Topic: Price or difficulty ? (Read 1175 times)

The "marginal miner" being forced out is a different discussion than the "broken" transaciton fee issue.  In time FPGA will displace GPUs.  It will take time but it will happen.  GPU are 2MH/W to 3 MH/W.  Underclocked and undervolted running optimal rig you are still <5MH/W.  45nm FPGA are 20MH/W.  By 2013 28nm FPGA products will be out getting 40MH/W (thats a GH/s on 25W).  If someday someone made an sASIC @ 28nm you could expect >100MH/W.  With custom ASIC 200MH/W+ isn't unrealistic.  40GH/s per chip @ $80 and 20W is possible @ 28nm.  From there it will only improve with Moore's law.   GPU mining will end just like CPU mining has all but ended.  I "think" we are still many years away.

The transaction issue is different one.  Imagine there is no subsidy all you get paid on is transaction fees.  Would you exclude a paying transction?  You are already solving the block.  The amount of extra work to include another transaction is negigble.  In other words including 100K transactions is only marginally more difficulty than including one.  If you are doing the heavy work of solving a block (who's difficulty is unrelated to # of transactions) you might as well include all paying transactions.  Excluding a free transaction is a simple choice and in time most economical miners will do that.  Excluding a paying transaction never makes sense.  If you exclude it when you solve the block you get paid less.  Someone else incldues it and when they solve a block they get paid more.  No matter how "cheap" the transaction fee is 0.01 BTC, 0.000001 BTC excluding it simply means doing the same amount of work for less revenue.

Current transaction fee rules allow a fee as low as 1 satoshi (if no anti-spam rule applies).  Given that a user can expect most/all of hashing power to include their transaction in a block for 1 satoshi so why would they pay more?  A miner can expect other miners to include a paying transaction if they don't so why would they exclude it? In time you can expect (non-spam prevention) transaction fees to settle at 1 satoshi each.  Which puts revenue at a negligble amount and will make almost all mining uneconomical.  

This can partially be solved IMHO by having a minimum fee.  Free transactions are still possible but if you include a fee it must be at least x.

Do you have any thoughts on how this will affect gpu miners who actually have to pay for electricity?
Fpga miners will naturally continue mining i guess.

As a gpu miner who does Not pay for elect i am mining just as long as it is worth my time. For example i just RMAd a card
Which cost me 15 dollars for a processing fee as well as shipping, crappy sapphire 5830 company, and if this happens more as time goes on with shrinking profits i will say it is time to quit. Now add in costs for elect, if i was paying it, and i could see many people quitting.

And as the other person mentioned i do pay attention to diff.... But if one waits too long older radeon cards loose their value.
Eventually one needs to decide...

You are missing a big part of the decision equation. What will other miners do, what will the difficulty be? If the btc/usd stays the same and 60% of hashing power drops out you'll make more money.

The question you should always ask yourself is "Am I going to be more or less efficient than the marginal miner?"

If you pay a reasonable price for electricity and value the time it takes you to manage your operation you should probably stop at some point unless you want bitcoins and have trouble buying them.

The "forced transaction fee is simply an anti-spam measure.  As block rewards decline miners (or more specifically) will start to decide that transactions they will include.  For example pool xyz could announce they won't include any free transactions.  If other pools follow suit then transaction fees will increase.  There is a problem in that the network/protocol allows 1 satoshi transaction fees.  Most likely no pool will reject a paying transaction and there is no reason for sender to pay more that will keep transaction fees unsustainably low.

The good news is we have plenty of time to establish a functional "fee economy" as hashing power (and thus difficulty) is likely 2x to 3x higher than it needs to be relative to the economic value it protects.  Bitcoin right now is like a private army with 10,000 troops, state of the art equipment, gunships and tanks to protect a baseball card collection.

Has anyone figured out how much trans fees will make up when the reward drops to 25? 1-4 btc would be my guess i suppose. Looks bad for miners cause i do not see the btc price much above what it is now. I need to check in on this to determine if i should sell my stuff.

I just looked at btcguild found blocks and it is pretty darn low. .05 if lucky per block.

I don't see the BTC/USD going up 6X without the difficulty going up less than 2X. 

Though with it being about 10 months until block 210,000 when the block reward will drop to 25 BTC if there were a spike I could see why miners would be hesitant to see if the level sticks before investing in additional capacity.  I still would say difficulty first though.

Difficulty, because difficulty didn't have time to adjust to price last time.  This time, we'll approach 30 more slowly.

to me, looks like difficulty will get there first