A few thoughts - Tuesday, February 10, 2015
Posted: Tue Feb 10, 2015 10:29 am
A few thoughts for today:
Mining profitability stabilizes
The main reason I wanted to comment today was because I noticed how it is so remarkable that mining profitability has stabilized recently. After dropping to 4 cents per MH in November, profitability has risen to 5 cents and has remained stable for months. Lots of things have happened, and none of them seems to have caused the profitability to budge:
Given that it is currently the coldest part of the year, one would think that people would be willing to run miners at significantly less-profitable conditions, but that doesn't appear to be true. For example, since we would normally need to run an electric heater in the basement, the cost to run these ASICs is actually negative, and will continue to be so until high temperatures average 60 degrees on April 20.
Therefore, the reasons for the increase of profitability must lie in the initial production of ASICs. One possibility is that ASIC manufacturers are no longer able to make money after the crash, and have reduced R&D into new models to continue operations. If all manufacturers behave the same way, the market reaches an equilibrium. This happened in the hard drive market in 2012 and 2013, when hard drive manufacturers like Seagate rested on their laurels without advancing to the next technology. But Intel did not, and 512GB solid state disks are now common. It wasn't until a competitor decided to make the solid state disks huge that the hard drive manufacturers decided to start making investments again. Now, 10TB drives are available and declining in price.
Another is that 5 cents is just below the value where a new ASIC will break even. That means that nobody can buy ASICs to start new mining operations, but those who already own them are laughing all the way to the bank. We'll see if this is born out by events, but I suspect that profitability will increase to 6 cents by March, as people who would otherwise consider starting mining see the summer approaching in a few months and even fewer of them decide to buy in. By April, we could even see 7 cents.
Finally, an interesting possibility is simply, as I mistakenly did in 2012, miners have lost interest in cryptocurrencies and left the space. I left 10 bitcoins in Slush's pool, because I didn't think it was worth the bother to secure a wallet for $20 (they were later confiscated by Slush, who "expired" bitcoins after a year). If we assume that most people are not rational, which is a good assumption, then mining is also subject to the same whims that we see in the markets. People join and leave based on sentiment, just as they buy high and sell low based on sentiment.
Whatever the cause, this is a good time to be a miner. Historically, the best time to get into mining has always been at the low point of a cycle. Suppose that you had purchased a bitcoin in October 2013 and somehow managed to sell it exactly at the top, making a profit of $800. If you had instead used the bitcoin to buy a 700 Kh/s graphics card, you would have made $30/day in December for 20 days, and then an average of $5/day for the next six months ($1500 after electricity). Contrary to popular opinion, buying mining equipment at the very bottom of bubbles is a good investment because difficulty does not increase at the same rate as price does, and ASICs usually don't come out until after the bubble pops.
Development slows down
Recently, some people made a huge deal about Andreas Antonopolous stating that 2017 will be the year for bitcoins. Few of these people seem to remember that 2015 was or is also supposed to be the year of bitcoins. Antonopolous seems to take pride in moving the goalposts forward.
This is a common tactic in software engineering. If things don't come together quickly enough, then instead of declaring the project a failure, you just throw more money at it, give it more time, and hire more people. Given all that investment, the project can't possibly fail. The managers quickly find out that these things just put the project further behind schedule, as more requirements are added and developer time is diverted to training all the newbies and fixing their bugs.
The reason that most of these hailed bitcoin projects are vaporware is requirements creep. As I've stated many times before, the only way that projects are going to succeed is to put out a version and then improve it. OpenBazaar is a good example. They stated that they want to have 100% test case coverage. Testing is great and is a necessity for great projects, but let the level of testing be dependent upon the number of features and the usage of the system. When more people use a product, they uncover weird bugs because they all use the system differently, but when only a few people are using it, they are less likely to encounter issues simply because of statistical probability.
The other problem with 100% test cases is that many of the test cases are testing things that are not important or even noticeable to users. For example, is it really going to break a project if on IE with Windows XP, some text doesn't appear properly on the right side of the screen? A better approach is to make the system work well for 90% of users, acceptably well for another 5%, and to ignore the remaining 5% until time permits.
Do some initial testing, and then put the project out there. If bugs are discovered, then fix them and issue a new release. Except in life-critical systems, something is better than nothing. Having a product generates publicity and followers. As long as things are moving in the right direction, then the software will gain more stability and features over time. Right now, there are lots of projects with lots of talk after years of development, with nothing to show for them.
