Bitcoin Forum

This has been making me curious.  

With the block reward halving that makes most of Europe, much of America (Cali, NY) unprofitable for BTC mining with GPUs.  I assume GPUs are a majority of the BTC mining base.

But from looking at http://bitcoin.sipa.be/ I would say we are statistically about even with the normal swings we saw before.

So are people just mining anyway just to see if BTC prices go up?  At what point might we see a decrease in network hashrate?

Conspiracy theory on the "now normal" wild network hash swings:
Someone has ASICs running and keeps adding more as they notice GPUs exiting the market to keep overall network hashrate on par...
(I know that's been pretty much debunk'd but still I can't help but wonder...)

Do we need a poll?  Are you:
A Still mining
B Partially quit
C Turned off rigs

Based on what I see today seems like A would be the response...

Baffling, isn't it?

I'm suspecting inertia.  When the next power bill comes in and the bitcoins earned don't cover it there'll be some GPUs going dark.

The network hash rate is twice what it was in the $5 BTCUSD stable period, but prices are 2.4 times what they were then. I suspect for many its still profitable, not to buy new cards but to keep mining with old cards. As for me, I mine with my single card not for profit but just to participate in the mining.

Regarding Litecoin - same situation. Payout from Litecoin now roughly equals payout from Bitcoin.

After Bitcoin reward halved, Litecoin difficulty increased sharply (almost doubled), seems more than 1,000 GPU moved from Bitcoin mining to Litecoin mining...

Blocks   210543
Total BTC   10.514M
 
Difficulty   3438909
Estimated   3584782 in 1137 blks
 
Network total   26.699 Thash/s
Blocks/hour   6.51 / 553 s


Still looking at an increase in difficulty, almost halfway through the 2012 block period.
The BTC price is hovering between $12.48 and $12.59. Not much of an increase in price -- it was $12.30 before the Halving.

Seems that for each person turning off his rigs (few in number), there is twice as much hashing power being switched on somewhere else...

Very strange. ASICs?

There was a huge spike in hashrate just before the block halving as people tried to get the "last block" and the "first block" and then that stopped. Since then, the hashrate is pretty close to what it was before the spike. What's actually happening is the vast majority of miners are still hanging in there. It will just take time for the reality of mining at a loss to get through to them. There are an awful lot of miners that are just hopeful the value will go up, and there are still a heap of miners who honestly don't even know the reward has halved. It will just take time for the new steady state to be reached, but by then ASICs will have honest and truly hit, so there may well be no significant drop in hashrate at any stage. Don't underestimate the power of inertia on human nature.

tl;dr The network hashrate is an estimate. The longer the time period over which this estimate is taken, the more accurate the estimate is. Daily estimates are subject to significant possible error.

One thing everyone seems to forget is that the network hashrate is an estimate, and that with such an estimate are sometimes quite large confidence intervals. In this case the confidence intervals indicate the possible values the actual hashrate could be based on the measures hashrate. There's not much point looking at changes in the hashrate over a day unless they're quite large changes.

To give you an idea of how large the confidence intervals (and possible error) can be, I've posted a couple of charts. The shaded areas are the 95% confidence interval for the actual (unknowable) network hashrate. It not that obvious (given the differing scales) but the possible estimate error is much smaller in the first chart (weekly average hashrate) than in the second chart (average hashrate per 144 rounds).

Although the actual number of weekly rounds are used for the first chart, here's a general idea of how the estimate error decreases:

The 95% CI  for 144 rounds: 84.3% to 117% of the estimate
The 95% CI  for 1008 rounds: 93.9% to 106.3% of the estimate

From http://organofcorti.blogspot.com/2012/12/2nd-december-weekly-pool-and-network.html (http://organofcorti.blogspot.com/2012/12/2nd-december-weekly-pool-and-network.html)

http://2.bp.blogspot.com/-50-NOGueUqo/ULr4618d7BI/AAAAAAAAEBo/f9NHnk566L8/s1600/2.nethr.02122012.png

http://1.bp.blogspot.com/-5nCrDcyvgj8/ULr4_c51bZI/AAAAAAAAEB4/qZh-Jd_-pYE/s1600/4.hashratePerBlock.02122012.png

Lots of testimonials of people shutting down and selling GPUs but the evidence is lacking so far.
Maybe people believe in bitcoins a little too much. Willing to mine for a loss on price optimism.

Edit:
Ok now that I see the charts looks like a few pools have had some recent drop offs in hash rate.
I guess it will take a few more blocks to see any fall out.

If someone is optimistic about the future price of BTC and is mining at a loss, they should turn off their rigs and buy BTC instead.

I think many people are overestimating the number of GPU miners that became unprofitable due to the block halving.  A dual 5970 rig is still profitable up until $0.15/kWh.  It is winter for most miners, so the heat generated is a benefit and not a cost driver.

organofcorti --

Impressive charts and all, but one can't deny that difficulty is set to rise 4.86% as of the current estimate right now. And we're talking halfway to a reset, which is a pretty accurate reading.
You can't write off a 5% rise in difficulty averaged over 7 days to "variance".

