Bitcoin Mining: A Miners Tale
by Chris WasshuberI have been actively mining bitcoins since June 2014. I fell in traps, was scammed, insulted, lost money, made money, learned a lot and have now learned to wrestle the beast that is called 'bitcoin mining'. It is not an easy way to make money. It is also not guaranteed that you will make money. It requires time and effort to get to know all the good and bad, the things to stay away from and the strategies that work. In this article I will share the lessons I learned in the hope to get you up to speed faster than I did. Of course, this is only my personal story. Some will disagree with my conclusions, but such is life. I have no problem with people disagreeing with me. But everything I write here comes from first hand experience. I have done it, walked the path, put my own money on the line, took the risk. This is not some theoretical blah blah some writer dreamed up reading about bitcoin. You can find enough of that Googling.
Economics of Bitcoin MiningHere are a couple of basic considerations that you should keep in mind before you invest in bitcoin mining. The most important characteristic of bitcoin mining is that it is a zero sum game. It means that every day only 3600 bitcoins can be mined. (It is typically a bit more due to transaction fees which the miners get. In 2017 the mining subsidy will be cut in half and then there will be only 1800 bitcoins plus transaction fees available. The halving of the mining subsidy takes place every 4 years until all of the 21 million bitcoins have been mined.) The system automatically adjusts the mining difficulty about every two weeks (precisely every 2016th block discovered) such that a new block is discovered every 10 minutes. That means regardless of how many or how few people mine, regardless of what the total network hashing rate is, there are only 3600 bitcoins available to be distributed to miners. It also means the more who mine the less everybody will make. The economical consequence of this is that eventually the mining profit will be competed down to zero. As long as it is profitable new miners will join and reduce the profit for everybody until there is no profit left. However, this does not prevent that profits be made in certain periods. Nevertheless, the long term prospects of profitable mining are rather bleak. Most likely there will be a few large industrial miners who enjoy very low energy cost and overhead, who will be able to squeeze out a couple of percent of profit.
The real winners will be service providers and equipment manufacturers rather than miners themselves, just like during the gold rush. Of course, there will be a few miners who started early enough, who had access to cheap electricity, who will be able to make a profit. However, the largest profits can be attributed to the increase of the bitcoin rate rather than mining itself. See my comments further below. It is therefore important for you to figure out if mining is what you want to do, or if perhaps buying and holding bitcoins would be a better strategy.
A comment on inflation. The bitcoin system is setup such that only 21 million bitcoins can ever exist. This is a firm protection against inflation. Nobody can decide to 'print or coin' a few million additional bitcoins. However, in the period where we are still mining a substantial amount of those 21 million bitcoins, we are in an inflationary phase. Currently about 13 million bitcoins have been mined. Every day an additional 3600 bitcoins are being added to the system. These mined new bitcoins produce an inflationary pressure on the bitcoin rate. On the other side of the equation is the demand for bitcoins, which is produced by people buying bitcoins be it for investment purposes or to transact. The balance of these forces will decide if the BTC rate will go up or down. Once the mining subsidy has become relatively small we will see a much stronger increase of the bitcoin exchange rate.
You can look at it this way. Investing in mining is making a bet that few new people will take up mining, and thus the mining difficulty will increase relatively slower than expected. Investing in bitcoins directly is making a bet that more people start using bitcoins and investing part of their money into bitcoins which will drive up the value of a bitcoin. These two are indirectly linked. The higher the bitcoin rate becomes the more people will enter mining and will make mining less profitable, and vice versa. They can therefore be used to some degree to hedge each other.
Owning Mining HardwareFrom my initial research back in summer of 2014 I deduced that the most profitable way to mine bitcoins is to buy my own hardware, run it from home and cash the profit. On paper this looks perfectly fine. I am a computer guy. I know how to build, setup and run hardware. Shouldn't be a big deal, right? Big error. There are several issues I forgot to include in my analysis:
- Buying mining hardware is like walking through a minefield: There are many bad manufacturers and merchants out there. Some are outright thieves who will take your money and then vanish. Others will sell you overpriced old hardware. Others will delay your hardware delivery to a point where the difficulty has risen so much that you can never be profitable with it. Yet others are just bad retailers or sellers who will tell you one day one thing and then the next day another thing with zero customer support. The best manufacturer I have dealt with was Bitmain. I bought quite a number of S2, S3, and S4. But even the best I could find, Bitmain, disappointed me several times with how they handled broken hardware and how they went back on coupons they already issued me. I had hundreds of dollars in coupons from them which I was not allowed to actually use.
