If you've used one of the generic bitcoin profitability calculators, you've probably been surprised by the projected income from Bitcoin mining. The profitability numbers can be shockingly high for just buying a video card, putting it in your PC, and running a program 24/7.

But all of the online calculators I've seen omit the biggest hurdle: Difficulty increases.

The bitcoin network is designed to produce approximately 2016 blocks every two weeks. With high performance hardware becoming cheaper and more widespread and more users jumping on the mining bandwagon, the only way to keep the block production rate constant is to make it harder. Thus, every two weeks the difficulty adjusts to slow down or speed up production to match this 2016 blocks / 2 weeks target rate. This almost always goes up.

In fact, it goes up at an alarming rate. Sipa maintains some excellent charts of the increasing difficulty over time here:
http://bitcoin.sipa.be/
Notice how difficulty is increasing rapidly. (The link is very flaky, so check back often if it doesn't load right away).

If you're going to make any long-term projections about the profitability of mining hardware, you must take periodic difficulty increases into account. Right now, the next difficulty increase is predicted to be about a 15% jump at the next increase. This means a rig that pulls in 29 BTC per week right now (typical for an OCed 5870 running phoenix) will only be making 25.5 BTC per week. Add another 15% difficulty and that drops to 22.2 BTC per week. Six months from now (at 15%) that drops to 6.3 BTC per week.

On the other hand, the exchange rate has been increasing lately. If you believe this will continue, then the numbers above don't look as bleak. However, it also means that you'd be better off purchasing BTC rather than mining hardware. If the exchange rate really does increase, your BTC will increase in value. Meanwhile, your hardware will only depreciate and produce less BTC over time.

There are some other costs and hurdles that miners tend not to consider when calculating profitability:

1. Downtime - This isn't just the 100% downtime when you're rebooting your computer or updating your miner application. Using your desktop will slow the miner down somewhat. Perhaps a lot if you're gaming or watching a video. Even dedicated mining rigs will have to contend with periodic pool outages or network hangups bringing production down from 100%.

2. Frictional mining costs - Unless you're solo mining (in which case you'll be dealing with variance) you're going to give up some small amount to whatever pool you join. Fees, stale shares, etc. Figure 3% or so of your output.

3. Frictional conversion costs - Unless you're going to be buying things strictly in BTC (not likely for a miner) you're going to have to pull the money out into your native currency. Conversion costs on small amounts will be painful. Conversion costs on larger amounts won't be as bad, but you'll still lose at least 1% and probably more.

4. Cooling costs - Summer is fast approaching in the northern hemisphere. Cooling a house or apartment is difficult in many regions, and even more so with computers pumping heat into your room. If your computer consumes 1000W, then it is ultimately dumping 1000W of extra heat into your room. If you don't think this is significant, go run a 1kW hairdryer in a small space for a few minutes and see how much warmer it gets.

5. PSU Inefficiencies - Any power supply won't be 100% efficient. The only semi-accurate way to gauge power consumption is to use a Kill-a-watt or equivalent at the plug and measure the power directly. If your video cards + mobo + CPU draw 1000 watts from your power supply, it's probably drawing 1200 watts from the wall. Your electricity costs are 20% higher than you expected.

Bitcoin mining is a race to the bottom, and if you don't already own a capable computer then you've probably missed the boat. You will always be up against people with lower (or no) electricity costs, and people who are buying hardware cheaper than yours. If you're counting on the increasing exchange rate to buoy your mining rig profitability, then you really should be investing directly in BTC, not buying mining hardware. Increasing exchange rates will only drive more people out of the woodwork who are willing to mine for meager profits. Even worse, there are plenty of people who don't grasp the concept that electricity and heating aren't free, and will mine for negative profits.

So before you drop cash on a new mining rig, think long and hard about whether or not you'll find it useful for anything other than heating your house in the winter.

I don't know why it says that; I just did a ~$500 transaction through mtgox yesterday and it went through fine.

So it seems like askrate is how long the mining client will wait (in ms) before it requests more work from the server IF the server has not already sent it work? Great, thanks for the help, I'll try it out.

884509627386359275033751967 943067599621731590401694134 434007629683591574337516791 197615733475195375920401694 343151239621353184932676605 800621596380716399501371459 954387507655892533875618750 354029981152863950711207613

Maybe this has been addressed, but I couldn't find it...

Twice now, when I have left  phoenix running connected to slush's pool, I have woken up to find phoenix waiting on workorders from the server. It only drops to 0Mhas/s for maybe a second or less before it gets a new work order then goes back to normal until it finished that work. Restarting phoenix fixes the problem.

Would this be something wrong with the pool's server? Or phoenix? And what can I do to stop it? Obviously I could make sure to restart phoenix before leaving it running for a while but that is more of a bandaid.

Hmm... Sage advice: If you use .bat or .cmd scripts, put the word pause at the bottom, on a line by itself.
That way it waits for a keypress before closing the window on you.
jedi95 should probably put that in the first post.

I got it to work. I forgot to define the device. rofl.

OK, I wrote simple watchdog which will restart bitcoind once it crash again, because I'll be out for following two days. This surely helps and I'll definitely compile patched bitcoind on monday.