Market volume is significantly higher than the rate of coin generation and that trend will only continue the subsidy will be cut by 50% again in 3 years and one would certainly hope year over year market volume will continue to rise.  Miners have very little pricing power today and it will only be reduced in the future.  Difficulty follows price, price doesn't follow difficulty.

BFL is largely to blame with their >1 year "pre-orders".  However the good news is pre-order timeframes are coming down.  

BFL >1 year delivery time
KNC started taking orders in June? and will ship in Oct (last day of Sept is close enough to Oct) = 4 month delivery time
Bitfury started taking orders in July and delivered early Sept = 3 months delivery time
HF started taking orders in Aug and will ship in Nov = <3 months (last batch 1 order was Aug 24)

Second batches should tighten that schedule further. As that shrinks and the network gets larger there becomes more visibility on what difficulty will be the day you get your unit.  The combo working together will be very powerful.  Still I agree on the larger point and the one reason I find mining unattractive at the current time is the largest factor on your return is the actions of others.

Who said anything about 2.5x basic ROI.  In the real world best prices are based on large volume, and lean operations.  Something like buying $200K worth of ASICS and operating at $0.05 per kWh and even THEN realistically shooting for <30% per year return on investment. 

As for people buying ASICMiners today.  They do it because the "day 0" ROI% is still high.   Sure that will rapidly decline to nothing and they probably will never make up the purchase price but the
"day 0" return is still good.  Looking at current difficulty an Eruptor blade makes about 1.1 BTC per month.  People can always delude themselves into thinking future difficulty won't grow or it wont' grow that fast.  However it is another thing to have negligible or negative ROI% on the first day.  When difficulty is 10x higher and an Eruptor Blade makes 0.01 BTC per month after electrical costs, based on "day 0" return and ignoring future difficulty growth I don't expect to see them flying off the shelves.

It is more complex than that.  Lets look forward a little bit.

My guess is that by end of Q1 2014 we are probably looking at 16 Ph/s.  That would be difficulty of 2.2 billion.

At this point ROI will be significantly reduced.  Even if ASIC companies slash prices to the bone electricity will be making up a larger and larger portion of gross revenue.  An Avalon miner at $0.10 per kWh will be spending 80% of gross revenue on electricity.  All of these things will weigh into future sales.  The next 16 PH/s isn't going to be "easy money" even optimistically the ROI% (using the term correctly) will be measured in % PER YEAR.

During that slow climb from difficulty 2.5 billion to difficulty 5.0 billion the more efficient devices will allow miners to keep a greater portion of the gross revenue.

For example at difficulty 5 billion, $0.10 per kWh, and current exchange rate:
Avalon & AsicMiner - negative operating revenue.  would need electricity <$0.056 just to break even ($1 in electricity produces $1 in coins)
BFL (65nm) - roughly break even higher difficulty will require going idle or selling unit to someone with lower costs
KNC - power cost is 36% of gross revenue
Bitfury - power cost is 18% of gross revenue
Other 28nm players - power cost is 17% of gross revenue

Note at this point I don't think it makes sense for anyone to place any NEW order (analyzing existing orders is more complex because a lot depends on early growth, timing, etc).  It is very likely every single company will be forced to cut prices 50% to 75% or more over the next three months to move units.  Since new units aren't going to reach break even in the next 90 days it makes more sense to wait for the inevitable price drops.

https://bitcointalk.org/index.php?topic=281279.0

The simple version.  Sure if you think difficulty will grow exponentially (doubling every 2 months) until we hit 50 billion or so and everyone even a Conterra miner with no AC costs and $0.05 per kWh electricity is mining at double digit losses then sure efficiency doesn't matter.  However the more realistic scenario is that as difficulty piles up and the "day 0" returns drop sales will slow.  ASIC providers can cut the costs of the chips but that only means efficiency becomes even MORE important.  There will be an inflection point where the growth curve flattens.   The laws of physics ensure that will happen otherwise in roughly 4 years mining will use more energy than the entire human race uses for all other purposes combined.

Difficulty and hashrate and perfectly linear.   If network hash rate doubles, then difficulty doubles (eventually), and the GROSS return on x GH/s is cut perfectly in half.  The net return is reduced by more than 50% because the electrical cost remains fixed.

Based on existing (not future) pre-orders we are looking at 6 PH/s min by end of Dec.  If you want to be more conservative you should consider 8 PH/s.  

1 PH/s ~= 140 mil difficulty
1 billion difficulty ~= 7 PH/s

If you want the exact formulas

hashrate = (2^32 * difficulty) / 600
difficulty = (600 * hashrate) / 2^32

Remember the network/protocol has no way on "knowing" the hashrate (and neither do any sites which chart it).  We can only estimate the hashrate based on the frequency that blocks are found.

