Bitcoin Forum

Bitcoin => Bitcoin Technical Support => Topic started by: mestar on June 03, 2011, 11:50:05 AM



Title: 5.70 fee on a 70 BTC transaction?
Post by: mestar on June 03, 2011, 11:50:05 AM
I want to send 70 bitcoins:

"This transaction is over the size limit. You can still send it for a fee of 5.90,..."

What is that, like 8% fee? 


Title: Re: 5.70 fee on a 70 BTC transaction?
Post by: SomeoneWeird on June 03, 2011, 01:49:52 PM
It goes towards the nodes the are processing the transactions, but I agree, 5.70 for that is a bit over the top. Nothing we can do though.


Title: Re: 5.70 fee on a 70 BTC transaction?
Post by: wumpus on June 03, 2011, 01:55:36 PM
Huh that's crazy. This is worse than paypal  ??? When did the fees change from 0.01 ?

Edit: If it's still 0.01/kB, this means your transaction is 590kB? How could a transaction become that big?



Title: Re: 5.70 fee on a 70 BTC transaction?
Post by: Alex Beckenham on June 03, 2011, 01:57:38 PM
Huh that's crazy. This is worse than paypal  ??? When did the fees change from 0.01 ?

That's per kb. So perhaps he's sending a 570kb transaction, which if it's 70 BTC could mean he's consolidating a truckload of incoming 0.01 transactions.


Title: Re: 5.70 fee on a 70 BTC transaction?
Post by: mestar on June 03, 2011, 02:44:00 PM
This is the transaction http://blockexplorer.com/address/14L8qNvBPBLax1rkDawZBHay6wDvHryeSC

Does anybody know how does the 'suggested' amount for the transaction work?  (Or should I try to find it myself in the source code.) 

It's really unintuitive,  I've tried it a couple of times, each time getting a different suggested fee amount, even for 7 BTC, then I finally managed to get it trough for 0.59, which comes to something close to 0.9%.   Weird. 


Title: Re: 5.70 fee on a 70 BTC transaction?
Post by: barbarousrelic on June 03, 2011, 02:46:47 PM
I was under the impression you could send any size transaction with any size fee or no fee, and it would depend on which miner processes the next block whether it would be confirmed in that block or not. (And that if the next miner didn't accept it, it would then have a chance of being confirmed in the next block, and so on.)

Could someone please correct me?


Title: Re: 5.70 fee on a 70 BTC transaction?
Post by: Gandlaf on June 03, 2011, 02:50:08 PM
This is the transaction http://blockexplorer.com/address/14L8qNvBPBLax1rkDawZBHay6wDvHryeSC

Does anybody know how does the 'suggested' amount for the transaction work?  (Or should I try to find it myself in the source code.) 

It's really unintuitive,  I've tried it a couple of times, each time getting a different suggested fee amount, even for 7 BTC, then I finally managed to get it trough for 0.59, which comes to something close to 0.9%.   Weird. 


Isnīt this a bit of a problem in the long run. I do(more or less) get the technical issues that lead to the differences in transaction fees. But is this really a way to run a monetary system? Shouldnīt transfers of the same amount incur the same fee, regardless of how the bitcoins behind it are structured? Wonīt completely arbitrary fees make the system unusable for commercial transactions(after all the main purpose bitcoin was invented for)?


Title: Re: 5.70 fee on a 70 BTC transaction?
Post by: mestar on June 03, 2011, 02:51:39 PM
I thought the same, but this is not possible with the 3.21 client, and it was the same with the 3.19. 

If you click "No" in the "Would you like to pay this fee?" dialog, the transaction is simply canceled/discarded.


Title: Re: 5.70 fee on a 70 BTC transaction?
Post by: goatpig on June 03, 2011, 02:56:45 PM
I was under the impression you could send any size transaction with any size fee or no fee, and it would depend on which miner processes the next block whether it would be confirmed in that block or not. (And that if the next miner didn't accept it, it would then have a chance of being confirmed in the next block, and so on.)

That assumption is true. It is up to miners to integrate any transactions they wish to the blocks they're attempting to solve. I think that warning is an implemented (but not enforced) guideline about transaction fees in the standard client. He can still try to submit the tx for free, and chances are one of the miners that integrate tx regardless of fee will take care of it(~99% of the network right now).

What this warning shows is how much you'd be expected to "tip" in order to have that tx processed in the late future.

Also, the message gets me thinking the guy is using a 0.005 BTC/kb fee client and that his tx is actually 1,180kb wide, which is why the client is bitching about "size limit", which is currenctly capped at 1mb per block.


