4
Is the invention of a faster computer (let's say a quantum computer) or the lack of advancements in computing power (let's say it takes a decade to make any progress greater than incremental increases on currently technology) capable of affecting the amount of time it takes to halve the block reward?
In other words, does the block reward get halved every 4 years, regardless of how much the hardware available is or isn't progressing?
5
The block halving takes place every 210,000 blocks. The difficulty retargeting mechanism makes it so that 210,000 blocks take approximately 4 years, but this is not exact. If the hashrate is rapidly growing, it will take a little less than that.
For example, if the total network hashrate doubles every year, it will take about 1 month less. If it doubles every week (increases by a factor of 4.5 quadrillion every year), it will take only 2 years.
2
The short answer is yes, provided the difficulty factor feedback loop (adjusted after 2016 (14*24*6) block awards which nominally maps to every two weeks for whatever next generation hardware gets used) remains stable, which it should from basic control theory 101.
The reason has everything to do with mathematics and the convergence of the geometric series to limit the number of Bitcoins to 21M.
Go to http://www.basic-mathematics.com/geometric-sequence-calculator.html. Enter 10.5M (number of Bitcoins mined during the first four years 50*4*365*24*6) into the first field, 0.5 into the second field (number of Bitcoins awarded gets scaled back every 4 years), 8 into the third field for n (roughly the number of 4 year time constants between 2009 & 2040), and select Sn and hit calculate. As n approaches infinity, the geometric series converges to 21M.
so the answer is yes, it will get halved approximately every four years, regardless of hardware advancements or lack thereof. do i have this right? – Gaia – 2013-10-27T05:16:55.230
@Gaia: Correct. – Meni Rosenfeld – 2013-10-27T05:28:37.707
hmm, just saw your edit: "if it doubles every week, it will take only 4 years". So the 4yrs can be affected only down, not up? – Gaia – 2013-10-27T05:35:51.413
3@Gaia: 2 years, not 4. Anyway, if hashrate is going up it will take less time, if hashrate is going down it will take more time. – Meni Rosenfeld – 2013-10-27T07:52:42.307
Yes, I meant 2 not 4 yrs. So is the answer now "no, the halving period does vary according to advancements (or lack thereof) in hardware" ? – Gaia – 2013-10-27T16:46:09.827
@Gaia: Not sure what you're asking. As I explained, the more rapidly the hashrate is increasing, the less time halving would take. The change is very small for any reasonable hardware progress. Note that to get to 2 years you need the hashrate to multiply by 4500000000000000 every year. – Meni Rosenfeld – 2013-10-27T22:20:26.660
I think you do understand what I'm asking. I get that if HW progression is geometrical, we could eventually end up with 2 instead of 4 years. What if HW doesn't progress much, could the 4 become 5 or 6 years? This is the question "So the 4yrs can be affected only down, not up?" – Gaia – 2013-10-28T04:01:54.623
@Gaia: If the hashrate is static, the time will be pretty much exactly 4 years. Only if the hashrate goes down it can take more than 4 years. Note that the network only sees the hashrate, it doesn't know if increases in hashrate come from technology advancement or more money poured into mining. – Meni Rosenfeld – 2013-10-28T06:28:39.637