Bitcoin's transaction fee lowered by 4000%

[Debonaire217]

I probably start with a question, have you noticed the decrease in transaction fee of average bitcoin transaction of today?

In the past 2017, the demand for bitcoin is nearly the same for today. And according to this resource, the bitcoin's average fees today is $0.50 dollars in comparison to $20 dollars wayback 2017.

Moreover, the reason behind this decrease and improvement we have is the SegWit and payment Batching. Before, it seems that SegWit isn't providing the best solution to solve the problem of huge transaction fees. But for today, we can see its effect since the transactions of this time is somehow the same with 2017.

In relation to the news concerning bitcoin-related searches, the demand for bitcoin seems to increase significantly over time. I just not sure if SegWit could handle an all time high and huge amount of transaction since we are near to a serious event of bitcoin halving.

There's no doubt that most of the cryptocurrency enthusiast are now suggesting a transaction using SegWit wallet address and since we now have the information regarding its effect on the bitcoin's transaction fees. I don't think we still have a reason to stay with the traditional bitcoin address.

[mocacinno]

Segwit and batching aren't responsible for a 4000% drop, i guess you're just comparing the fees of a large transaction in the middle of the 2017 spam attack (that fueled bidding war for fees) and the fees of a normal transaction right now.

A couple years ago, somebody was just spamming the mempool with thousands of transactions, so in order to have a decent chance of getting your transaction into a block, you had to outbid all those spam transactions... The fees were insane, but the attack only lasted a couple of months (if memory serves me correctly), before and after the attack the fees have always been very reasonable.

[Debonaire217]

20 fee for transaction. Are you sure? I dont remember that

Correction by mocacinno
Basically it somehow depends on the amount of BTC you are transacting with, that's why it could reach up to 20 dollars at that time, plus, if the market is congested, transaction fees could significantly increase. That is why it is advisable that if we transact, we should make sure that the market isn't congested too much because the amount of transaction fee might be high. So, if we are transacting using batch and with the technology that Segwit could offer, our transaction will be more efficient.

[mocacinno]

--snip--

Basically it somehow depends on the amount of BTC you are transacting with, that's why it could reach up to 20 dollars at that time, plus, if the market is congested, transaction fees could significantly increase. That is why it is advisable that if we transact, we should make sure that the market isn't congested too much because the amount of transaction fee might be high. So, if we are transacting using batch and with the technology that Segwit could offer, our transaction will be more efficient.

The fee does not depend on the amount of BTC you're sending. It depends on the size of the transaction.
The size of the transaction depends on the number of inputs, the number of outputs, which kind of addresses were funded by the inputs and which kind of addresses you're funding with the outputs.

A transaction using 1 input with a value of 100 BTC creating one output of 99.9999 BTC will be a lot smaller (thus cheaper) than a transaction using 100 inputs of 0.01 BTC to create 99 outputs of 0.01 BTC (eventough only 1 BTC was transferred in total, vs 100 BTC in the first case). In this example, it doesn't even matter if the second transaction was between segwit wallets, or batched...)

[stompix]

But for today, we can see its effect since the transactions of this time is somehow the same with 2017.

No, the number is not the same a the peak in 2017



https://www.blockchain.com/en/charts/n-transactions?timespan=all&daysAverageString=7

Those 50k transactions mean everything.
You have a maximum capacity that is not used currently, there are still blocks that are not full

617358   1.234.954 bytes
617357      753.964 bytes
617356   1.306.193 bytes
617355      943.467 bytes

When that capacity is reached fees will go up, and it all depends on how much the capacity might be exceded.
If it's only by 1% then people it won't affect the tx too much, just that some will start batching their transactions and stop spending dust, if it goes to 100% then expect an increase to the same levels as 2018. Although that would also come with more LN adoption.

20 fee for transaction. Are you sure? I dont remember that

Yeah, it was even worse for a brief period

[NeuroticFish]

Segwit and batching aren't responsible for a 4000% drop, i guess you're just comparing the fees of a large transaction in the middle of the 2017 spam attack

The 3 causes for the high fees were the spam attack, the constantly big change in price (triggering various bots) and the lack of batching.
I remember that Coinbase and other "friends of Roger Ver" were accused to cause quite a problem (some have telling numbers of 70% or so for the transactions) by refusing to implement batching.. or at least delaying it.