That configuration ranks up there as far as being on the high end of Mhash/J.  A whole lot of hardware is probably at a third less.   And few care to mine just to break even -- unless the waste heat serves a purpose as you described.

 - http://www.bitcoinx.com/profit/

The number of people whose marginal electric rate is $0.07 per kWh or less is a pretty small number.

I just don't get those guys mining at 0 profit, it's way better to sell the cards and buy coins with the money if they believe so much in bitcoin price.

I just shut down all my rigs and cleaned up each piece of each machine. Took about 3 days!

I have just 2 cards running now for a total of 770 MH/s.

I don't want to run these clean cards, because getting them all dusty again for $1 or $2 a day just isn't worth my time. Maybe if I were 16 again, sure. Maybe if it weren't 80 degrees in the part of Texas I live, sure.

But $1 or $2 a day isn't worth it.

And when you factor in the difficulty going up even before ASICs hit the market, I feel like I'm just bowing to the inevitable. I'm just getting out about 3-6 weeks before I'm forced to.

I'm not worried about Bitcoin and the network. It will do just fine without me.

I might rebuild an efficient rig or two for Winter heat -- but Winter hasn't started yet. Winter in this part of Texas lasts about 4 weeks, and we might not get one this year  ;D

The hashrate is actually increasing, the difficulty is estimated to increase by about 5-6% at the next retarget. Though that estimate seems to be decreasing... People probably haven't turned off yet because they haven't gotten their power bills, or maybe most of the network doesn't pay for their electricity, or maybe they believe in Bitcoin and don't want to leave it just because they wouldn't be profiting.

I can.

The 95% confidence interval for 1008 blocks is 0.94 to 1.063. I think you'll find a 5% increase is contained within those limits. Even the difficulty change can be affected by variance - the 95 % confidence interval for the hashrate mean over 2016 blocks is 0.957 to 1.044.

 A difficulty change mostly reflects a change in network hashrate, but if the change is in the 95% confidence interval, then we can't be certain that any of that change is due to a change in network hashrate.

I agree with your mathematics and reasoning, but still...
With variance like that, who needs people adding more capacity to the network?

Sorry, I don't follow. Did you mean "who knows if more people are adding capacity to the network"?

I had about 3 GPU rigs running at one point, then it became unprofitable for me, but I still had one rig running in the garage. It went on for months and months, and then one day it just died. And that was the end of the mining for me. Have been thinking about getting some new gear, but at the moment I haven't ordered anything.

Since there's been less than 140 blocks produced in the last 24h, I would say that the hashrate is going down.

I was paraphrasing the saying, "With friends like this, who needs enemies?"

In other words, a 5% diff increase is pretty significant, especially when margins are razor thin. It doesn't really matter WHY it happened, even if it's raw chance.
If variance can jack up difficulty by 5%, it doesn't matter that no one is adding capacity.

I find it odd that variance isn't just-so-happening to bring the total hashrate DOWN, even though several people have stated their rigs are now silent.

I personally took 5 GH/s off the network several days ago, and others have done the same.

Ok. I think you were misunderstanding me. When the 95% confidence interval is included, if difficulty goes up by say 4%, this could be the result of:

* no change in network hashrate, and all due to variance (network was "lucky")
* a 4% increase in hashrate and average luck
* an 8% increase with really poor luck.
* a 0.6% decrease in hashrate with really good luck.

All it means is that we can't know if the network is stable, decreasing a little, increasing or really increasing hashrate. The network hashrate is just not a very accurate indicator of what miners are doing in the short term.

Perhaps an analogy might help. When mining on vardiff at say 15SPM the variance over a day is low, but per hour your hashrate bounces all over the place since you aren't sending in solutions very often. The network is receiving solutions at the rate of 0.1 per minute, so the hashrate is going to have even more variance.

It evens out in the long run, it's just more of an issue for miners now that most are mining on very thin margins.

I'm curious if any one knows how long it will take to show a hash rate drop if indeed there is one.

I think there are some people legitimately shutting down but I think there's *just* been a lot of talk, perhaps to convince others to do the same?
I think there's a bit of a game of chicken going on now.
If others shut down I won't have (Halve) to...
Also you could buy BTC OR you could make them at the cost of electricity.

Maybe it is a good thing people are so optimistic about BTC prices that they are willing to mine at a loss?
But adding hardware to network at a loss? that's just crazy.

That's quite a good question. Basically, if the change is greater than your upper or less than your lower confidence interval bounds, then you can say the network hashrate has changed by at least the amount it exceeds the upper or lower bounds, with such - and - such confidence.Your choice of confidence interval directly determines how large the the upper and lower bounds are, and how long it takes to reduce them.

I'll post an example when I can.

i don't think the hashrate will ever drop.

people still want a piece of the pie -

however small that is.

they are all chasing a possibility, a probability -

a chance


for the same reason people have mined gold at a loss -

for centuries.

the reason people aren't stopping is they suck at math. We had our net income cut in half not profits! So people may have been making 20 dollars a day and now think, I'm only making 10 dollars a day when it reality its more like $5. They also are not figuring in their hardware is losing value about as fast as they are making money. They are also a few people like me who are playing chicken. But at this point I think I'm close to done. Maybe I can beat the fire sales to ebay.

