Topic: How to use Segwitaddress.ORG ? (Read 669 times)

That was from https://dedi.jochen-hoenicke.de/queue/more/#2h
It's the second graph down on the page, usually easiest to see on that one.


I'm not quite understanding your question here. Have a look at the site and watch what happens when a block is found and you should have a very good idea what a good fee level is at the moment. If not can you expand on this question a bit and I'll try and help.


When you receive that payment your wallet will have one input. Then if you haven't spent it the next time you get a payment you will then have 2 inputs and they will be listed separately on that tab. Think of it like cash, if someone gives you a 10 cent coin and then you get another and so on until you have 10, you do not have a dollar bill you have ten 10 cent coins. It doesn't matter if they are sent to the same address or a different address in the same wallet, every transaction you receive is a separate input.


There is only one fee. If you preview a transaction on Electrum it will show you a size. This is actually the virtual size as that is all you need to know about and what the fee is calculated on.
If you bump up the fee you are simply replacing the transaction with another one with a higher fee. Normally a wallet will not let you do this but with the Propose Replace-By-Fee feature, you can.


Not too sure what exactly you want to know but try this
https://en.bitcoin.it/wiki/Weight_units

Where can I get this picture? Link please.



That's totally understandable, but then, there were few bands in the image with the exact same size.
However, as I always send 10 - 15 sats/byte less than the exact amount as needed here, how does it guarantee that it will come in that band only?



Still unclear about this. Let's take an example. I currently have nothing in that wallet, I will receive something this weekend in a fixed address (single bech32 adddress only). Does it still divide inputs to more than one even if I use a single address to send coins to just one or more than one addresses?



What exactly happens when I bump it up? Does both the transaction fee sizes, i.e.; the normal fee and the virtual fee both change in sizes?
Can you explain about some % part, as based on what percentage a transaction's weight unit is decided? Where can I find more info about that?

If you watch the bitcoinfees.earn.com site you will see when a block is found the bars for unconfirmed reduce in size. There will be a point on the price scale where that doesn't happen for prices lower, so the lowest priced bar that decreased is the point which you are looking for. I find it easier to see on the dedi.jochen-hoenicke.de site as it is the colour band that gets bitten into on each block. You can put your mouse over the band to highlight the price.

So in this picture of right now:



The 120+ sat per byte band is regularly getting little bits taken out of it.


I'm not suggesting you can get more discount, I was just trying to explain how it works. It is the witness data for each input that is discounted. That's where the saving comes from. Using more inputs makes a larger (therefore more expensive) transaction. The penalty for using lots of inputs is much less using Segwit than it is on a legacy transaction.

Something you can use on Electrum is

View > Show Coins

and a new Coins tab appears. There you will see all your inputs listed. Electrum selects automatically which ones it will use in a transaction but here you have the advanced option to select that for yourself by using right click and spend.


Yes, that was what I was trying to say. I probably confused the issue by getting into too much technical detail.
There is also a bit of confusion caused by different block explorers and software using different terminology. Some will quote it in 'Weighted Units' rather than giving 'virtual size'. All it really means is the transactions are not really smaller but they are treated as being smaller both for fee calculation and for how much data is allowed in a block. So it is cheaper to use and increases the capacity of the network.

You should be seeing that normal transactions are 30% to 40% less bytes than they would be if you were not using Segwit.

Edit. Something else I just remembered. As you are new to Electrum go to Tools > Preferences then under Fees change Propose Replace-By-Fee to Always. Then you can always bump up the fee if you ever get a stuck transaction.

How is it even possible to check what fee/byte transactions are being taken in each block?
I only try to check that what fee level has the least transactions waiting to be confirmed, and send 10 - 15 sats/byte lesser than that amount (as it automatically increases after compression of the transaction). Still, sometimes need to wait longer as it's all mainly about luck nowadays. 



But how will I be able to put more inputs if I have everything stored on a single address?
Does it mean that I will be getting no such "discounts" that you mentioned below?



By these examples, are you trying to say that by using SegWit, I will not be getting anything like reduction in fee/byte in real, but due to it being done through a SegWit address, it will reduce the size of the transaction (which is actually the core purpose if I'm not wrong) and give it a virtual size, and that fee/byte (whatever the amount will be), will be multiplied by virtual size rather than 'rawtx'?

That was just the optimum fee at the time of that transaction. I use https://bitcoinfees.earn.com/ and https://dedi.jochen-hoenicke.de/queue/more/#2h to determine what fee/byte to use. Ignore the recommended fee and look at the data to see which fee level is the lowest that is getting included each time a block is found. Right now bitcoinfees.earn.com recommends 460 sat/byte but I would use 120 sat/byte and it would be confirmed quite quickly.

The saving with using Segwit is not that you can use a lower fee/byte but rather in the way the size of the transaction is calculated. Using that same transaction as an example you can see the btc.com block explorer shows a 'rawtx' size of 194 bytes and a 'virtual size' of 112 bytes. The fee is calculated on the virtual size. The implementation of Segwit replaced the 1Mb block size limit with a 4000 weighted unit block size limit. All data in a transaction is weighted at 4 per byte except witness data which is weighted at 1 per byte. So when you use a Segwit transaction you are receiving a weighting discount on the input scripts. The more inputs to a transaction the more discount. The outputs have no effect.

Edit: Something interesting that demonstrates the point just happened on the blockchain. A block just went through with a size of 1,588,136 which is unusually large.
https://btc.com/00000000000000000033a9758f74dd1c027e96d5be10f2e68dbda63feb9c7a94

Looking at what was in the block there are a large number of transaction like this one, all with 300 Segwit inputs and all going to the same output address.

https://btc.com/1a1997b7a16c3ea2f1dc21b95a8f838efb1146ddab3b4140ae69d7957c69557e

The transaction has a 'rawtx' of 51,497 bytes but a virtual size of only 27,309 bytes giving a saving of 47% to whoever was sending these. Looks like it is HitBTC.

No.
The outputs (destination addresses don't matter) as long as you're sending FROM a Segwit wallet.