- Operating mining hardware requires constant oversight: I can't tell you how often I have to reboot machines, make changes to settings, install new firmware, etc. Some units do run without much problems, but about 50% of the machines I owned had issues of one kind or another. One day a machine would stop hashing for no apparent reason. A reboot didn't fix it. I had to mess around with it for a full day until a combination of settings (frequency, voltage it was a Bitmain S4), was actually working. Other machines loose their pool settings. Some hashing ASICs stop working after a time. All these issues mean that what you think is a 1TH/s machine actually turns out more like a 0.8TH/s machine, because of downtime or malfunction. Add to this all the time you need to spend monitoring and fixing your machines. If you have nothing else to do then that is probably fine. But if you think it is a setup and forget kind of deal then forget about it. It isn't.
- Can you compete with industrial miners? As an individual who wants to buy a couple of machines, as I was, you will always be behind the big miners and mining companies who have millions to spend, can make special deals with hardware manufacturers to get the latest hardware before you can get it at prices you can't them for. Some industrial miners manufacture their own hardware. These are in an even better position. Don't be a fool and think that these companies wouldn't mine at least a couple of weeks before they make the latest technology available to customers. Of course, they do. It is a big advantage that makes them millions.
- Know your energy cost: I knew my kWh cost going in, which is very high here in Massachusetts (18c/kWh residential). I knew what it would cost running my machines from home. What I underestimated was the increase in difficulty and the drop in the BTC rate, neither of which are under your control. With a high energy rate you are among the first to be unprofitable. Unless your kWh is 10c or cheaper I would not recommend you run your own machines from home. Alternatively you could find a data center that will host your bitcoin mining machines. There are a couple of them out there. Check them out. I had a couple of my machines at a data center in Troy MI. It isn't really a big improvement to running them from home. You still own the hardware. You still have to monitor it and inform them to reboot or fix, which typically takes days. At home you could do it faster. But with a data center at least you do not have to do it yourself. But be prepared to spend time with email, following up, checking, remind them that it is still not working, etc. All the time the machine will not mine. Since time is so critical with mining due to the ever increasing difficulty, all these delays really cut into your profitability.
- Once bought mining hardware is hard to sell: Bitcoin mining is a fast changing field. New semiconductor technologies, new startups, new money and investors, new media reports, all of these fuel bitcoin mining. And since bitcoin mining is a zero sum game, there are currently only 3600 BTC being mined every day no matter how large the total hashing power is, you might end up sitting on old hardware very quickly. Selling mining hardware is not that easy. At least I ran into a number of issues with buyers backing out of deals, making unjustified chargebacks, and other lowlife scam artists. That means once you have made your investment into hardware it is not easy to change course, sell the machines, or get out of mining entirely. The best place to sell mining hardware is ebay. Ebay has its own problems with its own crowd of tricksters, but I was able to sell some of my machines there.
Cloud HashingPlease note that after I wrote this blog I found out that PBMining was a scam. I lost several thousand dollars. Do not buy any of their contracts. Another bad experience mining bitcoins. It is quite frustrating. Read the section about PBMining with that in mind.
If owning your own hardware is not the best choice for many interested in bitcoin mining, could cloud hashing be the right thing to do? I think yes, at least it is for me, but you will need to find the best offers. Many cloud hashing deals I looked at were losers from the start. It was clear that one would never be profitable unless BTC goes to $10,000 in the coming months, which it won't - trust me. I am a believer in bitcoin and I think at some point a bitcoin could be valued at $10,000 or more, but this will take years and requires that bitcoin takes many hurdles which isn't a certainty at all.
