Bitcoin Forum

I have generated a spreadsheet that should allow user input of all potential variables to consider when deciding whether to invest in mining hardware or in BTC directly.

Play with it, and tell me what you think.  It seems that the wisest choice could be either, and it greatly depends on difficulty levels and expected increase (or decrease) in the price of BTC.

v1.1 - Difficulty should now accurately affect BTC production.
v1.2 - Added in calculations to show profits if production is traded daily.
v1.3 - Added invest/withdrawal/pool fees fields, added uptime percentage field.
v1.31 - Calculation/bug fixes
v1.32 - Calculation/bug fixes

Download the spreadsheet (v 1.32) (http://www.justinbporter.com/bitcoin_mine_or_invest_1.32.xlsx)
If you find any errors in the calculations, PLEASE let me know so I can fix it!

If you found this helpful, please consider donating BTC to 1CjoEtypZhqkBLSdfpnphNYjWyokourkxe

Glancing at your photo, I don't see how that could possibly be accurate unless I am misunderstanding what C9 (Total BTC generated) means and what value you are referring to in C10.

well there is a reason why I'm selling my rigs now. For me the extra little bit is not worth the hassle of managing hardware and keeping em up and what not. I was gonna just sit on the coins for a while anyways so I'm just gonna buy. Been buying at a steady 3.3ish for a while now which was nice.

Another article on a popular site or a media outlet covering bitcoin on news and it will be another 50% difficulty IMO.

Total BTC generated (C9) includes difficulty increases.  By the end of 4 months, at 40% difficulty increase every 2 weeks, you aren't generating very much.

Total value generated is simply what you would have in USD at the end of 4 months if you kept all bitcoins generated.  This also takes in to account the rise (or fall) in value of BTC to make it worth whatever it is at the end of 4 months.  At 10% every week, one BTC would be worth $19.6 at the end of 4 months.

Ahh, I got it now. Thanks!

Another interesting note...

Assuming the difficulty is increased to 154068, as bitcoincharts.com is predicting, it would only take an average rise of 30% difficulty in the future to make buying bitcoins more profitable than mining.  So many of the naysayers are right, provided the difficulty levels continue to rise at 40% rate.

I need to make the difficulty calculations more concrete and accurate though.  Will work on that now...

Fixed the difficulty level issues.  You now have the option to set how many days from now the next difficulty level will change.  Also, difficulty now only changes BTC production at each difficulty level changeover period - so it's no longer divided into daily increments.  The time period until the next difficulty level is also adjusted according to difficulty increase/decrease.  The higher the next difficulty level, the more quickly that difficulty level will be reached.

TL;DR:  Difficulty should be accurate now.

It is quite interesting to see that after four months or so there is not much happening anymore to the amount of coins you own. You just have two different amounts of BTC - and the one bought seems to be always higher than the one mined for current hardware prices.

You could include a "sell immediately" strategy for comparision, though it has much lower performance.

I will put a link to this thread on my site.

Yes, but again, those values shown in the screenshot (and by default in the file) include very high difficulty rates, which may or may not be the case in the future.  It seems that the flip-flop between whether it is smarter to buy or to mine comes down to difficulty rate, and there will eventually be a difficulty level reached when it will be pointless to invest in new hardware, as the only way it would maintain a better profit than investing directly into BTC is if the value of BTC rises very slowly (< 2%/week).

Thanks for the link phelix!  I do have a sell immediately calculation in my latest version, just haven't uploaded it yet.  It's fascinating to see that as another comparison point...  It's not always the worst option.

Very nice work. Thank you for sharing!

Some comments: The next difficulty jump is actually over 41%, not the 20% like the default in this model. This alone makes a massive difference.

Also, you might consider adding some fields (to both sides) for exchange rate fees. For example, the 0.65% Mt. Gox charges going each way, then any fees for moving your local currency into and out of the bitcoin ecosystem.