Off-topic #1. BCash narrative is still sometimes based on "cheap fees" exactly because of this.
Off-topic #2. I don't know why bitcoiners still use Coinbase when they were known for favoring BCash and working against Bitcoin.


But I have to agree that excepting the spam attack period the mempool had its cooldown periods (sooner or later).
Even now a big variation in price causes a spike in fees, but it's smaller and it's usually getting absorbed within a day. We'll see what happens in the next FOMO bubble.

[Ucy]

Segwit and batching aren't responsible for a 4000% drop, i guess you're just comparing the fees of a large transaction in the middle of the 2017 spam attack (that fueled bidding war for fees) and the fees of a normal transaction right now.

A couple years ago, somebody was just spamming the mempool with thousands of transactions, so in order to have a decent chance of getting your transaction into a block, you had to outbid all those spam transactions... The fees were insane, but the attack only lasted a couple of months (if memory serves me correctly), before and after the attack the fees have always been very reasonable.

I wonder if the spam attack still happens or is it reduced? Been long fees went up that high.
I wonder exactly how that was fixed(assuming it has been fixed). IP ban? Or through some other means?

[mocacinno]

--snip--
I wonder if the spam attack still happens or is it reduced? Been long fees went up that high.
I wonder exactly how that was fixed(assuming it has been fixed). IP ban? Or through some other means?

Well, it's been a long time since the mempool was flooded in such a dramatic way as it was flooded in 2017-2018 (don't remember the exact time period, just that it's been a couple of years). It's possible that the "spammer" is still creating transactions, but if he does, it's no longer to a degree that causes issues with other people.

There is no real way to ip ban spam transactions. Transactions are broadcasted trough the network, nobody knows who created them. Ip's are not recorded on the blockchain, nor are they transmitted together with a transaction. If you really wanted to ban "spam" transactions, you'd have a hard time, since there is no difference between a spam transaction and a "valid" transaction (spam transactions are valid aswell).

You could ban peers you suspect being the source of spam transactions, but all nodes would have to use the same blacklist (and the spammer can just use a different ip to broadcast his spam, so you'd have to update the blacklist every couple of minutes). If one of the nodes wasn't willing to ban the suspect peer, the suspect would still be able to broadcast to the node that didn't use the blacklist, and this node would still relay the spammers transaction unhindered.
A second thing you could try is to convince miners not to include transactions using unspent outputs funding certain address lists, but nobody knows the complete lists of all addresses belonging to the spammer's wallet, nor would the miners be willing to help you (since they're missing out on income if they do this).

Neither of these options is bulletproof or feasible, and all options would depend on incomplete, centralised "blacklists" that would have to be honored by the complete network (all nodes and all miners), and one option would even cost the miners a substantial amount of income, so it's not likely they would be willing not to include valid transactions that have a higher fee per vbyte just because somebody says it's spam.

If we put aside the assumption most people have about who the "culprit" is, and why he/she spammed the network, only a couple possible explanations could be given as to why the attack stopped:

Either the attack was deliberately, and it stopped because the culprit reached his/her goal, made his/her point or ran out of funds
OR
The attack was unintentional, and the "culprit" found out he was the reason why the mempool was filled to the brim (and the fees were out of proportion) and he/she stopped spamming... Or maybe it was unintentional, and the "culprit" just needed several thousand tx's confirmed, and once this was done he was content.

[Artemis3]

Segwit and batching aren't responsible for a 4000% drop, i guess you're just comparing the fees of a large transaction in the middle of the 2017 spam attack

The 3 causes for the high fees were the spam attack, the constantly big change in price (triggering various bots) and the lack of batching.
I remember that Coinbase and other "friends of Roger Ver" were accused to cause quite a problem (some have telling numbers of 70% or so for the transactions) by refusing to implement batching.. or at least delaying it.