Several miners have used that phrase "playing chicken".

Yes, playing chicken -- with a train.

You have to get off the road when you get close to that train (ASICs) anyhow, so why not get off a month or so early, and reclaim the space (quiet, money, etc.) used by the mining operation?

Yes I agree the train is coming. I guess we will really see how bad people's math is soon enough!

/me can't resist saying
Choo choo

I will react to it when the first ASIC is verified and out in the wild. When and if.

Ente

We'll maybe the traders are really good at math and the price of bitcoin has the reward halving included in it? Maybe our awesome summer profits was a pre block halving bonus that will only happen every four years. And scraping by is the norm.

Or, you know most bitcoiners don't sit and watch their rigs 24/7 So they aren't in a rush to turn them off instantly. The ones switching to alt chains did it before the drop. Most are taking a wait and see and going for what they can get before asics hit.

I think it's hard to let go... It definitely will be for me too when ASIC kill my FPGAs (speculation x12.5 higher difficulty makes them unprofitable and ASIC brings x40 MH/W so I expect atleast x40 difficulty!)

I see a nice drop in the last few days:
http://bitcoin.sipa.be/speed-lin-10k.png

Or is that deepbit being offline? Website is down, no idea about the mining. Anyone can confirm?

Ente

..deepbit mining acting strangely too not completely down but almost.

Anyway I'm sure 95% percent of miners have backup pool so for me this is not the case of dropping hashrate.

I don't know.. The last time deepbit was down was a year ago, one day offline because of DDOS? I don't think many have a backuppool set up. Hell, many miners won't even have noticed yet! :-)
Anyway, the "difficlty" is in free fall, from 4m down to 3m already.. Let's see how this one goes!

Ente

Wow that is a big drop there. Maybe this is the shut off proof that we have been expecting. I'm gonna go check out the price forums to see if supply is adjusting price much. Still lean times for miners for now.

Never ceases to amaze me how many miners set up precisely one pool without a backup, at the highest fee, and virtually never update anything, nor check on whether it's working properly. Mining profitably just isn't like that...  ::)

down to august levels now

http://blockchain.info/charts/hash-rate (http://blockchain.info/charts/hash-rate)

23 T hash... Not bad. Hoping for July levels say 13T hash. I guess the lower it goes the better for the first ASIC guys!

looks like it's going back up.

i guess i was right.



you can't stop the hash!



i'm doing it at a loss right now -

and don't care.


it's heating my toes at least.

The price on used GPU's in this forum should be a good measure of the perception of future difficulty / price levels,

The market for used GPU's are existing miners and new entry miners.

So either people are believing that price is going to the 20-25$ range soon, or they are expecting a crash that will make difficuty go down to a level that compensate for the block reward halving.

If I was serious about mining now I would buy all FPGA's I could get my hands on, as they are priced favourable now MH/$ vise. That would still be a gamble that asic's will not take over (either because of delays or a smaller market than people think).

So everybody are waiting a little longer and running their GPU's while the price is still rising. If we see a serious price drop, people will shut their rigs off and their will be a fire sale of GPU's

Sigh.

Variance.

Spend some time looking at a network hashrate chart on a log scale and you'll figure it out. Or just read the posts of mine above.

7xxx cards are still profitable as they use less wattage per hash, it's the 5xxx and 6xxx cards that have a razor thin profit margin, if any at all.
plenty of 7970's making money though.

I wish the difficulty would drop quicker, since that would squeeze more out of FPGAs before ASICs...

Looks like a temporary weekend upsurge in total network hashrate.  More likely, than not, it'll go back to upper teens by Monday.

I don't think the estimated network hashrate is what you think it is.

5xxx and 6xxx are not profitable at all
7970 is razor thin like $0.33c per day right now, that's a long time to get payback on a card.

Finally we see hashrate drop! Now if it could just go down enough for ROI on my FPGAs before ASICs hit that would be very welcome!

Please turn your "environmentally unfriendly" heater GPUs off! ;)

You probably aren't following this thread then:


then following that:

wtf, now it's going back up again?!

EDIT: Ah of course, price went up 1 $... GPUs are back online...

And because of those coming back to BTC mining after giving LTC mining a try.

But yes, with the exchange rate up 10% and difficulty dropped by a small amount ... today it is marginally better than it was two weeks ago.

Though the ASIC onslaught continues unabated.   Not every promised ASIC design needs to ship to finish off GPUs, ... just one actually shipping will do that.

Or even some souped up FPGA monster.  Such as what yohan says they've got.  From another thread:

There seems to be a decent drop in difficulty, I hope it keeps going down to get more GPU profits before the Asics come in!

What makes people think that GPU miners will quit all at once, and no one will replace them? This isn't an orchestrated event. It's thousands of individuals discovering slowly that they might be unprofitable, and others acquiring FPGAs and replacing the computing power. There are also miners who have paid-for electricity. They have no incentive to stop mining now - especially not with all the GPUs from the miners who quit because they do have to pay their electricity bill hitting the markets and decreasing the price of used GPUs.