The pros for cloud hashing are that you are not anymore the owner of hardware and thus all the troubles that come with owning hardware fall away. No more rebooting machines, no more daily or even hourly checking that everything is hashing properly. But this convenience comes at a price. Your profit margins will not look as good on paper than with owning your own hardware. But if you factor everything into the equation it will become pretty clear that cloud hashing makes a lot more sense for most. I found two places where I have bought and am holding hashing power. As things look I will continue to build up my positions there. The two places I can recommend are:
CEX.ioCEX.io is primarily an exchange, where you can buy bitcoin for Dollars and Euros, where you can change bitcoin into litecoin or darkcoin and vice versa. But it also offers giga-hash (GHS). You can buy hashing power with fiat currency or cryptocurrency, primarily bitcoin, litecoin, darkcoin and dogecoin. As soon as you own GHS it will mine for you and the bitcoins are being deposited in your account. On top of it they also do merged mining where at the same time they mine for other SHA256 currencies like Namecoin. This is a little sweetener that adds to your income.
CEX charges a maintenance fee which is $105/TH currently. Back in summer it was at $260/TH, which suggests that they could continue to lower their fees in the future. In a way, this is logical, because if the maintenance fees are higher than what one can earn mining there, their market for GHS would collapse. Why would anybody buy GHS if it would mean a certain loss of money? Mining hardware improves with lower and lower power consumption. Next generations could lower power consumption to 0.1-0.2W/GH. Right now we are somewhere in the 0.5-0.7W/GH range. CEX will upgrade their hardware at some point and then they will be able to lower maintenance fees accordingly. The beauty for you as customer is that you don't need to do anything. Once they lower the maintenance fee you will benefit with your entire holding of GHS at CEX. No selling of old hardware. No buying and installing of new hardware. I consider this a very important advantage.
The other huge advantage at CEX is that you have a liquid market for GHS available to you. At any moment you can sell or buy GHS at the market rate. While the market rate might not be exactly what you want at that time, it allows you to add or dispose of GHS very quickly and conveniently in pretty much any amount down to fractional GHS. No ebay listing, no waiting days if you get a bid and then dealing with potentially problematic buyers. It is a big plus. The downside is that the prices for GHS are usually higher than if you buy your own hardware. But it depends. Right now for example you can buy GHS at $0.7/GHS or sometimes even lower. This is not that much over the market price and lower than hardware from some manufacturers. It pays to monitor the market and be informed. You will be able to tell when it is a good time to buy and when it is a good time to sell.
CEX also offers a trading API through which you can automate buying and seeling based on your own ideas, strategies and algorithms. If you know how to program it is another plus for CEX. I am a programmer and I am exploring their API to automate my trading.
In summary, while prices of GHS and maintenance fees are not as low as you can find in other places, the ability to trade as well as mine is a unique combination that you will not find anywhere else. If you know what you are doing you can earn a lot more with trading than with mining.
PBMining.com(After I wrote this I found out it is a scam.) PBMining.com (PiggyBack Mining) is a completely different offering. You buy 5 year hashing contracts at a fixed BTC price. As of this writing one GHS costs 0.0014 BTC. You can start with as low as 25 GHS. The pros of this offering is that the GHS prices are typically lower than at CEX, and most importantly there is NO maintenance fee. It is by far the most profitable cloud hashing offering I have found. However, you can't sell or trade a contract once bought. And after five years it terminates. You are not the owner of your hashing power. You are really only renting it for 5 years. However, in 5 years the difficulty will be so high that the earning power of GHS will have dropped to dust. 5 years is plenty. If you are not profitable after 1 year or so, then you will not be profitable anyway. If the difficulty continues to rise as it has risen in the past, and the BTC rate will not increase by leaps and bounds, then you will make the bulk of your money in the first months. After that you will only earn peanuts. That is why timing is so crucial with bitcoin mining. To continue earning from mining you also have to reinvest some of what you have mined into new hashing power.
