Topic: ASICMINER: Entering the Future of ASIC Mining by Inventing It - page 696. (Read 3917468 times)

 huge swings are possible once you are less then 10% of the total pool 1 block in 1 day  is not that uncalled for.

6 blocks in 1 hour is the 'norm'  for the entire pool or about 144 blocks a day.  If you are 10 percent of the pool 14 blocks is common place or your norm.

 But here is a simple way of doing your odds. if you are at 10% of the network.  vs 5% of the network
  
1 miss is .90                                                                                       .95
2 miss is .81                                                                                         .902
3 = .729                                                                                             .857
4 = .656                                                                                                        .814
5= .5905                                                                                                       .773
6 = .5314   so at 10% you are about 47/100 to do 1  block in the first hour                .735 at 5% your are about 25 -75 to do 1 block in the first hour
7 = .4783                                                                                                      .698
8 = .4304                                                                                                      .663
9 = .3874                                                                                                      .630
10 = .3486                                                                                                     .598
11 = .3138                                                                                                     .568
12 = .284 at 10%  your are about 72/100  to do 1 block by 2 hours                        .540  at 5% your are about 46 /100 to  do 1 block by 2 hours
13 = .254                                                                                                       .513
14 = .2287                                                                                                     .487
15 = .2058                                                                                                     .463
16 = .185                                                                                                       .439
17 = .1667                                                                                                      .417
18 = .1501  at 10 % you are about 85 /100 to do 1 block by 3 hours                        .396  at 5% you are about 60/100 to do 1 block by 3 hours


the point is at 3 hours 10 % of the network vs 5 % of the network  allows a big difference for more then 2x as some would think.

Yes this is variance. You can verifiy it after 3 days...

The whole network suffers from bad luck  sometimes. We could have 1400.093 Thash/s right now or 886.093 Thash/s... Its just variance

Candoo is right, variance really can be this "sharp".

it is because the swings are too sharp against variance. That is why most of us here are speculating about events happening in the farm.

Why do we keep repeating this... Its called variance

Last 30 hour hashrate reading is at 58.23 and the 20 hour is at 73.90. Perhaps some new hardware is being deployed?

it makes sense if you consider that he doesn't mine with the hardware before selling it.  So, if he mines for 2 months with the hardware, then what?  Who buys Gen1 hardware in a Gen2 world? 

Why not just sell it right now and get the money in hand?

What you don't consider when trading off selling versus self-deployment is the overhead and time it takes to get a datacenter at this energy density up.

Also, our hashrate rather dropping then rising could mean the near-future deployment of gen2 hardware, although this is only scheduled for november and a sell-off now would not make sense.

To offshoot the bitcoin network just reached 1 petahash so assuming growth projection and estimated hash coming online according to Friedcat's timelines it does get interesting.

Q: How much hashing power does Bitfountain plan to deploy within this year?
Friedcat: 800-1000T within this year.

http://siliconangle.com/blog/2013/09/18/bitcoin-weekly-2013-september-18-bigpoint-games-to-accept-bitcoin-network-exceeds-1-petahash-a-second-armory-technologies-raises-600k/

In other words probably 25-30% of network by the time the hardware comes online might be doable but I digress we will need to wait.

I am getting it mainly from one of the resellers - canaryInTheMine - he claims that he sold around 15TH/s. There are around 5 big resellers, so it's safe to estimate that 4 other sold together at least the same amount as the biggest one.

Where are you getting the "sold" estimate from? Income? How much roughly has selling 30 - 50 TH/s raised?

in house hashing ~45TH/s
franchising ~15TH/s
sold ? 30 - 50 TH/s ?

https://bitcointalksearch.org/topic/m.3165200

he didn't say he was going to deploy the whole 200 TH/s.  How much hardware has been sold?

I agree, he said that, and not long ago. Yet its getting towards the end of September, and we are still at 50TH/s per week, absolute max. So where is it? If it was being franchised out, it would be getting paid out as dividends. Its not.

What happend to the hardware franchising?

I think it's pretty likely Friendcat will have modelled his projections and optimised his strategy accordingly. I do recall him posting that hashrate growth for the network was under his predictions.

So about that sell-versus-mine tradeoff, there are not one but two considerations:

1) Time preference, already mentioned above several times.  This is more or less obvious: it may be preferable to get slightly less immediately than to get slightly more later, depending on how much "slightly" and how much is "later".  This is not only about APR but about market situation as well: when you have sufficient funds for current needs, your time preference is more likely to be "later but more" than when you need to gather more financial resources quickly.  Here franchising is treated as mining (if I understand correctly, franchising pays off over time, just like self-mining).

2) The less obvious factor here is anticipated network hashrate growth.  By favoring mining over selling they would be betting against hashrate growth relative to market (average public) expectation.  And by selling more they are clearly anticipating that the network will grow faster than the average public expects.  And here franchising is treated as selling (the price is affected by customer's network growth expectations).

So we can conclude that if AM is selling new units instead of deploying them, then it can only mean that at least one of the following statements is true (more likely both):

A) They have time preference for "less but sooner", which may indicate they are accumulating funds for massive final gen.1 production or for launching gen.2.

B) They anticipate higher network growth than the market expects.  This can assure us that they are trying to be well-prepared for the coming 28nm massacre.

Note that franchising volume doesn't indicate much on itself because it belongs to the opposite sides of decision.

As you see, I'm still bullish because both conclusions look optimistic to me.  Now with the current share prices I seem to be against the market, but sadly I don't have enough capacity to buy even more shares.  Regardless, an update from friedcat would be very nice.

Can you copy paste what you wrote there here? It seems to be a private forum.

I'm quite concerned about scaling too, but apparently most people aren't.