A field for 'pool fees' would be great for those mining in a pool like deepbit, where 3% of the take goes to the pool operator.

An 'uptime' factor is also a worthy addition, especially for anyone who plans on gaming a few hours per day. Even dedicated systems may only get 98% due to connection issues, pool disruptions, etc.

Again, excellent work!

Cool, thanks for the suggestions JJG!  I think they're all worthy of adding in as extra fields to further increase accuracy of the calculations.

Updated to 1.3... includes all of JJG's suggestions and phelix's suggestion to include calculations for selling immediately (or in this case, daily, since hourly would be a bit insane and impractical.

Next update, I'm going to notate all of the calculations in comments on each calculated field, so that others can more easily double-check my work (and I can double-check it myself while writing out the notations as well).

I recommend 6% for Expected weekly rise (fall) in USD/BTC
this will lead to a 20$ btc value for jan 2012. with 10% i doubt we will reach 70 usd/btc for january 2012.

I'll also recommend to use 8.6% for difficulty increase, it is the historical value so far.

It's a tough call to say where BTC might end up by the end of the year.  But, that's why the fields are changeable.  :)

with those 2 numbers you will see that it's more profitable to mine and hold. with my 5870 i've a 12 month profit of 27000$ (mining) against 17300$ (investing)

What difficulty rise did you use?

Also, I'd strongly suggest that everyone run many different scenarios to see how each decision will fare across several cases. For example, don't forget to run scenarios in which BTC value falls slightly over time as well.

how do you calculate the 6% USD/BTC rise per week and the 8.6% / 2 weeks difficulty increase?

I get quite different values:

usd/btc mtgox
2010-10-08 ~0.1
2011-05-08 ~3.8
~30 weeks
increase factor: 38
--> 1.129^30 ~ 38
--> 12.9% per week

difficulty (after huge gpu jump)
2010-08 ~200
2011-05 ~150000
~39 weeks / 2 = 19.5
increase factor: 750
--> 1.404^19.5 ~ 747
--> 40.4% per two weeks

I considered a target price of 20 for jan 2012, with 6% you have it, 20$/BTC for jan it's expected if prices will continue to rise with an exponential growth. The growth isnt linear, i simulated a cycle with 4 weeks of big expansion and 8 weeks with price retracement as we seen in the last months, i think that the most probable target it's 18-20 for jan 2012.

8.6% it's calculated forecasting an exponential increase of difficulty using historical values as a starting point, we will probably hit 1,400,000 difficulty for jan 2012 with this pace.
Indeed 6% for usd/btc weekly growth factor it's a conservative value, if a real boom kick in then 10% isnt impossible thus 70$/btc for jan 2012:) instead of 20.
cheers

8.6% for your difficulty increase values doesn't even begin to make sense in the current environment. You can't include the pre-GPU era and the pre-BTC/USD-runup era in your forecasts, as it was a different game back then.

The last increase was 30%. This increase is over 42%. The rate does not appear to be slowing down one bit.

If you had used 8.6% in your forecast 1 period ago (before the 30% jump), it would predict difficulty would be jumping up to 17.9% more (for the 2 periods) at this next jump.

Instead, after the 30% and 42% increases it will be up over 82%.

That's 82% actual vs. your 17.9% predicted. And every other increase will be on top of the increases in the near future.

Another example: 4 difficulty increases of your 8.6% yields a total difficulty increase of 39%.   4 difficulty increases at the current 40% rate yields a total difficulty increase of 384%. Even if growth cools off to 8.6% after these increases, that 8.6% is still building on top of that 384%.


In short: Any predictions must use the current difficulty increase rate for the near future, otherwise they are invalid.

Yes you're right, i'm checking using the last few data points of difficulty increase.
this is what i obtain
http://img84.imageshack.us/img84/843/unledhl.png

around 1,7 mln difficulty for november, 6 months from now. i used the "gpu" historical data.
To match this new value, i've to use 18% instead of 8.6% as "Expected average change in difficulty" in the spreadsheet.