Title: Re: 5.70 fee on a 70 BTC transaction?
Post by: Raulo on June 03, 2011, 02:57:50 PM
I want to send 70 bitcoins:

"This transaction is over the size limit. You can still send it for a fee of 5.90,..."

What is that, like 8% fee? 

Next time, do not get 0.1 BTC pool payouts but larger chunks. Everybody now needs to store your 55 kB transaction in their blockchain. Transaction fees are a way to force people to economize their transaction behavior and use larger BTC value transactions.  


Title: Re: 5.70 fee on a 70 BTC transaction?
Post by: barbarousrelic on June 03, 2011, 03:01:13 PM
I was under the impression you could send any size transaction with any size fee or no fee, and it would depend on which miner processes the next block whether it would be confirmed in that block or not. (And that if the next miner didn't accept it, it would then have a chance of being confirmed in the next block, and so on.)

That assumption is true. It is up to miners to integrate any transactions they wish to the blocks they're attempting to solve. I think that warning is an implemented (but not enforced) guideline about transaction fees in the standard client. He can still try to submit the tx for free, and chances are one of the miners that integrate tx regardless of fee will take care of it(~99% of the network right now).

What this warning shows is how much you'd be expected to "tip" in order to have that tx processed in the late future.


A similar thing happened to me before when someone accidentally sent me 100 1-BTC transactions. I tried sending them back all at once, and it seemed to me that it was not possible to send them without the transaction fee. mestar's post above concurs: If you click "No" that you don't want to pay the fee, the transaction is cancelled.

I ended up paying the guy back with two free 50-BTC transactions, but it still seems like something is wrong.

And I agree with Gandlaf - it is very confusing if the transaction fees can't be anticipated by the amount of the transaction. If the user can't easily tell whether he has 100 individual bitcoins or 100000 of .001 bitcoins, it's going to be frustrating and confusing for him to try to send without knowing the number of kilobytes the transaction will be and therefore how much of a fee he should pay. How are merchant-customer agreements on fees going to be set if they are unpredictable?


Title: Re: 5.70 fee on a 70 BTC transaction?
Post by: Dobrodav on June 03, 2011, 03:15:29 PM
That happens if you received your bitcoins in small sums, on lots of different addresses. Solution - its still possible to use 0 fee, or at least real low fee, per Kb. Well, you should use that opportunity and accumulate that bitcents on another own account.


Title: Re: 5.70 fee on a 70 BTC transaction?
Post by: Gandlaf on June 03, 2011, 03:20:38 PM

A similar thing happened to me before when someone accidentally sent me 100 1-BTC transactions. I tried sending them back all at once, and it seemed to me that it was not possible to send them without the transaction fee. mestar's post above concurs: If you click "No" that you don't want to pay the fee, the transaction is cancelled.

I ended up paying the guy back with two free 50-BTC transactions, but it still seems like something is wrong.

And I agree with Gandlaf - it is very confusing if the transaction fees can't be anticipated by the amount of the transaction. If the user can't easily tell whether he has 100 individual bitcoins or 100000 of .001 bitcoins, it's going to be frustrating and confusing for him to try to send without knowing the number of kilobytes the transaction will be and therefore how much of a fee he should pay.

Having thought about it a bit a bit more, I think its more than just confusing. In a certain way its debasement http://en.wikipedia.org/wiki/Debasement (http://en.wikipedia.org/wiki/Debasement) after all if you look at the situation in the OP, mestars 70 BTC are effectively worth only ~64.3 untainted BTC(potentially even less, given that these fees will be incurred for any future transfers as well; Transfer it 10 times at ~8% fees and you end up with ~30BTC). This is how money dies.
Furthermore http://en.wikipedia.org/wiki/Gresham%27s_law (http://en.wikipedia.org/wiki/Gresham%27s_law) says that everyone will hang on to their "good money"(untainted/freshly mined BTC) and will be inclined to get rid of their tainted BTC asap. Given that the "tainted" BTC incur transaction fees precisely because they take a lot of work to process and everyone has an incentive to get rid of their tainted BTC first, this should nicely ratchet up the overall workload.


Title: Re: 5.70 fee on a 70 BTC transaction?
Post by: eturnerx on June 03, 2011, 03:23:57 PM
That is confusing. Is there a way to consolidate bundles of coins to avoid the project with too many tiny bundles of coins?


Title: Re: 5.70 fee on a 70 BTC transaction?
Post by: Garrett Burgwardt on June 03, 2011, 03:34:54 PM

A similar thing happened to me before when someone accidentally sent me 100 1-BTC transactions. I tried sending them back all at once, and it seemed to me that it was not possible to send them without the transaction fee. mestar's post above concurs: If you click "No" that you don't want to pay the fee, the transaction is cancelled.