I like PBMining primarily for the simple process they offer and the low prices. Every week they sweep the mined funds into your wallet. It is automatic and you have no worries with the hardware. But you are locked in. You can't sell GHS. That is the primary downside. It is even worse than owning hardware which you can sell, albeit with hassles and work. The PBMining contracts can't be sold. So whatever you invest there be sure that you don't need the money and that you can wait for whatever mining may or may not bring, because you have no idea how quickly the mining difficulty rises and what the BTC rate does. In the end your investment into mining contracts at PBMining could be a loss proposition. Never forget that bitcoin mining is a risky investment. There are no guarantees of a profit. This is particularly important with an offering where you can't back out from.
You might ask how come PBMining can make this offer without charging a maintenance fee? Don't they have to pay for electricity? Yes, they do have to pay for electricity, but my hunch is that they are selling these mining contracts to hedge against their own mining operation. By taking the other side (selling hashing power rather than buying it) they can effectively hedge against higher than expected increase in mining difficulty. Here is how this works. Based on what they are paying for mining equipment and their cost of electric power one can determine a mining difficulty increase up to which point they will be profitable with their own mining operation. Let's assume that rate is 14% every 2016th block (about every 2 weeks). Meaning, as long as the difficulty does not increase by more than 14% whenever it is adjusted their mining will produce a profit. But if the difficulty would increase by more they would loose money. To hedge against that loss they will sell you a mining contract that is profitable up to say 12% (lower than their own profitability of 14%). Should difficulty increase a lot more they have to pay out a lot less to you than you paid for the contract. This profit they make on your contract is used to offset their own mining losses. And should the difficulty not increase that much then they will use their own mining profits to compensate for the profit of your contract. In either case, as long as the rates and amounts of the contracts they sell are adjusted correctly, they will make a fairly fixed and stable profit of a couple of percent. It is a pretty clever scheme. While some of you might resent that they are making a fixed profit no matter what, they are offering to all of us who can only take one side of the deal a pretty good mining contract. They are also giving up the possibility to earn higher returns if the mining difficulty does not increase that much. You, the contract holder, can potentially make a lot more money if difficulty stays low. We still have to accept a level of risk that comes with the contract, but as I wrote above, any other cloud hashing offering I looked at was more expensive.
Here is a little bit of math to quantify their offering. Essentially they are offering an annuity that pays out less and less over time. It has a hard stop at 5 years, but since payments will have decayed so much in 5 years we can assume it will run forever. The error we make is somewhere in the 5th or 6th decimal if not smaller. Payout P will look something like
P = a (1 + q + q^2 + q^3 + ...) = a/(1-q)
where a is the ratio of your hashing power compared to the entire network hashing power multiplied by the bitcoin subsidy and q is the inverse of the increase in difficulty. Say you buy a 1TH contract. The total network hashing power is currently about 285000TH. And in one period where difficulty stays constant the total mining subsidy is 2016*25BTC, or 50400BTC. This means a is equal to 0.17684BTC. The break even point can be calculated if we set P equal to the price of the contract, which currently is 1.4BTC for 1TH. From this we calculate q=0.87369, which is equivalent to a difficulty increase of about 14.5% every adjustment period. You might think that a 14.5% increase is pretty high, but consider that the average of the increase over the last 1.5 years was about 20%. On average the difficulty has increased by about 20% every adjustment period. Your contract would certainly loose money. In the very recent past, the last weeks, difficulty has not increased by that much. The increase was only about 5%, which is unusually low.