Off-topic #1. BCash narrative is still sometimes based on "cheap fees" exactly because of this.
Off-topic #2. I don't know why bitcoiners still use Coinbase when they were known for favoring BCash and working against Bitcoin.

But I have to agree that excepting the spam attack period the mempool had its cooldown periods (sooner or later).
Even now a big variation in price causes a spike in fees, but it's smaller and it's usually getting absorbed within a day. We'll see what happens in the next FOMO bubble.

Yes i remember it was right after the price peak of 2017, in January 2018 a low fee transaction could easily take weeks or even a whole month, those days were very crazy, who knows if that also fueled the bear sentiment, and sure certain altcoin promoters have used that excuse ever since.

Still, after around March 2018 such a transfer fee crisis has never occurred again. And then came Segwit etc.

As for coinbase i'm sure i said it back then, it was a very American thing. Elsewhere its irrelevant, but Americans love it for whatever reason.


The other problem i see with transaction prices is that by then many wallets had the "auto detect the optimal fee so it takes n block(s) to confirm". Too many wallets have that enabled by default, so it isn't particularly difficult to provoke a chain reaction that feedbacks it self Its like the attackers have a wiling army of bots worsening the problem.

Yes, once upon a time wallets didn't guess anything, the transfer price was always manually set and things worked fine. But now they all want to "confirm within the next block"...

This is not even Bitcoin's fault but most Bitcoin wallets happily adopted the practice, contributing to the problem. This is why i keep telling people to always force the wallet to use 1 sat/B unless its absolutely urgent. If you plan ahead you can do 99% of your transactions without a hurry, relax and check again tomorrow and for sure your coins will be there, no matter if you paid 3¢ or 1$ in transaction fees.

And when the time comes when we can finally have 0.1 sat/B transactions, I'll switch to manually always use that too.

One welcome side effect of the LN adoption is that it should reduce the onchain traffic somewhat. Again, i would leave that only for exceptional situations (like using a credit card for emergencies only).

[kryptqnick]

I've noticed that the transaction fee has been quite low for a year or so, but nothing special about the last couple of days. I am used to seeing that the normal fee is less than a dollar, and priority fee is a bit more than a dollar (when you send 0.02 or so). Yeah, it's a big improvement in comparison with 2017, but it seems to me that it's actually a temporary thing that is explained more by the network not being that busy rather than significant improvements. If the price jumps to $20k again and people get much more active, the same problems will come again... Scalability is still the biggest obstacle to Bitcoin's adoption as money, as for me.

[todiboa]

I probably start with a question, have you noticed the decrease in transaction fee of average bitcoin transaction of today?

In the past 2017, the demand for bitcoin is nearly the same for today. And according to this resource, the bitcoin's average fees today is $0.50 dollars in comparison to $20 dollars wayback 2017.

Moreover, the reason behind this decrease and improvement we have is the SegWit and payment Batching. Before, it seems that SegWit isn't providing the best solution to solve the problem of huge transaction fees. But for today, we can see its effect since the transactions of this time is somehow the same with 2017.

In relation to the news concerning bitcoin-related searches, the demand for bitcoin seems to increase significantly over time. I just not sure if SegWit could handle an all time high and huge amount of transaction since we are near to a serious event of bitcoin halving.

There's no doubt that most of the cryptocurrency enthusiast are now suggesting a transaction using SegWit wallet address and since we now have the information regarding its effect on the bitcoin's transaction fees. I don't think we still have a reason to stay with the traditional bitcoin address.

SegWit really changed Bitcoin for the better. Compared to 2017, Bitcoin has become incredibly profitable in terms of fee

[franky1]

segwit came into effect in summer 2017

the highest fee ever seen came after segwit.. (december 2017)
why
because segwit forced many transaction types to cost 4x usual amounts.
so if someone sending a legacy transaction paid $20 in december 2017.. if the exact same circumstances and demand was to occur in 2016. they would have only paid $5

also the reason for the $20 was due to a exchange price of $20k(winter) compared to $5k(summer)
so again same situation as just described. but taking the exchange price into account when calculating sats per bytes.
if there was same demand.. but in 2016 and without the exchange price.. the most you would have paid was $1.25

yep. the peak price of fee's should have been $1.25. but DUE TO SEGWIT's 4x of fee's and the whole price  rise drama on exchange caused the TEMPORARY peak of a $20 fee.