I ended up paying the guy back with two free 50-BTC transactions, but it still seems like something is wrong.

And I agree with Gandlaf - it is very confusing if the transaction fees can't be anticipated by the amount of the transaction. If the user can't easily tell whether he has 100 individual bitcoins or 100000 of .001 bitcoins, it's going to be frustrating and confusing for him to try to send without knowing the number of kilobytes the transaction will be and therefore how much of a fee he should pay.

Having thought about it a bit a bit more, I think its more than just confusing. In a certain way its debasement http://en.wikipedia.org/wiki/Debasement (http://en.wikipedia.org/wiki/Debasement) after all if you look at the situation in the OP, mestars 70 BTC are effectively worth only ~64.3 untainted BTC(potentially even less, given that these fees will be incurred for any future transfers as well; Transfer it 10 times at ~8% fees and you end up with ~30BTC). This is how money dies.
Furthermore http://en.wikipedia.org/wiki/Gresham%27s_law (http://en.wikipedia.org/wiki/Gresham%27s_law) says that everyone will hang on to their "good money"(untainted/freshly mined BTC) and will be inclined to get rid of their tainted BTC asap. Given that the "tainted" BTC incur transaction fees precisely because they take a lot of work to process and everyone has an incentive to get rid of their tainted BTC first, this should nicely ratchet up the overall workload.

You don't get those fees every time. Just when you make a gigantic transaction, as was done here. This was ~500 .1-.2 inputs, and one output.

In the next client, fees will be more manageable but still, an alternate client that handled fees way better would be pretty cool :)

-Garrett


Title: Re: 5.70 fee on a 70 BTC transaction?
Post by: Gandlaf on June 03, 2011, 03:47:10 PM

You don't get those fees every time. Just when you make a gigantic transaction, as was done here. This was ~500 .1-.2 inputs, and one output.

In the next client, fees will be more manageable but still, an alternate client that handled fees way better would be pretty cool :)

-Garrett

I get that you probably will not have fees quite as high as the 8% in this example at every further step. But the principle still applies: X BTC "made up" of many small amounts will incur higher transfer fees than X "whole" ones. Effectively one composite BTC has less purchasing power than a newly mined one due to transfer fees.
What I wrote before in my last post about debasement and Greshamīs law(the incentive to spend composite coins and hoard whole ones) still seems to apply. I donīt see how you can successfully run a currency if some units with the same notional value have less  purchasing power than others.

What am I missing?


Title: Re: 5.70 fee on a 70 BTC transaction?
Post by: barbarousrelic on June 03, 2011, 03:48:39 PM
Fungibility is a crucial characteristic of any money, and it seems that Bitcoin is not completely fungible. I hope this can be fixed, but I fear it's inherent.


Title: Re: 5.70 fee on a 70 BTC transaction?
Post by: Dude65535 on June 03, 2011, 04:00:45 PM
Fees for a transaction need to be based on the load the put on the network by that transaction since those fees help support the network.

Coins split into very small amounts can regain their full purchasing power if you use small free transactions to combine them into larger chunks. This would take some time though as if you tried to do it quickly the short time between spends would prompt a need for a fee as well.


Title: Re: 5.70 fee on a 70 BTC transaction?
Post by: Alex Beckenham on June 03, 2011, 04:26:00 PM
Effectively one composite BTC has less purchasing power than a newly mined one due to transfer fees.

Just as a comparison, let's say you're a hot-dog vendor and a customer wants to pay you with 570 actual pennies. You'd probably want to ask for a surcharge for the inconvenience.


Title: Re: 5.70 fee on a 70 BTC transaction?
Post by: Gandlaf on June 03, 2011, 04:33:26 PM
Effectively one composite BTC has less purchasing power than a newly mined one due to transfer fees.

Just as a comparison, let's say you're a hot-dog vendor and a customer wants to pay you with 570 actual pennies. You'd probably want to ask for a surcharge for the inconvenience.


Exactly, or maybe I would just refuse to accept them. According to the same logic, freshly mined BTC should command a premium over those of unknown composition!
And thereīs your break in fungibility. Letīs assume I run a shop, what do I do? Raise my prices by x% to compensate for the risk? Refuse to accept certain bitcoins? 


Title: Re: 5.70 fee on a 70 BTC transaction?
Post by: bitoption on June 03, 2011, 04:51:03 PM
Clearly the fee is way, way too high. Can we all agree on that?

