instagibbs>
a reorg might permanently invalidate a tx chain, which in general is frowned upon as a design decision in Bitcoin
<Taek>
it's also more complicated than you want a typical user interacting with
<Taek>
they just want to know that their transaction is going to go through, and in some (strict or casual) timeframe
<instagibbs>
fwiw it may make perfect sense in a no-reorg situation like federated chains
Chris_Stewart_5 has joined #bitcoin-wizards
JackH has quit [Ping timeout: 240 seconds]
kexkey has joined #bitcoin-wizards
<Taek>
instagibbs: you can solve the issue with a mandatory timelock
<Taek>
imagine that you set a deadline on a transaction at 6 blocks from now. In doing so, you also get an implied timelock that's 86 blocks forward on that output
<Taek>
this follows the miner rule of needing to wait 80 blocks, that's where I'm pulling the number from
<Taek>
or, I guess it wouldn't need to be 86 exactly, it could just be 80 forward from whatever height the txn gets confirmed
<Taek>
this eliminates the risk of invalidating tx chains
JackH has joined #bitcoin-wizards
thrmo has joined #bitcoin-wizards
danrobinson has joined #bitcoin-wizards
kmels has joined #bitcoin-wizards
tromp has quit [Remote host closed the connection]
tromp has joined #bitcoin-wizards
Aaronvan_ has joined #bitcoin-wizards
<Taek>
adam3us: Displaying users fees in terms of dollars instead of btc might help, but when the btc price goes up people's willingness to pay high fees will also go up because they will feel richer
<Taek>
really, I feel like we should instead build a mechanic around figuring out the max a person is willing to pay, and then otherwise using market mechanics to drive the fee as low as possible
<Taek>
I'm pulling together a more comprehensive proposal, will post in a few
AaronvanW has quit [Ping timeout: 260 seconds]
GoldenAngle has joined #bitcoin-wizards
GoldenAngle has quit [Changing host]
GoldenAngle has joined #bitcoin-wizards
GoldenAngle has joined #bitcoin-wizards
licnep has joined #bitcoin-wizards
BashCo has quit [Ping timeout: 246 seconds]
Dyaheon has quit [Ping timeout: 246 seconds]
Dyaheon has joined #bitcoin-wizards
BashCo has joined #bitcoin-wizards
paveljanik has joined #bitcoin-wizards
mkarrer_ has quit []
tromp has quit [Remote host closed the connection]
abpa has joined #bitcoin-wizards
tromp has joined #bitcoin-wizards
tromp has quit [Remote host closed the connection]