At what point do you start solo mining?

[crazyates]

At what point is it considered more profitable to solo mine, rather than mine at a pool? 1 GH/s? 10GH/s? 100GH/s?

[the joint]

At what point is it considered more profitable to solo mine, rather than mine at a pool? 1 GH/s? 10GH/s? 100GH/s?

It's never "more profitable" in terms of probability.  However, it can be "worth it" to solo mine if you don't expect to be waiting ages for a block.  Realistically, I probably wouldn't bother with solo mining unless I was pulling at least 25 BTC per day at a pool (i.e. I would expect to solve a block solo every 2 days). 

But hey, if you feel lucky, go for it 

[crazyates]

At what point is it considered more profitable to solo mine, rather than mine at a pool? 1 GH/s? 10GH/s? 100GH/s?

It's never "more profitable" in terms of probability.  However, it can be "worth it" to solo mine if you don't expect to be waiting ages for a block.  Realistically, I probably wouldn't bother with solo mining unless I was pulling at least 25 BTC per day at a pool (i.e. I would expect to solve a block solo every 2 days). 

But hey, if you feel lucky, go for it 

Maybe I will once I get my ASICs. 

I've been mining for about a year, and I've found a number of blocks (abeit small number) since starting. Looking back and doing the math, I've been paid < 60% compared to if I had found those blocks solo mining. Maybe I've just been lucky, but idk. That's why I was asking.

[Stephen Gornick]

At what point is it considered more profitable to solo mine, rather than mine at a pool? 1 GH/s? 10GH/s? 100GH/s?

 - https://en.bitcoin.it/wiki/Difficulty#How_soon_might_I_expect_to_generate_a_block.3F

Currently difficulty is 2,694,048.

From the wiki page:

  time = difficulty * 2**32 / hashrate

Thus
 hashrate = difficulty * 2**32 / time

When the difficulty is going up, you really get penalized if you didn't solve a block before the difficulty goes up.  So you might want to target a block every four days, let's say.

4 days = 4 * (24 * 60 * 60) = 345,600 seconds.


 hashrate = 2,694,048 * 2^32 / 345,600

 hashrate = 33,480,463,118 hash/s  or 33.5 Ghash/s.

[stick_theman]

After you buy a BFL SC Mini-Rig at 1TH? lol

[os2sam]

At what point is it considered more profitable to solo mine, rather than mine at a pool? 1 GH/s? 10GH/s? 100GH/s?

When/if the network hashrate goes to 400+Ths those of us that have GPU rigs may as well solo mine since we'll get virtually nothing from mining on a pool.  Leave it to random chance on solving a block.  That will also help keep the network decentralized which is a good thing.

If you are one who can afford a high end ASIC/FPGA Rig/Farm, keep mining at your favorite pool and keep getting a steady income.
Sam

[crazyates]

At what point is it considered more profitable to solo mine, rather than mine at a pool? 1 GH/s? 10GH/s? 100GH/s?

When/if the network hashrate goes to 400+Ths those of us that have GPU rigs may as well solo mine since we'll get virtually nothing from mining on a pool.  Leave it to random chance on solving a block.  That will also help keep the network decentralized which is a good thing.

If you are one who can afford a high end ASIC/FPGA Rig/Farm, keep mining at your favorite pool and keep getting a steady income.
Sam

Lol we have opposite views on this then. You're saying smaller operations should solo mine, but larger farms should pool mine. I've always thought of it as the opposite, where pools are more attractive to those with smaller hash  speeds (just so they can get something), but there becomes a certain point as you grow that it's better off to solo farm.

- https://en.bitcoin.it/wiki/Difficulty#How_soon_might_I_expect_to_generate_a_block.3F
Currently difficulty is 2,694,048.
From the wiki page:
  time = difficulty * 2**32 / hashrate
Thus
 hashrate = difficulty * 2**32 / time
When the difficulty is going up, you really get penalized if you didn't solve a block before the difficulty goes up.  So you might want to target a block every four days, let's say.
4 days = 4 * (24 * 60 * 60) = 345,600 seconds.
 hashrate = 2,694,048 * 2^32 / 345,600
 hashrate = 33,480,463,118 hash/s  or 33.5 Ghash/s.

So you think a block every 4 days is about where it becomes beneficial?

[the joint]

So you think a block every 4 days is about where it becomes beneficial?

It will be beneficial if you are getting lucky and solving blocks in less than 4 days.  It will not be beneficial if you are unlucky and solve a block in several weeks/months.

Solo mining is great when you start getting 100+ BTC in 4 days (when the average is 50 BTC in 4 days).  Solo mining sucks when you get 50  BTC in two weeks (when the average is still 50 BTC in 4 days).

[Graet]

Rather than solo mine and increase variance a lot of larger miners do this
https://bitcointalk.org/index.php?topic=78031.0
to reduce variance

um, even some pools do this actually 

[os2sam]

At what point is it considered more profitable to solo mine, rather than mine at a pool? 1 GH/s? 10GH/s? 100GH/s?

