Topic: [Aug 2022] Mempool empty! Use this opportunity to Consolidate your small inputs! - page 41. (Read 89029 times)

If you are one entity and need that many transactions, it may make more sense to combine the transactions into one... also called batching. Don't do 50 or 100 separate single transactions, do 1 transaction with 50 to 100 inputs/outputs as needed, wait a few minutes to accumulate a bunch, ask the user if they are willing to wait to pay a lower fee.

Depends on who the culprit is and how much he pays for it.
If it's Binance and they pay their usual 60-100sat/b you know they will most likely fill the next block and you're left behind if it's one of those tumblers that have appeared lately with their 2-3sat/b then they will be at the bottom till the mempool clears. If they pay the average, then yeah, they will have them distributed one by one as they find space in each block.

We just had one of those spikes, we had an average of 60 sat/vB in the block after it but in the last hour we had 13 blocks, so the mempool is empty with the fee ~3 sat/vB in the last one. To be honest, at these fees it doesn't make much sense to spend that much time on researching unless you have a lot of dust to deal with!  What's a few extra cents, you're losing more if the price goes down by 1% while you do all that math.

Correct me if I'm wrong, but in the case of a so-called burst of transactions, they are not divided evenly between blocks? I don't remember who exactly published it, but a couple of months ago I saw a test here on the forum that simulated 300 thousand transactions, which were ultimately evenly distributed between three or four blocks, (I confess I don't remember what the essence of the test was, but I remember that it was somehow connected with testing the size of the mempool).

I disagree. Most people uses the mempool data wrongly, for eg. reading the chart after successive blocks gives a fee rate that is far lower. You can't predict the future, but it doesn't help if the user don't know what to do with the data. Simply telling them to aim for the top 1vMB of the chart with their fee rates is insufficient.

If you know what you're doing, sure go with the mempool charts. If you don't, then it would often be a better idea to just follow the floating fees given by your client. Fees estimator, or estimation algorithms in general uses the data from the mempool, grouping them into buckets for analysis and showing an estimate to the user. It is really not as inaccurate as people think, though admittedly the way it works makes it less responsive than looking at the mempool yourself. At the start, Core's estimation wasn't great but I've seen quite an improvement over the past few versions (actually got better after 0.15 IIRC). It doesn't reflect the state of the mempool in real time, that isn't what floating fees should do.

I've abandoned that for mempool.space and one of the things that made me do to were exactly the fluctuations, it shows way better the incoming transactions for the last two hours if there is no huge spike in them and they are all around the median capacity you know that there is the risk of binance or some other big exchange dumping half a block worth of transactions just about now. On the other hand, if you see that there have been two or three of those spikes reaching 7-10k vb/s  you know that at least for a few hours there is little chance of another set of waves hitting the mempool.

But that is just for the mempool stats, their estimator is also pretty inaccurate, it takes into account the median fee for the next block, and since right now at block 689888 somebody has sent a 2875 sat/b the estimate is complete bs.

Sometimes that is true, like these days that the bitcoin market is "calm". Other times it is not possible to predict, for example when the price is making sudden moves. Or even in some cases major altcoin pump and dumps require bitcoin moving to and from altcoin exchanges creating a lot of spikes that can not be predicted.

I couldn't agree more. That is why visual representation of mempool such as https://jochen-hoenicke.de/queue/ are always the best for fee estimation. You can see the mempool size, the fluctuations and in times of fee growth you can even estimate the rate at which fee is rising.

I think, generally speaking, the number of transactions and their fees can be easily predicted if you know the amount of time since the last block was mined. I think it is fairly unusual to see large spikes in the number of transactions.

This would result in fee estimators consistently underestimating the cost to get a transaction confirmed. If a block was found 2 seconds ago, the fee estimator will show a minimum required fee for a next block confirmation that, if used, may result in a user waiting a day for confirmation. On average, half of all transactions will be broadcast 5 minutes prior to the next block is found, and the mempool will generally change during this time.

Fee estimators could be improved if users could see, and adjust assumptions made by the estimator.

That is the "human behavior" that we cannot predict not the time between blocks. Even if blocks were mined at fixed intervals (like every 60 seconds) we still wouldn't have been able to predict whether in the next seconds there will be a surge of new transactions paying higher fees or not.

Fee estimators should work on size not count and shouldn't try to predict the future.
Basically they should tell you that with current mempool situation when the next block is mined it will clear ~4MB from mempool and will contain down to x sat/vbyte fee and anything lower will remain unconfirmed. That way it won't matter if someone was paying very high fees for 100 transactions.

It is difficult for fee estimators to accurately predict the minimum required fee because the time until the next block is always unknown. If a block was just found, the mempool will have fewer high-paying transactions, but it is reasonable to expect that high-paying transactions will appear in the mempool over time until the next block is found. This makes estimating the required fee difficult because you don't know when the next block will be found, and in cases when a lot of mining capacity is being taken on/offline, the fee estimator may not even know how long until the expected time of the next block.

A lot of posts in this thread, and in posts in other, similar threads talk about the required fee to get a single transaction confirmed, but this fee rate would probably not be appropriate if you need 50, 100, or more transactions confirmed in nearly every block, as many larger bitcoin businesses may need. A large business who pays the minimum required fee might get a portion of their transactions confirmed quickly, but would quickly see their transactions pile up as being unconfirmed, which will cause problems. I think a business that consistently needs many transactions confirmed in every block would need to pay something closer to the median fee rate. This has the potential to cause confusion to some of the fee estimators out there.

