Spreadsheet: Invest in BTC or mining hardware? - page 3.

JJG

member

Activity: 70

Merit: 20

Quote from: Enky1974 on May 09, 2011, 10:31:48 AM

Yes you're right, i'm checking using the last few data points of difficulty increase.
this is what i obtain

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.

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.

SgtSpike

legendary

Activity: 1400

Merit: 1005

Quote from: phelix on May 09, 2011, 10:46:56 AM

Quote from: SgtSpike on May 09, 2011, 02:23:25 AM

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 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.

phelix

legendary

Activity: 1708

Merit: 1019

Quote from: SgtSpike on May 09, 2011, 02:23:25 AM

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

Enky1974

sr. member

Activity: 254

Merit: 250

Yes you're right, i'm checking using the last few data points of difficulty increase.
this is what i obtain

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.

JJG

member

Activity: 70

Merit: 20

Quote from: Enky1974 on May 09, 2011, 09:50:52 AM

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.

Enky1974

sr. member

Activity: 254

Merit: 250

Quote from: phelix on May 09, 2011, 09:33:45 AM

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

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

phelix

legendary

Activity: 1708

Merit: 1019

Quote from: Enky1974 on May 09, 2011, 04:19:44 AM

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.

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

JJG

member

Activity: 70

Merit: 20

Quote from: Enky1974 on May 09, 2011, 04:28:32 AM

Quote from: SgtSpike on May 09, 2011, 04:26:40 AM

Quote from: Enky1974 on May 09, 2011, 04:19:44 AM

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.

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.

Enky1974

sr. member

Activity: 254

Merit: 250

Quote from: SgtSpike on May 09, 2011, 04:26:40 AM

Quote from: Enky1974 on May 09, 2011, 04:19:44 AM

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.

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)

SgtSpike

legendary

Activity: 1400

Merit: 1005

Quote from: Enky1974 on May 09, 2011, 04:19:44 AM

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.

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.

Enky1974

sr. member

Activity: 254

Merit: 250

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.

SgtSpike

legendary

Activity: 1400

Merit: 1005

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).

SgtSpike

legendary

Activity: 1400

Merit: 1005

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.

JJG

member

Activity: 70

Merit: 20

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!

SgtSpike

legendary

Activity: 1400

Merit: 1005

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.

phelix

legendary

Activity: 1708

Merit: 1019

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.

SgtSpike

legendary

Activity: 1400

Merit: 1005

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.

SgtSpike

legendary

Activity: 1400

Merit: 1005

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...

trentzb

sr. member

Activity: 406

Merit: 251

Ahh, I got it now. Thanks!

SgtSpike

legendary

Activity: 1400

Merit: 1005

Quote from: trentzb on May 07, 2011, 10:14:31 PM

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.

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.

Topic: Spreadsheet: Invest in BTC or mining hardware? - page 3. (Read 58708 times)