Mining profitability stabilizes
The main reason I wanted to comment today was because I noticed how it is so remarkable that mining profitability has stabilized recently. After dropping to 4 cents per MH in November, profitability has risen to 5 cents and has remained stable for months. Lots of things have happened, and none of them seems to have caused the profitability to budge:
- Pools come and go
- New ASICs are always being released
- The price of bitcoins has fallen significantly
- Altcoins get pumped and dumped
- New exchanges open and old exchanges close
Given that it is currently the coldest part of the year, one would think that people would be willing to run miners at significantly less-profitable conditions, but that doesn't appear to be true. For example, since we would normally need to run an electric heater in the basement, the cost to run these ASICs is actually negative, and will continue to be so until high temperatures average 60 degrees on April 20.
Therefore, the reasons for the increase of profitability must lie in the initial production of ASICs. One possibility is that ASIC manufacturers are no longer able to make money after the crash, and have reduced R&D into new models to continue operations. If all manufacturers behave the same way, the market reaches an equilibrium. This happened in the hard drive market in 2012 and 2013, when hard drive manufacturers like Seagate rested on their laurels without advancing to the next technology. But Intel did not, and 512GB solid state disks are now common. It wasn't until a competitor decided to make the solid state disks huge that the hard drive manufacturers decided to start making investments again. Now, 10TB drives are available and declining in price.
Another is that 5 cents is just below the value where a new ASIC will break even. That means that nobody can buy ASICs to start new mining operations, but those who already own them are laughing all the way to the bank. We'll see if this is born out by events, but I suspect that profitability will increase to 6 cents by March, as people who would otherwise consider starting mining see the summer approaching in a few months and even fewer of them decide to buy in. By April, we could even see 7 cents.
Finally, an interesting possibility is simply, as I mistakenly did in 2012, miners have lost interest in cryptocurrencies and left the space. I left 10 bitcoins in Slush's pool, because I didn't think it was worth the bother to secure a wallet for $20 (they were later confiscated by Slush, who "expired" bitcoins after a year). If we assume that most people are not rational, which is a good assumption, then mining is also subject to the same whims that we see in the markets. People join and leave based on sentiment, just as they buy high and sell low based on sentiment.
Whatever the cause, this is a good time to be a miner. Historically, the best time to get into mining has always been at the low point of a cycle. Suppose that you had purchased a bitcoin in October 2013 and somehow managed to sell it exactly at the top, making a profit of $800. If you had instead used the bitcoin to buy a 700 Kh/s graphics card, you would have made $30/day in December for 20 days, and then an average of $5/day for the next six months ($1500 after electricity). Contrary to popular opinion, buying mining equipment at the very bottom of bubbles is a good investment because difficulty does not increase at the same rate as price does, and ASICs usually don't come out until after the bubble pops.
Development slows down
Recently, some people made a huge deal about Andreas Antonopolous stating that 2017 will be the year for bitcoins. Few of these people seem to remember that 2015 was or is also supposed to be the year of bitcoins. Antonopolous seems to take pride in moving the goalposts forward.
This is a common tactic in software engineering. If things don't come together quickly enough, then instead of declaring the project a failure, you just throw more money at it, give it more time, and hire more people. Given all that investment, the project can't possibly fail. The managers quickly find out that these things just put the project further behind schedule, as more requirements are added and developer time is diverted to training all the newbies and fixing their bugs.
The reason that most of these hailed bitcoin projects are vaporware is requirements creep. As I've stated many times before, the only way that projects are going to succeed is to put out a version and then improve it. OpenBazaar is a good example. They stated that they want to have 100% test case coverage. Testing is great and is a necessity for great projects, but let the level of testing be dependent upon the number of features and the usage of the system. When more people use a product, they uncover weird bugs because they all use the system differently, but when only a few people are using it, they are less likely to encounter issues simply because of statistical probability.
The other problem with 100% test cases is that many of the test cases are testing things that are not important or even noticeable to users. For example, is it really going to break a project if on IE with Windows XP, some text doesn't appear properly on the right side of the screen? A better approach is to make the system work well for 90% of users, acceptably well for another 5%, and to ignore the remaining 5% until time permits.
Do some initial testing, and then put the project out there. If bugs are discovered, then fix them and issue a new release. Except in life-critical systems, something is better than nothing. Having a product generates publicity and followers. As long as things are moving in the right direction, then the software will gain more stability and features over time. Right now, there are lots of projects with lots of talk after years of development, with nothing to show for them.