When to sell a miner - evaluation of a miner's projected return

[Altoidnerd]

New miners who are confused on network parameters and how to tell if a miner is a good buy, or when to sell it, or just need some information, I wrote an article about how to calculate the expected return of a miner and checked it over with some other experts.  

http://altoidnerd.wordpress.com/2013/11/07/when-should-i-sell-my-bitcoin-mining-hardware-bitcoin-mining-hardware-resale-value-vs-projected-return/

It should be educational, as long as I hope it doesn't discourage you from mining!  I still mine myself, because getting involved with bitcoin in all ways is best.  

Cheers

[mvidetto]

I'd have to agree with most of the article however he uses 14 days for difficulty changes, when in reality when difficulty is skyrocketing its much less like 10-12 days.  This will definitely throw of the amount of btc production of the miners throughout the lifetime.

[sushi]

At least 1 guy may got an okay deal on KnCMiner Jupiter!

[KGBSlim]

I have a Jupiter to be delivered this month, hopefully I break even at least.

[wmcleod]

I'm currently holding 2 Jupiters.  Trying to sell one to hedge myself but it's hard finding buyers who will buy with BTC (or at least a portion of it).  Even harder finding face-to-face buyers (posted on Craigslist).

i'm not too concerned selling them though, right now they're turning a big profit, but I think the network hash is set to skyrocket so i want to protect myself against this. 

[KGBSlim]

How much you trying to get for a Jupiter?

[wmcleod]

How much you trying to get for a Jupiter?

have them up for $7300 USD, but not getting many firm offers around what i'm looking for.  lots... and lots of low-ballers.
buyers market right now i think.

[KGBSlim]

Well hold onto it, id be running the shit out of it if I was you. 1THs can bring some coins in for another month easy!

[wmcleod]

Well hold onto it, id be running the shit out of it if I was you. 1THs can bring some coins in for another month easy!

oh it's running.  it would be blasphemous to hold onto it and not use, I think.

[KGBSlim]

Well hold onto it, id be running the shit out of it if I was you. 1THs can bring some coins in for another month easy!

oh it's running.  it would be blasphemous to hold onto it and not use, I think.

LOL I meant forget about selling it! Just wait, make some coins now and when all the let down Hashfast and Cointerra peeps get BFL'd you can bang that off for 7300.00 then!!!

[ScaryHash]

Well hold onto it, id be running the shit out of it if I was you. 1THs can bring some coins in for another month easy!

oh it's running.  it would be blasphemous to hold onto it and not use, I think.

LOL I meant forget about selling it! Just wait, make some coins now and when all the let down Hashfast and Cointerra peeps get BFL'd you can bang that off for 7300.00 then!!!

BFL'd ... now that's verb.

BFLed = to be lied to, cheated, duped, and no product delivered for one year, while using your money for personal gain.

That's like, worse than a ponzi scheme. 

[-ck]

BFL'd ... now that's verb.

BFLed = to be lied to, cheated, duped, and no product delivered for one year, while using your money for personal gain.

That's like, worse than a ponzi scheme. 

http://www.urbandictionary.com/define.php?term=BFL

[Badonkadonk]

BFL'd ... now that's verb.

BFLed = to be lied to, cheated, duped, and no product delivered for one year, while using your money for personal gain.

That's like, worse than a ponzi scheme. 

http://www.urbandictionary.com/define.php?term=BFL

oh god i love the interwebz

[ScaryHash]

I had not even seen that one...lol

Pretty funny !

[zvs]

How much you trying to get for a Jupiter?

have them up for $7300 USD, but not getting many firm offers around what i'm looking for.  lots... and lots of low-ballers.
buyers market right now i think.

what exactly are you considering a 'low-baller'?   you're insane if you think ASICs are a 'buyers market'.  How many of the chumps buying ASICs now to mine bitcoins do you think will come out ahead?  esp. those godawful USBs.

now, i don't know exactly what you're considering a lowball offer, but if it's something around $4000-$4500, you should probably rethink your valuation of  that jupiter as it's a bit off

oh, just for the hell of it, i'm not interested in buying a jupiter for bitcoins, but if you live within 500 miles of dallas, austin or within 100 miles of tulsa, memphis & possibly some other areas (there are lots of people I could ask), i'd pay $4500 USD for a jupiter (assuming it's not some lemon and can do at least 500ghash/s).    I'd possibly be able to make that $4500 back (and if I was *really* lucky, maybe a few hundred extra) by using it for a month or 6 weeks or so, then reselling it & consider it a nice break-even investment for a hobby.   and, no, I wouldn't pay any more than $4500... which is why I put $4500 and not $4000.

there's no way I'd make a profit @ $4500 if I kept mining on it until it became dust unless BTC went up significantly (in which case, the better off party would be the one that spent that $4500 on bitcoins instead).

If you look at it from a 'wtf should I have done to make the most money' perspective, the best case scenario in regard to purchasing bitcoin mining material is for the bitcoin price to remain stable.  worst case is for it to go down -- then you lose money, because buying ASICs at the prices they're at now is a bit loony to begin with.  

[Rannasha]

New miners who are confused on network parameters and how to tell if a miner is a good buy, or when to sell it, or just need some information, I wrote an article about how to calculate the expected return of a miner and checked it over with some other experts. 

http://altoidnerd.wordpress.com/2013/11/07/when-should-i-sell-my-bitcoin-mining-hardware-bitcoin-mining-hardware-resale-value-vs-projected-return/

It should be educational, as long as I hope it doesn't discourage you from mining!  I still mine myself, because getting involved with bitcoin in all ways is best. 