Updated Cointerra efficiency from 0.7 J/GH at the wall (0.55 J/GH at the chip) to 0.8 J/GH at the wall (0.6 J/GH at the chip) based on this statement.

Updated (lowered) Cointerra efficiency based on this statement.

A couple new questions for HF.

1) If one purchases the expansion/upgrade board will it ship with the original unit on the original order date as a single device with two boards OR will the original order ship (w/ upgraded PSU and cooling) ship on the original order date and the expansion/upgrade board at a later date based on order number?

2) Can the heatsinks on the FET be removed?

3) What is the size of the radiator used in the Babyjet?

One interesting thing that the expansion board shows us is the max cost for MPP.

The MPP is raw chips only not assembled boards, HF hasn't provided details on estimate cost to turn a raw chip into an assembled board.  The expansion board is $1500 and that includes the cost of the chips and some (unknown %) profit margin.  It does provide an upper limit for the MPP.  My guess is that someone (either HF or third party) will be able to provide boards at $500 plus ASIC chip.

It has no effect.

Ouch yeah that will throw the average off.   If we exclude the last 10 days then average fee per block in Sept was 0.28 BTC per block or 1.1% of total miner compensation, not as impressive as 2.5% but still fees as a % of BTC revenue tripled in a year.   One interesting note is excluding those accidental high fee txs, the average fee per tx in BTC terms actually declined.  That makes sense because the min mandatory fee was reduced 80%.  An interesting factoid is that compared to a year ago a higher % of tx are paying a fee which IMHO is a good sign.  So more txs, lower fee per paying tx but greater portion of tx paid some fee.  In USD terms, fees paid to miners rose from $300 per day in Oct 2012 to ~$5000 per day in Sept 2013.

http://blockchain.info/charts/transaction-fees-usd?showDataPoints=false&show_header=true&daysAverageString=7&timespan=&scale=1&address=

Higher tx volume = higher tx per block = higher total tx fees per block.  Even if the fee per tx remains low.  Measuring it in dollars makes no difference.  If fees are 20% of miner total compensation then are 20% of the total compensation in BTC or USD regardless of the exchange rate.

Fees per tx don't have to be high for the total revenue to rise.  As a complete hypothetical in USD terms today miners collect (collectively) ~$180M USD in combined fees and subsidies.  PayPal has a tx volume of ~50 tps and today Bitcoin has a tx volume of ~0.6 tps.  If Bitcoin grew to be as large as PayPal and the average fee per tx was $0.02 that would be $31B a year in miner compensation or >150x higher than today even if the subsidy was zero.


Not possible to say.  How many tx will there be per block 100 years from now?  However if we look forward just 3 years and say IF tx volume keeps growing, and % of txs which are paying grows, and the block subsidy is halved then the % of miner total compensation will rise.  Even if the % of tx that pay a fee doesn't increase AND tx volume doesn't' increase after the subsidy cut rise fees be ~5% of miner total compensation.

In one year the % of miner compensation paid by fees for Bitcoin has increased from 0.3% to 2.5%.
http://blockchain.info/charts/transaction-fees?timespan=1year&showDataPoints=false&daysAverageString=7&show_header=true&scale=1&address=

Oct 2012.  
30 BTC per day in fees. 144 blocks per day.  
Average fees per block 0.21 BTC.  
Subsidy per block 50.
Total miner compensation per block 50.21 BTC
Fees as a share of total compensation: 0.4%

Sept 2013.  
90 BTC per day in fees. 144 blocks per day.
Average fees per block 0.62 BTC.
Subsidy per block 50.
Total miner compensation per block 50.21 BTC
Fees as a share of total compensation: 2.5%

This trend has been continuing for a long time.  If we look back further to say Oct 2011 fees were ~ 0.02 BTC per block or 0.0004% of total miner compensation.

The rise in fees as a share of total compensation is due to three factors
a) avg fee per tx.  while the min fee has gone down by a factor of 100x the % of tx paying a fee (any fee) has increased
b) number of tx per day/block.   ~ 100 per day in 2009, ~ 400 per day in 2010, ~5,000 per day in 2011, ~18,000 per day in 2012, ~45,000 per day in 2013
c) the block subsidy decreasing from 50 BTC to 25 BTC (and will decrease to 12.5 BTC in 2016).

As long as those trends continue it is possible that fees will be more than 20% of miner revenue not in a 100 years but by the time of the next subsidy.  As an example.  TX volume in 2016 is 5x higher than today or 300K tx per day.  Average fee per tx is 0.004 BTC (double today).  Block subsidy is reduced to 12.5 BTC.   That puts fees per block at ~4 BTC, the subsidy is 12.5 BTC for total miner reward of 16.5 BTC.  Fees are 4/16.5 = 24%.