this is not due to legacy transactions being bad. this aint to do with bitcoin being broke . nothing to do with segwit making things cheaper. but purely due to hype being glorified as how a temporal event being wrote out as a 'average' price.. (facepalm) to then pretend things got better..

no.. after segwit .. things got worse.. then it recovered
too many stupid people think the $20 fee and the $20k exchange rate are normal long term averages to base things off of.. sorry they were not. they were stupidly short term flukes and should not be used to base average expectations off of to then compare.

wise up people. stop jumping on the segwit hype train.

cant beleive people saying todays fee's of $0.50 is ok..
in 2015 people were creating hell when feels were over $0.20
yet strangely.. saying fees of $0.50 are now deemed as cheap

where did the 2015 promise of cheaper sub cent fee's disapear to.
come on 2009-2015 people wanted to pay under $0.01
devs promoted their 'fixes' to achieve that..
yet here we are 2020. being told $0.50 is acceptable.

screw it. lets take a $10k car. from 2015.. price it temporarilly at $200k for just a couple months.. and then start advertising that the car is at a great deal and worthy paying $50k for it..

or wise up and ask why the heck isnt it $5k or atleast the same $10k
soo many people are duped into the stupid advertising ploy without realising the numbers dont actually meet expectations when they really think about it

wise up people

[hatshepsut93]

Segwit and batching aren't responsible for a 4000% drop, i guess you're just comparing the fees of a large transaction in the middle of the 2017 spam attack (that fueled bidding war for fees) and the fees of a normal transaction right now.

A couple years ago, somebody was just spamming the mempool with thousands of transactions, so in order to have a decent chance of getting your transaction into a block, you had to outbid all those spam transactions... The fees were insane, but the attack only lasted a couple of months (if memory serves me correctly), before and after the attack the fees have always been very reasonable.

If I remember correctly, spamming started around the beggining of 2017, as a preparation for SegWit2x, to force community to accepted a big block fork, but at that time the fees were around $1. The rise to $20 and even $50 per tx can't be solely attributed to spam, that would be insanely costly. There's a simple explanation for such prices - it happened at the height of 2017 bubble, so it's reasonable to assume that those were real transactions done by users who wanted to cash out their profits.

There's some truth to OP's statement, as spamming today would be much more costly, but it's misleading to say that $20 fees are gone because of SegWit. I'm sure we will see high fees again during another peak of the market, there's already a strong correletion between price going up and amount of transaction temporarily following it.

[figmentofmyass]

Well, it's been a long time since the mempool was flooded in such a dramatic way as it was flooded in 2017-2018 (don't remember the exact time period, just that it's been a couple of years). It's possible that the "spammer" is still creating transactions, but if he does, it's no longer to a degree that causes issues with other people.

the memorable mempool backlogs in 2016 were likely caused by a sophisticated spam attack originating in august-september 2015. https://bravenewcoin.com/insights/bitcoin-spam-attack-stressed-network-for-at-least-18-months-claims-software-developer

the mid-late 2017 backlogs were less likely to have been "spam" attacks per se. they were highly correlated with rising bitcoin price and heightened exchange activity throughout the year, climaxing in december when the bubble popped. bitmex, coinbase, binance and others were either drastically overpaying for withdrawal transactions to ensure next-block delivery, not batching transactions, or both.

this will happen again during the next bubble, especially since exchanges have done very little to optimize their withdrawal systems. 8 months ago, coinbase's CEO said batched transactions "should be coming out in a few months". *crickets*

everyday, bitmex clogs the network during USA business hours with its unoptimized mass withdrawal. https://twitter.com/ziggamon/status/1134490591264497664

If I remember correctly, spamming started around the beggining of 2017, as a preparation for SegWit2x, to force community to accepted a big block fork, but at that time the fees were around $1.

it all started way earlier than that. in 2015, coinwallet.eu kicked things off with their "stress test" of the network: https://bitcointalk.org/index.php?topic=1094865.0

this was shortly before the release of Bitcoin XT.

