I'm interested in mining. Buying feels too risky. What are the downsides of mining if I already own a decent GPU? Is it as easy as just joining a "pool" and leave my computer on overnight, and basically get paid to do it?
I have been mining since Dec 2010. I know what I am talking about, and let me tell you that at this very moment an AMD Radeon HD 7970 graphics card ($400) would mine the equivalent of $400 in 50 days:
600e6 (hash/sec) * 3600 (sec/hour) * 24 (hour/day) * 50 (days) / (2^32 * 7.7e6 (number of hashes to solve a block)) * 25 (bitcoins/block) * 200 (usd/btc) = $392
So it is very profitable to GPU mine right now. In fact, it has never been so profitable since June 2011 (previous Bitcoin bubble). However keep in mind a few caveats: (a) this assumes an exchange rate of 1 BTC = 200 USD which may or may not crash in the near future, and (b) this assumes a difficulty factor of 7.7e6 which is rising quite quickly (see the charts at http://bitcoin.sipa.be/, it rises by about 10-50% every 2 weeks or so, mostly due to ASICs).
Well, no. Mining is as profitable now as it's ever been, since the price is so high. Sure, you won't be mining 1 BTC a day any more, but when 1 BTC is $200 it doesn't matter.
Further, the announcement you linked to (by Butterfly Labs) isn't really all that accurate. They originally promised a ship date of early October. No products have shipped to date, and they've changed the specs and pricing on what they're going to offer a number of times. Some people think they'll never ship anything.
There are, however, other companies who have shipped ASIC mining chips recently, hence the skyrocketing difficulty. At this rate within 6 months it will no longer be profitable to mine bitcoin on GPUs.
In solo mining, chances are that the difficulty will increase faster than your GPU can increase the odds of solving a block. Statistically, it would already take years to find a block and collect just 25 BTC (and by that time the reward may be halved again to just 12.5). But right now, you can join a pool and still get some fractions of a bitcoin over time, and get some profit. It costs me $2.70 to run my GPU for a week and collect 0.05 BTC, or ~$10 at today's rates.
Are you just assuming that, or have you actually measured it? With my modest setup (a single Radeon HD 7850, mining only when I'm at work or asleep) I'm currently earning about 0.3 BTC/month, and the electricity only costs me about $5-$10.
At the current difficulty its hardly worth it. For example i had a Rig with 3x ATI 6870 running 24/7 in 2011 when Bitcons were around 25$. I think it made about 20-30 USD a day in profit. The same Rig would make $12 today by running 24/7 even at the current BC price of 218$. Subtracting electricity (expensive in europe) profit is fairly minimal and you have to keep the beast somewhere which is very noisy and pumping out hot air.
The risk is that you'll pay more in the electricity bill than you'd get for what you mined. You'd be competing against a crowd that is a step ahead (FPGA) and a few early birds two steps ahead (ASIC). So better first check the current calculations.
I checked a few days ago and it seemed that my 2011 MBP would earn me a few cents per day...
Mining means using electricity today and buying hardware yesterday, which will be paid tomorrow to get coins today. If you're trying to "invest" (probably a bad idea) and anticipate a rise then mine. Otherwise its less risky to buy.
If you're interested in it as a transfer technology (A very wise idea, especially for small-ish international xfers) then just buy/transfer/sell, mining is ridiculous.
For a long time coins were being mined and sold at about the cost of electricity for the most efficient miner. It would be interesting today to see what the premium looks like, probably a pretty scary number. Back when the difficulty factor was a large two digit number I mined a hundred or so BTC in software (not GPU). It's a bit harder now.
Mining isn't free. You will need to purchase hardware if you want to mine in a capacity that's actually profitable. If those coins depreciate, then all that mining is for naught, same as if you had bought them with cash.
If you want to get by with the hardware you already own, look into Litecoin mining, then sell those Litecoins for cash or Bitcoin. Mt.Gox will support Litecoin transactions (indeterminate ETA).
Honestly I think the easiest path is to just buy coins and hope they continue to appreciate. I remember not too long ago buying a few when they were in their ~$90s USD, and then regretting buying them at that "high" price. Now it's more than doubled.
Don't listen to these people. If you already have one or multiple decent GPUs you can make money by mining right now. You wont generate many coins (between 0.01 an 0.05 a day) but the price is so high that this is still profitable. Do the calculations though and see how much your electricity costs.
DO NOT buy into the ASIC market right now though. The preorder waits are too long to accurately make ROI calculations, regardless of price. As more ASICs come online the difficulty will rise significantly making it hard to make back the money you spent on your ASIC.
It you are truly interested in mining with your GPU try litecoin instead. There isn't currently an FPGA or ASIC solution so GPUs are still the most efficient mining hardware. At the time of this post they are trading at ~$4.40 a coin.
Electricity is pretty cheap. The kind of gear required to make the electric bill a significant problem would be a bankrupting level of capital expense. I would estimate it would take a pretty large rack and tens of thousands of dollars to consume several kilowatts at which time hundreds of dollar energy bills would be a rounding error.
Also you need to correct for HVAC.
If its winter a KWh of BTC is the same heat as a KWh equiv of natgas, although the natgas is a little cheaper than electrical, by about a factor of 2. So spend $10 on a absolutely roaring system (About a KW continuously 24x7 which you're not going to be doing on non-dedicated hardware) and save about $4 on my gas bill so the NET is an expense of $6.
In the summer you have to correct for the overall system coefficient of performance, and people who have no idea what a COP is usually assume its "1" (legally mandated by local building code to be a minimum of 10 where I live, your experience may vary). So spend $10 on electricity and add an extra buck or so for HVAC cooling capacity.
What it comes down to is whether you will ever break even with what you put into your rig + all the ongoing expenses for what you get out.
> Electricity is pretty cheap. The kind of gear required to make the electric bill a significant problem would be a bankrupting level of capital expense.
Note that the initial values (60,000MHash/sec, and a very tidy profit in 3 months) are appropriate only for one of the specialised hardware options. If you start filling it in with realistic values for a PC with a graphics card (see, e.g., https://en.bitcoin.it/wiki/Mining_hardware_comparison) you'll likely find the calculations running rather less in your favour.
Profiting is also much easier when bitcoins are $188 each than it is when they are $10 each.