The graph is on money supply not value. ixCoin HAS gone down if you look at it on a long timeframe.  Down by 99% or more.  In the short term pump and dumpers can easily make moves of 100% or more .... until the dump.  None of that has anything to do with value.


The money supply is well known and predictable for both coins and is based on the subsidy algorithm.  ixcoin money supply went up faster (i.e. higher annual inflation) because it was designed to do that.  Still both of those graphs are only an estimate as due to network growth blocks can be found faster than expected over a short period of time.

ixcoin will die.  It has no support, no infrastructure, no services, no merchant adoption.  The supply will go max monetary inflation to 0% inflation overnight.  If it was robust and heavily adopted it might be able to absorb that kind of shock but it isn't so it will do as well as a egg going from 50 mph to 0 instantly when it hits a brick wall.

That is the the kind of thing that would be useful to look into before heavily acquiring the coin.

There won't be a sudden and sharp drop in supply.  The largest nominal drop in daily new generation has already occured.  Ever subsidy cut going forward will have less and less of an impact.  Already today the effect of new coins on trading is minimal.

The space is due to the way some browsers parsing the html code.

I see this as the raw html

Notice the span injected in the middle of the signature.  Not sure how you got part of the signature ended up inside a different div tag but different browsers are going to parse that differently.  Chrome adds no spaces or line breaks but the html "standard" is more like a guideline and different browsers = different outcomes.  Not sure the exact reason the forum's parser interprets a long string that way but as indicated above using code tag is a workaround.

Using code tags:

Produces the following html:

This  & /thread.

Asking a user to come up with a password with sufficient entropy is a challenge.  That is why key stretching should be used in any key derivative function.  Asking the end user to ignore trusted and peer review cryptographic systems and "roll his own" almost always ends in catastrophic failure.

It isn't difficulty to come up with a cryptographic system that you (the creator) can't break.  It is very difficult to come up with a system which remains strong in the face of crypto analysis.

I asked KNC to confirm that the 1.6 J/GH is at the wall but they did not.  Given that the at the wall efficiency will depend on the user supplied power supply efficiency I think the most probable scenario is that the 1.6 J/GH represents the entire system DC wattage.  At the wall would be ~10% lower for the average 80Plus Gold PSU.  Unless KNC states in clear terms the at the wall/plug efficiency we won't know until some units get to users.

Difficulty of 442 Billion.  That seems realistic to you?  Really?

442 billion * 2^32 =  1.89 x10^21 hashes per block.
25 BTC per block = 7.59 x10^10 GH per BTC.

Let be optimistic and assume the average miner has 1 J/GH efficiency and $0.10 per kWh electrical rate.
7.59x10^10 GH/s * 1 J/GH = 7.59 x10 J = 21,039 kWh = $2,103.90 per BTC in electrical cost (plus another up to $1K in cooling cost).
Even if BTC went up to $500 ea would you spend $2,000 in electricity to make a $500 worth of BTC?


It may not, but you just look silly showing "projections" of the network going to >3,000 PH/s.  If it isn't a wise investment ... then don't buy it.

Of course they do BUT there is two reasons why pre-orders (by all companies) will continue:

a) A seller which waits 30-90 days to make the sale when the unit is ready risks having the sales lost to OTHER competitors who offer preorders.  Simple version only so much capital will be thrown into hardware this generation and waiting gives the seller absolutely no advantage.  How much hardware will be sold?  I have no clue but obviously that number is finite (in the near term) and every sale counts towards that total. 

b) you probably won't like the available to ship tomorrow prices.   Say you (personally) owned a bitfury 400 GH/s rig and your estimate (based on difficulty growth) is that it will net 120 BTC over the next year.  Would you sell it for less than 120 BTC?  You might discount it to account for the risk (120 BTC in hand is worth more than projected 120 BTC) but the discount likely will be small.  Maybe you would sell 120 BTC future revenue for 100 BTC today but you certainly wouldn't sell it for 50 BTC or 30 BTC.  The same thing will happen with hardware companies.  They can do their own projection.  So say HF has a 100 Sierras lying around in Dec, and they have some smart people run some projections on future earnings, lets say hypothetically they estimate it will net 200 BTC before falling below break even.  Now they probably aren't going to sell it for 200 BTC.  Buyers can do the same projections and likely will pass so they will need to discount it "some" but the idea that they (or anyone) would discount it so much that it would be massively profitable (i.e. pay 20 BTC, get 200 BTC over next 18 months) is illogical.  They could just use them to mine for themselves.