I think you forgot to add/subtract some cells in the sums for "total net profit" and "total net profit daily" in version 1.3.

Hmmm... any idea what I am missing?  I double checked it, and I did find one mistake in the net profit if traded daily... I wasn't including the uptime percentage in that calculation!  Other than that though, everything looked good.  To me, at least.

18% is still going to give you way optimistic results, especially in the near term.

You cannot use anything other than 40% if you want realistic results. If you're going to extrapolate that far out and assume that increases will fall in the future to 18%, then you're going to have to modify the spreadsheet to have a 'difficulty increase' that decreases by a % every month.

for short time frames the traded daily value gets much better than the stashed total profit. I think the cost of the cards is missing from the "total net profit daily". in the "total net profit" is the salvage hardware value included?

You're right, thanks!  The latest spreadsheet should fix all of those issues.  Salvage value was not being included on the total net profit calculation, and initial capital expense was not being included in the total net profit if traded daily calculation.

Is it possible to lower the requirements for this spreadsheet ie to work on Excel 2003 without installing compatibility addon?

Here you go:  http://www.justinbporter.com/bitcoin_mine_or_invest_1.32.xls (http://www.justinbporter.com/bitcoin_mine_or_invest_1.32.xls)

It gave a warning about the conditional formatting, so I'm not sure if the red/green highlights will still work, but the numbers themselves should be ok.

Error 404 - Not Found :D

Gah, my bad.  Try it again.

With OpenOffice it works like a charm  :D

I'm refusing to install JRE ;D

Lolwut?  Office requires JRE?

OpenOffice has and does. Although I think there is a way to install it without installing JRE as well if you're willing to unpack the installation files yourself.

Your spreadsheet is very depressing.  It's so easy to forget how fast an exponential growth moves, I guess that's how we got a debt based currency.  In 6 months my 6 Giga hashes will get me 0.08 bitcoins per day!?

What makes no sense is bitcoins need to be $1000 in 6 months for me to continue mining.

Sorry to be the bearer of bad news...  :(

I would advise people against investing in new mining hardware at this point.  It is still very profitable TODAY, but with the exponential growth not expected to decline any time soon, mining is going to have slim margins, at best, in the not-so-distant future.

Thats why I think the best place to profit is with BitCoin Business Davinci =)  BTW this is Brad from email we've been chatting back and forth about silver and bitcoins.

404 errors.

Should be fixed.  I am in the middle of switching web hosts.

Very nice tool, thanks!

cans someone explain the "expected average change in difficulty" and the "next difficulty starts in this many days" fields?

Thanks in advance.  What are realistic numbers to plug in here?

this spreadsheet looks great.

could you add a "traded weekly" and traded bi-weekly" summary, or something similar?

Because daily trading is unlikely unless you have 5 Ghash/s+ capacity I think, the amount of BTC you get is too low and the transaction fees and time required is too high.

Hmmmm, I'll have to think about a good way to do that trueimage.  It'll be complicated to add such a calculation in.  Either extreme is easy to calculate, but when I have to hold the price constant in the calculation for X number of days, well, there's just not an easy way to do it that I can think of.

Hi everyone, this is my first post. I finally did some reading about BitCoin today. I am impressed! This thread was very helpful.
I think rather than buying hardware, or buying BTC directly, I will try selling development work. I didn't see this discussed in this thread, as I take it most of the participants here are miners.

Anyone have any experience doing this? I've previously done dev work and been paid through paypal, but I've been burned badly by that route (as I'm sure you all can imagine).

Wheee! New hobby!

An allegory? The merchants who sold the gold miners the tools they needed to mine the gold often made more money than they who did the actual mining. Except for the few who were first to find gold and subsequently become fabulously wealthy, most gold miners were lucky to break even.

