Topic: Won't transaction fees become wildly unpredictable as bitcoin grows? (Read 2035 times)

We can't predict what the predictability will be.

I will always include 0 fee transactions in my mining. Many other miners will too. You may have to wait hours or days for your transaction to be included in a block though. If you submit a 0 fee transaction, and it's taking too long to be included in a block, there is nothing stopping you from resending the transaction with a fee to get it pushed through. Then the 0 fee transaction will be rejected and priority given to the transaction with the fee.

You'll notice if your transaction doesn't go through because it will never get any confirmations (aka included in a block).  Transactions only get rebroadcast so long.  You'll basically have to double spend it with a greater fee if you want it included sooner.  I'm sure eventually a zero fee transaction will get included in a block, but it may take a while, and you may need to rebroadcast it several times.

So if miners start requiring a transaction fee for every transaction, and I decide not to to pay a fee when I send coins, will my transaction simply be dropped? Will I be notified if this happens?

The taxi analogy is great.

You just have to remember that with bitcoin, it doesn't really cost you anything to add a transaction to a block. It costs a lot to make the block, but once you found it, it doesn't cost you anything.

Let's say you paid 100 USD in electricity bills to find a block. Yay!
You have 500 txn sitting on your desk. 50 of them are offering a 5% fee (say that's 150 dollars). 100 are offering a 2% (say that's another 50 dollars). 250 are offering a 0.1% fee (and that's another 10 dollars). The rest don't offer any fee

You can now say one of the following:
1) I'll only process txn with 5% and up - you'll make 150 dollars
2) I'll only process txn with 2% and up - you'll make 200 dollars (150+50)
3) I'll process all txn with any % - you'll make 210 dollars (150+50+10)
4) I'll process all txn, with or without a fee - you'll still make 210 dollars

They all cost you the same - 100 dollars for electricity. Which will you pick?

If you want to be a gold taxi, you'll make people want to pay more txn fees, but you'll make less money. Only if you own a big monopoly can you start fixing prices. But if you're trying to make a monopoly on mining - you're fighting Moore's law and they can't make their monopoly too devastating or else people will just start using a bitcoin competitor. Virtual currencies are relatively easier to switch than traditional currencies.

That's why I believe that txn fees will not be unpredictable - they will be totally predictable. They will cost exactly how much electricity costs. If electricity prices rise, so will txn fees rise.

I feel like at some point in the future due to losses and the value that everyone gives to it, BTC can eventually become very pricey for one coin.

Let's say 1 BTC = 1,000$ USD then the .005 BTC transaction that we look at now (0.1175$) would be worth a lot more later (5$). So I would hope that the fees would be reduced. Of course it's just a hypothetical solution.

The example I figure is a great number of people awaiting in a queue their turn to jump in a taxi cab and go from airport to hotel..
The people showing more money can jump the queue.
More people is willing to pay and more is paid in total, more cabs are available and the queue become shorter.
If the normal queue is 1 hour, someone will be interested in pay 1 BTC to have its transaction confirmed immediately (say 5 minutes).
Then, there will be people willing to wait 10 minutes at max that will pay 0.1 BTC for their transaction.
Then the common folk that have no problem with awaiting 1 hour that will pay 0.01 btc
Then the people not paying anything that will wait what it get to be confirmed.

Longer the queue, higher the fee people is willing to pay to have their transactions confirmed in a shorter time

If all the taxis that come along use only slightly different prices there will me some demand for the more expensive ones because you don't know when you'll get another one coming by.

A 5% chance of delayed inclusion must be worth something. Also maybe there are 1% ers with a lower requirement that you satisfy by attacking the 5%ers fee too. I have no idea what the params will be, but it'll probably end up being pretty smooth.

That is very unlikely.  Even if 5% (the greedy miner wanting a % of the gross) excludes the transaction the other 95% wouldn't.  

Nobody given the choice of these two options would pick the latter
fee 0.01 BTC - 95% of miners working.  Estimated average 6 confirm time = 10*6 = 60 minutes
vs
fee 10 BTC - 100% of miners working.  Estimated average 6 confirm time = (10/0.95 + 10*5) = 60.5 minutes.

Even with 20% or 30% of global hashing power holding out for fees like that doesn't make sense for either the sender or the miner.  It is very likely that "exclusive" miner would earn less (and potentially nothing once subsidy is zero) by making unreasonable demands.  There may be variability in pricing but it is going to be bound by common sense.  i.e. one pool takes anything paying 0.001 BTC or more, another requires at least 0.005, etc.  Essentially increased price is based on increased utility and each miner can only increase utility by a very small amount.  People tend not to pay 100x more for something they value as being worth only 5% more.