I think if it were .057 BTC, or .57BTC nobody would be complaining. (Let's recall that's roughly $5 US today, a lot!)

I also think it's outrageous that the most recent client enforces this behavior -- let people float their transactions out there and wait longer for confirmation if they want to! That's how the fee was intended to function originally if I recall -- a way to incent miners to take it, and speed up confirmation of the transfer.

There should always be the notion of a 'bulk' transfer, 4th class postage which costs very, very little to nothing.


Title: Re: 5.70 fee on a 70 BTC transaction?
Post by: mestar on June 03, 2011, 04:54:59 PM
Exactly, or maybe I would just refuse to accept them. According to the same logic, freshly mined BTC should command a premium over those of unknown composition!
And thereīs your break in fungibility. Letīs assume I run a shop, what do I do? Raise my prices by x% to compensate for the risk? Refuse to accept certain bitcoins? 


I don't think this would be a problem.  Once they are consolidated, your transactions are back to normal size..

For example, the next transaction from my monster transaction was just 258 bytes:

http://blockexplorer.com/tx/597359518fbdcc36fcf6a6b5dedb0196111fbf64992ab8e1237633f9538ceb46#i951162


Title: Re: 5.70 fee on a 70 BTC transaction?
Post by: Dude65535 on June 03, 2011, 05:07:42 PM
Interesting, does this happen organically? Is there some mechanism through which coins are healed or does someone have to actively do it?
If there is such a mechanism, is the whole process(splitting up of bitcoins through transactions and "healing") divergent or convergent? Are more coins(/as many) being "healed" than are split? Or are more coins being split than reassembled?  In other words is the problem self correcting?

Two ways for it to happen you can spend them and pay the fee, they will arrive as a single unit. Alternately you can send small batches to another address you control (say at mybitcoin, mtgox or bitcoin running on another computer you own) and then send in larger chunks from there till you can send them all at once with no fee. You have to wait some time between stages of this process for the transactions to mature to the point where they won't need a fee just because of age.


Title: Re: 5.70 fee on a 70 BTC transaction?
Post by: Alex Beckenham on June 03, 2011, 05:08:27 PM
Letīs assume I run a shop, what do I do? Raise my prices by x% to compensate for the risk? Refuse to accept certain bitcoins?  

If you're running a shop which sell things for 70 BTC for example, then this issue of small change won't affect you directly. It'll affect the customer (who pays the MINER more in fees if using small change). You'll still just get one chunk of 70 BTC. What's your risk?



Title: Re: 5.70 fee on a 70 BTC transaction?
Post by: Gandlaf on June 03, 2011, 05:19:51 PM

Two ways for it to happen you can spend them and pay the fee, they will arrive as a single unit. Alternately you can send small batches to another address you control (say at mybitcoin, mtgox or bitcoin running on another computer you own) and then send in larger chunks from there till you can send them all at once with no fee. You have to wait some time between stages of this process for the transactions to mature to the point where they won't need a fee just because of age.

Thanks to all of you for the explanations. I think I understand: Large fees only occur at the point where lots of small payments are combined, subsequent transactions are not affected.
Sorry if Iīve wasted your time, but thanks for your patience.


Title: Re: 5.70 fee on a 70 BTC transaction?
Post by: tacomaster on June 03, 2011, 11:19:37 PM
I still don't quite understand. I have but .02BTC. an attempt to send .01 of my .02 coins shows the error 'this transaction is over the size limit, you can still send it for a fee of 0.01'. what gives?


Title: Re: 5.70 fee on a 70 BTC transaction?
Post by: bitoption on June 04, 2011, 12:27:20 AM
You need a different client. I would expect this to get amended in the next release; it's broken as-is: that behavior is inappropriate, although agreed on here in the forums.




Title: Re: 5.70 fee on a 70 BTC transaction?
Post by: imperi on June 04, 2011, 08:21:51 AM
Sorry if someone already made this clear -

If I wait over 24 hours, will I not have any fees with future transactions?

I have about 10 received payments, from a mining group, and when I sent 0.05 out to someone it made me give a 0.01 fee.


Title: Re: 5.70 fee on a 70 BTC transaction?
Post by: mapwow on June 09, 2011, 04:33:30 AM
If I wait over 24 hours, will I not have any fees with future transactions?

If your priority is high enough you will not have to pay a transaction fee. Things that increase your priority: 1) more confirmations on the coins you're using 2) larger value of the coins 3) smaller size of the data. Details are available at
https://en.bitcoin.it/wiki/Transaction_fees

You can also look into using a free transaction relay: https://en.bitcoin.it/wiki/Free_transaction_relay_policy