There are three factors for this historically low difficulty increase. One reason is the low bitcoin rate. For the last weeks one bitcoin was listed below $400 when it was for most of the year around $600. A low BTC rate means that less new miners are attracted because mining profitability is lower. It also means that existing miners will delay new mining hardware purchases, because hardware is pegged to fiat currency and thus would be more expensive if bought at low BTC rates. The second reason is that mining hardware has caught up with state-of-the-art semiconductor technology. In just two short years bitcoin mining went from CPU to GPU to FPGA to ASIC, and now with ASIC it has quickly cycled through the various semiconductor technology nodes to arrive at the most current one of 16nm. The newest machines scheduled to hit the market in 2015 will use 16nm technology. Beyond that bitcoin mining will be constrained like any other semiconductor technology. Improvements will come much slower. Even the newest 16nm node will not anymore mean that huge an improvement, because yields will be lower and manufacturing costs will be higher. Power consumption will be lower by about a factor 2, but GHS prices will not dramatically drop, perhaps even stay the same or go up. And the third reason for a slow increase in mining difficulty is the burning down of a mining facility in Malaysia. A couple of Peta hashes burned to the ground. With the concentration of mining in data centers and clustered facilities the risk of bigger chunks of hashing power going down is much increased. Natural disasters, errors and sabotage can easily take down a mining facility. For example, many mines are located in Iceland due to their low power costs and easy natural cooling. However, Iceland is also known for volcano eruptions an earthquakes. Those events could easily shut down a couple of mines from one day to the other. Bitcoin mining also attracts criminal organizations due to the relative anonymity of bitcoin transactions. Imagine such an organization would invest tens of millions into bitcoin mining. I am pretty sure they already have. Such organizations are known to kill people for much less money. I wouldn't be surprised if a few mining centers blow up one day which would be one way to ensure the profitability of their own mining operation.
Besides disasters and sabotage there is one important thing to keep in mind. Whatever miners earn from mining will at least in part be reinvested into mining. This means that mining difficulty will increase with mining profitability. The more profitable it is the more hashing power will come online increasing difficulty which in turn lowers profitability. This puts a lid on what one can earn with mining. It will never be crazy high and will tend to go towards zero as it becomes more industrialized.
BTC versus GHS: Owning Bitcoins rather than Mining BitcoinsOne fundamental decision you need to consider is if mining bitcoin is the best way for you to participate in the bitcoin market. For many people it would be much easier to simply buy bitcoins and then hold them for investment. You can buy bitcoins on a number of exchanges. At Coinbase you can link to your bank account (in the US) and buy with fiat currency whenever you wish. Right now BTC is below $400 which many consider a pretty low price. Buy some and hold it as you would hold investment in gold or silver. In a few years chances are that the price of BTC is higher. Then you can sell it at a profit.
When you own GHS and the BTC rate goes up you do not immediately benefit from it. First, GHS it typically more tightly linked to fiat currency, because hardware cost are linked to the rest of the economy which runs on fiat currencies. So when BTC goes up GHS comes down in terms of BTC. This means that simply owning GHS does not allow you to directly benefit from a BTC increase. You do benefit from ongoing operations of mining. Every bitcoin mined will be more valuable if the BTC rate is higher. But that is an advantage that takes time to manifest itself for the miner. It is not immediate. However, if you hold bitcoins an increase in BTC rate has an immediate effect on the value of your holding. The other problem for the miner is that an increase in BTC means that more potential miners are being attracted. When it is more profitable to mine many more will come which will lower the profit for everybody. Remember, it is a zero sum game. The more who are mining the less everybody makes.
I therefore think the most sensible strategy is to diversify your holding. Don't put everything into GHS and don't put everything into BTC. Maintain a holding of both and every now and then change the composition depending on the market situation. If BTC is low buy some more bitcoins, or consider selling some of your GHS for BTC if you are on CEX.io. If BTC is high I tend to sell some BTC and buy GHS for BTC.
Indirect BTC InvestmentYet another way for you to invest in bitcoin and or mining is to buy shares of companies who are operating in this market. This is similar to either buying gold coins or investing in a gold mine or mining company. Each one has its pros and cons. In Germany there is an industrial bitcoin mining company called Coinbau. I don't know if it has already started and if one can buy shares, but there your profit would be an indirect one from the success the company has mining for its own account. Or look up a company called Bitcoin Shop Inc with ticker BTCS. This is a retailer that has made bitcoin a central part of his business model. I can't say if this is a good company to invest in right now, but it is a way to invest in bitcoin without having to deal with some of the messy part of bitcoins and mining. The Winkelvoss brothers are trying to setup an ETF (exchange traded fund) for bitcoin investors. Once that is available it could become a convenient way to invest in bitcoins.
I hope that this was helpful to you. I will expand this article as I gather more information and experience. Please consider making a small donation to this bitcoin wallet address: 1CReEXCZmsrWHcTduW