Cheers

There are some errors in the calculation in that blog.

First, the daily mining income is computed by taking ones fraction of the network hashrate compared to the total hashrate and multiplying that with the amount of coins generated per day. However, for this last term, the target-rate of 1 block per 10 minutes is used. In the scenario of exponentially increasing difficulty (which is assumed later on), a rate of (1 + D) blocks per 10 minutes should be assumed (where D is the fractional difficulty increase per adjustment).

Since the global network hashrate is an estimate at best, a more accurate way to estimate personal mining income is to look at the average time to find a block for given hashrate and difficulty, which is:
time = difficulty * 2^32 / hashrate (source: https://en.bitcoin.it/wiki/Difficulty)
The amount of coins mined per day at current difficulty is then:
coins/day = 25 * 24 * 3600 * hashrate / (difficulty * 2^32)

Next, the length of each difficulty-cycle depends on the size of the adjustment. If the difficulty goes up with 20%, that means blocks were found 20% too quickly for the last 2016 blocks, so those 2016 blocks were mined in 14 / 1.2 = 11.67 days. This changes the outcome of the summation of the geometric series.

If we take X as the amount of coins mined per day at current difficulty and D the fractional difficulty increase per cycle (so 20% -> D = 0.2), we see that the length of each cycle will be 14 / (1 + D) days. During the first cycle, we mine X * 14 / (1 + D) coins. Each next cycle, this goes down by a factor 1 / (1 + D), so we obtain the sum (n from 0 to infinity):

Sum( X * 14 / (1 + D) * (1 / (1 + D))^n )
= X * 14 / (1 + D) * Sum( (1 / (1 + D))^n )
= X * 14 / (1 + D) * [ 1 / (1 - 1 / (1 + D)) ] (using the sum formula for a geometric series)
= X * 14 / (1 + D) * [ 1 / (D / (1 + D) ) ]
= X * 14 / (1 + D) * (1 + D) / D
= X * 14 / D

So when you count in the fact that difficulty adjustments come faster with higher difficulty, the expression for the total amount mined becomes quite simple: 14 / D times the daily income. So with an increase of 25% per adjustment, a miner will at most mine 56 times what it mines today in its lifetime.

There are obvious flaws with the current model of infinite exponential growth, but for the short term it's actually a pretty good approximation.

[Sythyn]

I have a Jupiter to be delivered this month, hopefully I break even at least.

I though KNCminer has already delivered all the jupiter pre-orders?

[KGBSlim]

October preorders were all delivered....As far as I know shipping hasnt started for November orders, it was delayed 5-10 days which was originally supposed to start on the 15th of November.

[Altoidnerd]

New miners who are confused on network parameters and how to tell if a miner is a good buy, or when to sell it, or just need some information, I wrote an article about how to calculate the expected return of a miner and checked it over with some other experts. 

http://altoidnerd.wordpress.com/2013/11/07/when-should-i-sell-my-bitcoin-mining-hardware-bitcoin-mining-hardware-resale-value-vs-projected-return/

It should be educational, as long as I hope it doesn't discourage you from mining!  I still mine myself, because getting involved with bitcoin in all ways is best. 

Cheers

There are some errors in the calculation in that blog.

First, the daily mining income is computed by taking ones fraction of the network hashrate compared to the total hashrate and multiplying that with the amount of coins generated per day. However, for this last term, the target-rate of 1 block per 10 minutes is used. In the scenario of exponentially increasing difficulty (which is assumed later on), a rate of (1 + D) blocks per 10 minutes should be assumed (where D is the fractional difficulty increase per adjustment).

Since the global network hashrate is an estimate at best, a more accurate way to estimate personal mining income is to look at the average time to find a block for given hashrate and difficulty, which is:
time = difficulty * 2^32 / hashrate (source: https://en.bitcoin.it/wiki/Difficulty)
The amount of coins mined per day at current difficulty is then:
coins/day = 25 * 24 * 3600 * hashrate / (difficulty * 2^32)

Next, the length of each difficulty-cycle depends on the size of the adjustment. If the difficulty goes up with 20%, that means blocks were found 20% too quickly for the last 2016 blocks, so those 2016 blocks were mined in 14 / 1.2 = 11.67 days. This changes the outcome of the summation of the geometric series.

If we take X as the amount of coins mined per day at current difficulty and D the fractional difficulty increase per cycle (so 20% -> D = 0.2), we see that the length of each cycle will be 14 / (1 + D) days. During the first cycle, we mine X * 14 / (1 + D) coins. Each next cycle, this goes down by a factor 1 / (1 + D), so we obtain the sum (n from 0 to infinity):

Sum( X * 14 / (1 + D) * (1 / (1 + D))^n )
= X * 14 / (1 + D) * Sum( (1 / (1 + D))^n )
= X * 14 / (1 + D) * [ 1 / (1 - 1 / (1 + D)) ] (using the sum formula for a geometric series)
= X * 14 / (1 + D) * [ 1 / (D / (1 + D) ) ]
= X * 14 / (1 + D) * (1 + D) / D
= X * 14 / D

So when you count in the fact that difficulty adjustments come faster with higher difficulty, the expression for the total amount mined becomes quite simple: 14 / D times the daily income. So with an increase of 25% per adjustment, a miner will at most mine 56 times what it mines today in its lifetime.

There are obvious flaws with the current model of infinite exponential growth, but for the short term it's actually a pretty good approximation.

Was I close?  I (should have) admitted the whole thing is an approximation i empirically found to be correct.