It would be like WU offering a 15 minutes service (like they do now) and 1 minute service (for 100x higher fees).

Price discrimination is real and emerges in many places. It would not be impossible for a miner with even a teeny bit of pricing power (pool with 5%?) to use a rule like .0001 + .01% based on the reasonable assumptions that large tx are willing to pay more and that they may care more about immediate entry to a block.

But it isn't going to be unpredictable, that doesn't help anyone. There will be software that take into account the details of your tx, the current network load, published miner inclusion rules, etc and make a very good guess about the appropriate fee based on your desires.

A fee of 1-10 BTC?   Doubtful (outside of mistakes).  Bitcoin is more like an ACH in that the amount of the transaction has no relevence on how difficult it is to complete.  A fee in the range of a few cents (2012 USD) is likely sufficient to ensure you are at the very top of the queue even on a 10,000 BTC transaction.  I don't see anything changing that results in Bitcoin needing significantly higher fees and certainly not 1-10 BTC.  If anything fee in nominal BTC are likely to decline.

Right now if there was no subsidy miners could earn that same amount if the average fee per tx was 0.07 BTC.  As transaction volume and the percentage of transactions paying a fee (any fee) increases the average unsubsidized cost will decline.

http://blockchain.info/charts/cost-per-transaction

Another way to look at it is that miners make ~$26M a year in gross mining revenue.  Lets say 10 years from now we would prefer miners make roughly $250M a year (the amount paid to miners will directly correlate to the cost to attack the network).  Currently transaction volume is ~1 tps (transaction per second).  PayPal is ~50 tps.  So lets say Bitcoin grows to be roughly "PayPal Sized" or roughly 1.6 billion transactions per year. If the network was paid 50% by fees (and 50% by block subsidy) then the average tx fee need to be $125M / $1.6B tx = $0.08.  Note I used USD not BTC for calculations because miners (and attackers) costs are mostly in dollars.  The amount of security it provides is closer related to the value of mining in USD terms.

The transaction fee is totally voluntary. Someone could decide to not pay any transaction fee and others could decide to pay a transaction fee as large as they desire.
The penalty for not paying a transaction fee is your transaction will be processed when there is free CPU cycles.
The penalty for paying a small transaction fee is your transaction will be processed with a low priority.
Higher the transaction fee you set, higher the priority the miners will give to process your transaction.

If you do not bother to wait a day or two, you could decide to pay no transaction fee.
If you want your transaction processed immediately (maybe your transaction is of 10.000 BTC) then you pay an high transaction fee (for example 1-10 BTC) and the miners will be glad to process your transaction before any other transaction with a lower fee.

It's a bit more complicated than that because of the exchange rate variance, but I agree that it will be fairly stable.

The block reward only changes every 4 years.  Currently the block reward easily covers mining costs (or miners would stop processing transactions w/o fees).  It will probably still be a few halvings before transaction fees come into play.  Once that is the case, they will probably be pretty steady except for a few months around the block reward decreases.  The larger miners will most likely post their fee requirements - and the smaller miners will follow their lead.  You'll probably be able to pay any fee you'd like, it will just correlate to how quickly it is added to a block.  The smaller the fee, the longer you'll have to wait.  But I think it will be pretty common knowledge what the current going rate is to get your transaction into the next block.  And obviously every 4 years the fees will increase until the block reward is practically none existent - then fees will become even more predictable.

No.

The Bitcoin network is constantly growing. Our current 1 MB block size limit is obviously unsustainable, and will soon be raised and (hopefully) eventually eliminated. Due to the size of the Bitcoin economy in the future, and the volume of transactions, miners will no longer need high transaction fees for every transaction. Transaction fees could eventually decrease to the point where microtransactions become feasible, and miners would still make much more than they do today.

If transaction volume is low, then miners will shut down and difficulty will decrease. If transaction volume is high, then miners will restart and difficulty will increase. Transaction fees may be unpredictable, but they will be both sufficient and manageable.

I was checking out a chart on Blockchain which many of us have seen before. It details the total BTC payouts per day of all the mined blocks.  Unfortunately as a newbie, I'm not allowed to post a direct link yet.

But this chart also brought a little long-term concern for me.  Many slush pools (and indeed practically everyone mining out there supporting the network) is doing so based on the relatively steady payouts from uncovering the initial 21 million bitcoins.  Except, as time goes on and these exponentially decrease, the primary incentive for mining at all will be the transaction fees. 

It seems that transaction fees are much harder to predict though, and are more closely based to the overall activity of the bitcoin economy.  Will this make it exceptionally hard to mine in the future?  Will it become analogous to playing the lottery?