When/if the network hashrate goes to 400+Ths those of us that have GPU rigs may as well solo mine since we'll get virtually nothing from mining on a pool.  Leave it to random chance on solving a block.  That will also help keep the network decentralized which is a good thing.

If you are one who can afford a high end ASIC/FPGA Rig/Farm, keep mining at your favorite pool and keep getting a steady income.
Sam

Lol we have opposite views on this then. You're saying smaller operations should solo mine, but larger farms should pool mine. I've always thought of it as the opposite, where pools are more attractive to those with smaller hash  speeds (just so they can get something), but there becomes a certain point as you grow that it's better off to solo farm.

If the network hash rate increases astronomically over one difficulty period like some are predicting then yes the larger farms should, obviously, solo mine for that short period before the difficulty increases astronomically to match the hash rate.  Then after that the drunken party will be over so the large farms will probably be better off pool mining.  At least until the next bump in technology.

Us small time miners will get virtually no income from pool mining, unless the Bitcoin value increases dramatically as well, so we may as well solo mine and hope for a block once in a while, or just shut down completely and quit mining altogether.  I'm too stubborn for that though .

Maybe once the ASIC's come out the current big time miners will want to sell their FPGA's and I'll by one or two of those and just let the mine in the corner for a few years and not think about it.

When everyone else is going with the flow someone has to go against the grain .
Sam

[crazyates]

Rather than solo mine and increase variance a lot of larger miners do this
https://bitcointalk.org/index.php?topic=78031.0
to reduce variance

um, even some pools do this actually 

Is this what the --balance option in CGMiner does? I've heard a lot of people talking about that option a while back, but I never understood the purpose. Thanks for pointing out that thread!

[os2sam]

Rather than solo mine and increase variance a lot of larger miners do this
https://bitcointalk.org/index.php?topic=78031.0
to reduce variance

um, even some pools do this actually 

Is this what the --balance option in CGMiner does? I've heard a lot of people talking about that option a while back, but I never understood the purpose. Thanks for pointing out that thread!

In my experience both balance and load balance are skewed heavily toward Ozco.  ckolivas says that's because load balance make more use of pools using rolltime since that has a much lower load.  So I guess Ozco uses rolltime?  I can't find any references to that on list of features though.

You would probably need to run multiple instances of your miner to distribute your hash rate reliably.
Sam

[Stephen Gornick]

So you think a block every 4 days is about where it becomes beneficial?

I think it will depend on a number of factors.  If I expect difficulty will be going down in the future I'm more inclined to lengthen that duration from 4.  If I expect difficulty will go up and go up by a large amount, I might want to make the duration target well under between now and when the next adjustment is to occur.

For example, say the difficulty adjustment happens in five days.  I can use the calculation from above to learn that with my hashing capacity I have a 50% chance of solving a block in three days, and an 80% chance in five days -- or whatever the results are after calculating the probability.  So if I want to use as my target a 70% probability, I then will know that I should continue mining solo for another half day or so, until that 70% is reached.

For any probability target, hashing capacity and difficulty level, I should be able to calculate the exact block at which I want to switch over to pool mining.

But there are other factors as well.  The fees for pool mining weigh in.  My cost of electricity weighs in as well.  If I have electricity included in my rent, I might be more willing to take the chance with a lower probability target as the risk of not getting a block is lower (simply just lost income).  Whereas if I'm barely profitable and I am unlucky, I still have to pay the electric bill out of pocket and didn't get any mining revenue to cover it.   So that's a deterrent.

So determining the "right point" for going solo is dependent on each individual's risk tolerance and expectations for the direction and degree of the next difficulty adjustment.

[jojo69]

I have found 5 blocks on a pool where I have only earned 83 BTC

dammed socialism

[Graet]

Rather than solo mine and increase variance a lot of larger miners do this
https://bitcointalk.org/index.php?topic=78031.0
to reduce variance

um, even some pools do this actually 

Is this what the --balance option in CGMiner does? I've heard a lot of people talking about that option a while back, but I never understood the purpose. Thanks for pointing out that thread!

In my experience both balance and load balance are skewed heavily toward Ozco.  ckolivas says that's because load balance make more use of pools using rolltime since that has a much lower load.  So I guess Ozco uses rolltime?  I can't find any references to that on list of features though.

You would probably need to run multiple instances of your miner to distribute your hash rate reliably.
Sam

we have been rolling time since we started using ecoinpool (end of last December)
it is one of those things all pool should be doing
If there are that few, I guess I should add it as a "feature" but it is a bit like calling headlights on a car a feature imo
cheers
Graet

[Shadow383]

If you had the sort of hashrate where solo mining would make sense (~100-200Gh/s) then it'd still be more profitable to either find someone who'll pay you over 100% (like me  ) or to set up your own system for hopping/leasing hashrate.