The only issue you'll run in to is there are a LOT of devs here.  But, since Bitcoin is such an up-and-coming thing, there are a lot of opportunities and new businesses springing up as well.  I have experience with being paid bitcoins as a dev, and it works quite well really.

Assuming electricity is free, how much longer do you think mining will be more profitable then trading?  :-\

Depends 100% on what you mean by 'trading'.

Buying/selling bitcoins is crazy volatile right now. I wouldn't touch it unless you really like gambling.

Hi.

Can someone confirm my calculation, that building a new typical rig with 2x6950 (~1,000$), is already less profitable then investing these into BTC (15.5$ ex. rate, default 10% monthly gain - though it pure speculation, I know).

Thanks.

even if its not "profitable", keeping mining is important for BTC safety/operational security.

There's no sure thing in the world of Bitcoin, especially lately.

Honestly, any projections are probably about as good as flipping a coin at the moment. Or perhaps you could just take your money to Vegas?

JJG, that's why I like the mining investments.  It's virtually risk-free, since the ROI is so quick, and even if you don't hit it, the hardware can be resold for nearly as much as the purchase price, especially if you are a frugal shopper.  But yes, it's still hard to say what will happen in the world of bitcoins.  The price could drop below $1.70 tomorrow, which would make my mining unprofitable.  Who knows...

IMO, the 6990's are a terrible investment, unless you just don't have room to have several rigs.  Going with a midrange card, (ideally a 5830/5850/5870, though those are difficult to find for decent prices anymore) will generally give a much better MHash/$$$ rate.  As it is, you are lucky to get 1Mhash/$ with a 6990, but a card like a 6950, which can get 400MH/s with a bios flash, is nearly 2Mhash/$.  Of course, system costs must be taken into account, but with the spendy PSU's required for a 6990, I just don't see the point in spending the coin on them instead of a cheaper alternative that gets more MH/s.

That's not quite true.  The 6990s pay themselves back just as quickly, have a higher production/watts metric and can give you a higher rack density if you're living in a country where rent is comparatively more expensive than electricity ;)

Also, the 6950 that can be converted into 6970s got sold out something like 10 billion years ago when everybody and their grandma heard about how they could save 150$ on their video cards.  The newer 6950s can't be made into 6970s anymore.  If you were to buy new cards, you basically can't find 5870/5970s anymore so you have to move back to either 6870s in some kind of optimized setup or you spend more and get 69-5/7/9-0's.

It seems that today, buying a few 6870's for $179 each gets you the same Mhash/s for less money.  Those 6990's still cost over $700 each. It would be more cost efficient to buy multiple mid-range cards, right? :-\

It depends on how much it costs per gfx slot in a system as well though.  I think the cheapest I could get it was about $100/slot.  So you have to take that in to account as well...

But since 6990's are just about impossible to find for sale now, going with a midrange card is probably your only option anyway.

Eh, lots of factors missing out but decent sheet.

It's the most complete sheet I've seen... what other factors would you like to see?

I would add the fees concerning the initial buy of bitcoins (could be "Fees to invest" but I'm not sure) and then the fees concerning the exchange of currency when you trade multiplied by the number of trades you plan to do.

I don't know if it's clear... I think the actual spreadsheet is for "buy bitcoin at once and sell it at once" but it could be "trade daily (buy/sell) with low risk but multiple time".

Mining with your GPUs is no longer profitable, Instead of buying high price asic hardware, just buy shares from BTC securities like btc-tc.com or bitfunder.com and you will get the same mining profit without the hazzle of hardware maintenance and electric bills

What do you mean my "shares" ?

I don't pay electricity so mining is still profitable for a long time in my case.

With the asics flood fast approaching, soon, day trading btc will be the only way to be profitable.

i don't pay for electricity too - what kind of gpu should I buy?

In today's market investing with bit-coins is better than investing in mining hardware.

even i've free electricity?