Trying to calculate profitability - weird outcome

[MiningBuddy]

Looks legit to me.

The question is... How much would you have made if you just bought Bitcoins instead of rig building?

See this spreadsheet, http://forum.bitcoin.org/?topic=7531.0

[wareen]

For example, I am right now duming some money into mining. Will likely get online with 20 servers the next 3 months - each with 2.3 gigahash around.

In 3 months, you serious about this?
You do realize that even with a mere 10% difficulty increase each time (which is more than optimistic) 3 months make up for over 50% of your total profit? With a 20% increase each time, you even lose 75%! Finally, with a 37% increase (which is not that unlikely) you would only earn 5% of somebody who started tomorrow.

No matter how powerful your equipment is - starting with two servers tomorrow and letting them run for 3 months is probably going to earn you the same as if you power up 20 servers in 3 months and let them run for 3 years (and that's with free electricity).
I'd really be interested in how you come up with a profitable business plan for this?

Why don't you take your strategic investment pool and buy Bitcoins instead? Especially if you plan on selling your mined coins for 100 USD and you obviously rely on the long term viability of Bitcoins.

Just askin'

[accod]

How did you calculate the power?  I'm tired so maybe I'm wrong but I think it should be more like $51.5 per 10 days.

[gentakin]

AFAIK, you don't need the hdd, so that's 60€ saved. Just use a flash drive with some linux distro on it. Sure, flash drives allow only a limited amount of write cycles, but OTOH they're cheap.

Don't forget you can sell the hardware when mining is no longer profitable for you.

If you're interested in a more long-term investment, you might want to look into FPGA mining. That will be future-proof even with the "Energiewende" coming closer now and energy costs likely to increase. FPGAs won't give you 1ghash/s unless you're willing to invest more, but they have pretty low energy costs.

[bcpokey]

God why do people make spreadsheets and why do other people take this crap seriously? These projections are totally bogus, we already had a huge ass thread about an "analysis of mining profitability" not 2 days ago, with some super "legit" looking spreadsheet showing why it would be completely unprofitable to mine. The first very first step in the projection was 100% wrong. They projected a 60% difficulty increase for the next step (this one is about 25%) and a 10% increase in coin price (this one was about 120%).

No one knows the future. This is the key component to making every spreadsheet projection completely worthless. A static rise in difficulty, a static coin price, none of these make any sense. The market is especially volatile right now.

I made a spreadsheet with some conservative values showing I could make ~100% profit return from mining hardware starting tomorrow, but I don't even want to post it as a counter argument because it's just going to feed into the idea that spreadsheets mean anything.

[dikidera]

what will be your hash rate?
what gpu are you buying?

since its impossible to get 6990s here in germany (all out of stock), i calculated with 3 5850s, which have 330.000 kHash/s, according to https://en.bitcoin.it/wiki/Mining_hardware_comparison.

combined hashrate would be 1.000.000 kHash/s. (as you can also see in the top-left of the excel sheet)


PS: using 2x6990 instead of 3x5850 wouldn't change the outcome too much.

A 5850 does 270, your calculations are wrong!

[bcpokey]

what will be your hash rate?
what gpu are you buying?

since its impossible to get 6990s here in germany (all out of stock), i calculated with 3 5850s, which have 330.000 kHash/s, according to https://en.bitcoin.it/wiki/Mining_hardware_comparison.

combined hashrate would be 1.000.000 kHash/s. (as you can also see in the top-left of the excel sheet)


PS: using 2x6990 instead of 3x5850 wouldn't change the outcome too much.

A 5850 does 270, your calculations are wrong!

Most 5850s can overclock to get ~325KHash/sec without much difficulty. If you get a particularly sucky one I suppose it's possible to not get any more than what you say.

[Sukrim]

AFAIK, you don't need the hdd, so that's 60€ saved. Just use a flash drive with some linux distro on it. Sure, flash drives allow only a limited amount of write cycles, but OTOH they're cheap.

Don't forget you can sell the hardware when mining is no longer profitable for you.

If you're interested in a more long-term investment, you might want to look into FPGA mining. That will be future-proof even with the "Energiewende" coming closer now and energy costs likely to increase. FPGAs won't give you 1ghash/s unless you're willing to invest more, but they have pretty low energy costs.

If selling hardware nets you less than 6 BTC are worth at the time you sell it (going with the buying 56 BTC vs. mining 50 BTC example here), you still lost.

FPGAs take a LOT of time until they become more profitable than GPUs, thanks to immense initial hardware costs. Think of timeframes of 2-3 years here. They might need less "overhead" (CPU, RAM...) though, so they might even pay off already after 1.5 years.

However this "payoff" calculation does not take into account growth - if you take this into account as well (and believe that difficulty will not settle at current values), getting the highest possible hashrate you can possibly get asap. will bring the best mining results.

[NetTecture]

For example, I am right now duming some money into mining. Will likely get online with 20 servers the next 3 months - each with 2.3 gigahash around.

In 3 months, you serious about this?
You do realize that even with a mere 10% difficulty increase each time (which is more than optimistic) 3 months make up for over 50% of your total profit? With a 20% increase each time, you even lose 75%! Finally, with a 37% increase (which is not that unlikely) you would only earn 5% of somebody who started tomorrow.

No matter how powerful your equipment is - starting with two servers tomorrow and letting them run for 3 months is probably going to earn you the same as if you power up 20 servers in 3 months and let them run for 3 years (and that's with free electricity).
I'd really be interested in how you come up with a profitable business plan for this?