Going a bit OT from the original point of the thread so we might want to continue the difficulty discussion here:
https://bitcointalksearch.org/topic/--5307087

But I did want to point out that as of NOW 8:30 PM EDT 5-July there are 3 blocks of transactions in the mempool the 1st block would require a fee of 70+ sat/vb to get into it. the 2nd one you can into for 4 sat/vb and there have been no blocks for the last 30 minutes.

There are some really messed up fee estimators out there. And some places just paying a fixed fee, which may or may not be messing up the fee estimators.

-Dave

I don't think the Chinese government getting miners to stop mining in China achieves anything of substance. If the miners are allowed to move their equipment overseas, the Chinese government would be hurting their own economy, and bitcoin would still be making it more difficult for the CCP to control it's population.

Information does not flow freely out of (or within) China. So if the Chinese government is seizing miners, there is a good chance this will not be publicly known. The cost to maintain a 51% attack is pretty low if you exclude the cost of the equipment. The cost of a 51% would be the cost to run the equipment, which is, with the most advanced ASICs is a fraction of the expected value of the bitcoin that would be mined absent a 51% attack.

We also had the effect of having 28% of all mined blocks have no transactions via the recent 28% drop in difficulty, and there were not major issues with transaction fees rising, or the mempool getting clogged.

Speaking abstractly, I can say with confidence that naked statistics do not reflect the real state of affairs, not to mention the fact that, earlier on my own experience, I was repeatedly convinced of this (we are talking about statistics). Continuing the thought in an abstract spirit, I can also invite all of us to think logically ... A 51% attack is the collapse of an entire industry, it is not even hundreds of billions of dollars, the success of such an attack would mean a trillion dollar infrastructure collapse. I mean, the sector leaders have definitely discussed such a theoretical scenario behind closed doors, and I think these 5-6 people have a plan of action in such a case, (I am not at all referring to shorting Coinbase stocks as a hedging tool, I mean an emergency consolidation or redistribution of m'power).

If we start reading reliable news coming from miners directly that someone (be it the government or anyone else), is starting to confiscate huge numbers of mining units, we could be closer to such a scenario. But it's still an expensive attack to maintain even if they gathered the required amount of hardware.

I think the Chinese government has reached its goal. Chinese miners have been shut down, they have seen what will happen to them, and their mining operations in China are now a thing of the past. With them out of the way, the communist party can now switch their attention to the digital yuan.

Not realistic but far more possible than it was two months ago.
We were at 170Exa, you needed 1.6 million S19 pro which nobody could manufacture with all this chip shortage, now we're at ~90 Exa with at least 80 Exa of that gear already laying around and wondering where to go, a lot of it for sale, you suddenly don't need a million and a half you need a hundred thousands.
Besides, the whole 51% is just like the killer blow, somebody mining empty blocks with 40%, and it's enough to make the network unusable...in normal conditions...which we're not having right now.

Looking at the fees and the mempool, nothing is normal, normally we would see low fees in the first two mornings of the week and that's it, now we're coming after a 27% drop in capacity for over two weeks, we're still getting 11% slower blocks two days after the adjustment and yet the mempool is completely empty.
And despite this, somebody is having fun sending 300sat/vb tx....

I don't think the difficulty dropping by this much, or that this much mining capacity going offline is "good news".

My assumption is that the majority (nearly all) of the drop in hashrate is due to Chinese miners going offline and that nearly all Chinese miners have already been taken offline. The CCP (Chinese government) does not like bitcoin because it makes it more difficult for them to control their population, and they have an incentive to harm bitcoin.

There is the risk of a government-sponsored 51% attack. This could mean that you cannot get your transactions confirmed no matter the fee you try to pay, or it could mean that you are forced to pay very high fees that make bitcoin undesirable to use.

I like that part of mining leaving China as well. What I don't like is a +50% less secure bitcoin network precisely for the reasons LoyceV mentioned in his post above. Although I don't find the threat of a 51% attack that realistic because it's almost impossible that all the missing hashing power could end up in the hands of the same individual or a malicious group who are going to use it for such a purpose. It's even less realistic that those who are still providing their computational power to the network do the same thing. A combination of 1 and 2 makes more sense, but I doubt it.     

Sure, but that is going to take time. It's not a one-click process. If the miners decide to sell their equipment, it's gonna take weeks to pack up, ship, and connect and configure at the new destination. If they decide to leave China to mine somewhere abroad, it might take even longer. Finding a new location, ensuring you understand and comply to local laws and regulations, finding an electricity provider, filling out the paperwork, setting up your farm...

Allow me to quote myself:

---

---

Fees are as low as they get now: 1 sat/vbyte. Use it

I don't find it scary at all, it's good news all around. There's now far less of a concentration of miners in China. The difficulty has dropped making mining more profitable for everyone else. The mempool didn't get excessively congested while it happened so fees weren't too bad. Eventually, all those mining rigs will find their way out of China and get redeployed somewhere else.

Looking at a chart showing the total network's hash rate is pretty scary. Looks like an altcoin after a huge sell-off. From around 150 TH/s at the beginning of June, the hash rate dropped by almost 50% in July. But it seems to be taking a turn for the better now after reaching a bottom at around 84 TH/s if the data I am looking at is correct. 


https://www.blockchain.com/charts/hash-rate   

Mempool is empty now. This may or may not last, depending on how will the hash rate "evolve".


(source: mempool.space)