[DarkDays]

Didn't most exchanges start using both Segwit and transaction batching within this time? That's gotta account for quite a big drop.

Also, most recipient addresses are Segwit these days.

But then again, trade volume has increased significantly, even since Bitcoin reached its peak back in December 2017.

Is it because most transfers are happening offchain now? e.g. virtual trades on exchanges, rather than direct wallet-to-wallet transfers?

The days of $50 transaction fees are likely long behind us in any case.

[figmentofmyass]

Didn't most exchanges start using both Segwit and transaction batching within this time? That's gotta account for quite a big drop.

the huge drop came after the 2017 bubble popped. huge inflows and outflows from exchanges (at inflated fee rates) ceased after that.

side note: this has a lot do with really bad fee estimation/overpayment by exchanges and wallets. it's not all about transaction volume. people tend to assume there is a linear relationship between transaction volume and fees but that isn't true.

a lot of exchanges (including the biggest ones like coinbase and bitmex) still aren't batching transactions.

[FlightyPouch]

SegWit really changed Bitcoin for the better. Compared to 2017, Bitcoin has become incredibly profitable in terms of fee

Well, in the past there are this jerk/s that spam transactions. At that problem, people are paying large amount of transaction fees so they could be quickly approved. At that time a lot of people are really pissed with the transaction fees and transaction time, I think that is the reason of this drop. It is not that profitable when we talk about fees since there are still a lot of complains regarding to fees especially to exchanges.

[gentlemand]

Let us wait for a major FOMO before deciding what's what. There is at least some batching and no Bitmain spamming but there is a hard limit that can't be gotten past and no sign of the madness that descended during that period. Fees have spiked violently upwards for brief spurts in recent times and that's a possible harbinger of it only being a matter of time before it gets sporty for long periods again.

[omone1]

The high transaction fees in 2017 was occasioned as I read by Mempool attack. So then we had a lot of congested transactions. I suffered this then, couldn't just sell my bitcoin whenever I wanted because of fees and slow confirmation. Some custodian wallets worked fine then but with a fixed fee, I have never been a fan of Custodian wallet because of the risk they pose no key not your money. In 2017, viabtc came to my rescue by using their free service in speeding up my transaction. 

I love the silent truth in Tones post that scaling is working as regards bitcoin but scammers refuse to admit it. Maybe he regards some altcoin founders who always lash out on bitcoin because of high transaction fees and slow confirmation in the past as been scammers.

[NeuroticFish]

Still, after around March 2018 such a transfer fee crisis has never occurred again. And then came Segwit etc.

SegWit is active since 24 August 2017

As for coinbase i'm sure i said it back then, it was a very American thing. Elsewhere its irrelevant, but Americans love it for whatever reason.

Quite bad if they care more about their "american things" than Bitcoin itself.

This is not even Bitcoin's fault but most Bitcoin wallets happily adopted the practice, contributing to the problem. This is why i keep telling people to always force the wallet to use 1 sat/B unless its absolutely urgent. If you plan ahead you can do 99% of your transactions without a hurry, relax and check again tomorrow and for sure your coins will be there, no matter if you paid 3¢ or 1$ in transaction fees.

The problem is usually the trading/arbitrage bots for which the time is indeed critical. They trigger on every significant price change and cause a spike.

The wallets are indeed badly set, some of them even having some automatic big fee, especially services. Of course, they need to make sure the transactions go through in a matter of hours, to keep the service credibility, but the big fees may be also a mix of laziness and bad programming which cause everybody some loses (and bad publicity to Bitcoin too). Just they are used all day long without big effect on the mempool size, so the spike problem is not there.

One welcome side effect of the LN adoption is that it should reduce the onchain traffic somewhat. Again, i would leave that only for exceptional situations (like using a credit card for emergencies only).

I see LN still as a beta. I was over enthusiastic about it when it started getting used, but the time is passing and it's still not final/used at a big scale.
The next version of Electrum having it may (or may not) be quite a leap forward, but we are still far from paying in the supermarket with Bitcoin and LN (which is imho the goal).