Why don't you take your strategic investment pool and buy Bitcoins instead? Especially if you plan on selling your mined coins for 100 USD and you obviously rely on the long term viability of Bitcoins.

Just askin'

Yes and no.

First, the 3 months if the final stage - first server goes online end of the week, possibly first two.

my strategic pool does not allow speculation - my failure actually IS to relabel this for speculations...
...and use the hardware to run option simulations. You can not imgine how freaking amout nof powers you need for full market volatility analysis over the CME group (that is all 4 exchanges) -these are about 200.000 active instruments to track my backup plan

[bcpokey]

God why do people make spreadsheets and why do other people take this crap seriously? These projections are totally bogus, we already had a huge ass thread about an "analysis of mining profitability" not 2 days ago, with some super "legit" looking spreadsheet showing why it would be completely unprofitable to mine. The first very first step in the projection was 100% wrong. They projected a 60% difficulty increase for the next step (this one is about 25%) and a 10% increase in coin price (this one was about 120%).

No one knows the future. This is the key component to making every spreadsheet projection completely worthless.

I made a spreadsheet with some conservative values showing I could make ~100% profit return from mining hardware starting tomorrow, but I don't even want to post it as a counter argument because it's just going to feed into the idea that spreadsheets mean anything.

My point as well. Which is why you should hedge.

But that does not mean that such spreadsheet projections are worthless. Because it serves to teach you how the different moving parts of a Bitcoin mining/investing venture interact, and what you should pay attention to.

I don't particularly believe in hedging. I feel like you should make a reasoned seasoned analysis for yourself about where you think bitcoin has been and is going and make a bet about what kind of risk you are comfortable with vs projected reward. Hedging can fall out of that analysis (providing a middle path for risk/reward) but it isn't necessarily a superior option.

Why I don't believe in spreadsheets even as a method of teaching is that it creates the illusion of some sound forward looking analysis. To me you can only look backwards at past performance and say "based on those factors and what is happening today, in the future it will probably..."

Or else get a maths PhD and invent some crazy ass formula that you definitely won't fit on an excel spreadsheet to predict market trends. Heh.

[StopLossLOL]

do you mining-guys really expect a price increase of 20+% per 10 days?

20% would bring us to 450usd/btc in 6 months. 30% to 2000 usd/btc.

[bcpokey]

One thing will become readily apparent to anyone who does the kind of back-testing you recommend, and that is there may not be enough data to make reliable projections. But that does not take away from the value of learning how the mining/investing markets behave and how their different parts interact. Such analysis is entirely accessible to someone without a PhD in mathematics.

It works like this: You try to make the best decisions you can based on the information you have. You take action based on those decisions. You experience pain.

That pain is the best teacher, only if it motivates.

It has taught me to hedge.
[/quote]

I agree 100%. The only place where I disagree is the way in which one learns. I believe one learns from experience such as what you mentioned. I do not believe one learns from create simple static models into the future ignoring the related variables that we've seen affect things until we reached the state we are in today. That is all. The experience and pain give you a much better 'feeling' (if you dont have that phd) for where things are going.

For you, this feeling may have taught you hedging. For me it may be different. But neither of these feelings came from a spreadsheet, that's all.


@OPs question: I'd really like you to get one thing out of what I've been saying, aside from the fact that no one knows the future (which is key point #1). Static models make no sense. Obviously a 20% increase in price every 2016 blocks doesn't make sense, but neither does a 30% increase in difficulty.

The two are inexorably linked, often with one following the other (sometimes switching which leads). They are not always proportional. They are not independent entities either. For example a yahoo news article about bitcoins brought in an influx of new people to bitcoins (maybe even you) and thus bitcoin surged in price. This will happen in spurts that are impossible to predict (unless you own yahoo). 2 months from now all these new people may be fed up with how confusing bitcoin is and that it's not a free money machine and instead of a 20% increase we will see a 50% decrease in coin price. Equally possible. You need to model to predict that too.

[StopLossLOL]

i never said that this model was perfect. it's just the best i can do.

to get things forward: how do you decide? just "feeling it" obviously doesn't work.

[MiningBuddy]

do you mining-guys really expect a price increase of 20+% per 10 days?

20% would bring us to 450usd/btc in 6 months. 30% to 2000 usd/btc.

I personally don't think so no.
I think mining will be totally worthless for the average Joe within the next couple of months, there's a lot of big guys with deep wallets and cheap electric setting up shop and expanding at a crazy rate.

For me? Whatever man, as long as I can make my electric bill & cover some of the difference in depreciation value when Ebay becomes flooded with cheap ATI cards I'll be happy. It's been a fun experiment.

[StopLossLOL]

if this trend continued, we would all be multi-millionaires next year.

i think once certain price-barriers are crossed, psychological issues come into play. people will get very hesitant when it comes to paying 1000usd for one bitcoin.

but thank you, chodpaba. this is in fact very interesting.

I think mining will be totally worthless for the average Joe within the next couple of months, there's a lot of big guys with deep wallets and cheap electric setting up shop and expanding at a crazy rate.

i wouldn't have any problems investing 100k in expensive hardware if i could expect a positive outcome. but so far, i'm not convinced at all.

[Findeton]

Looking at that chart, I long for the day we hit 25e6 difficulty (1000$/btc).

[Freakin]

Chodpaba: Where did you get your data for that difficult vs BTC/$ weighted average? 

[iMiner]

Does it look anything like this?

http://forum.bitcoin.org/index.php?topic=7427.msg184765#msg184765

[Freakin]

Awesome thanks.  You should post regular updates of the graph