What Kodak doesn’t want you to know about Bitcoin mining
Mining Bitcoin — A Comprehensive Analysis
With Kodak’s recent release of the KashMiner branded Bitcoin miner, it is prudent to evaluate how mining — and its profitability — dramatically changes over time. In this analysis, I show how for most people, mining Bitcoin is not the best way to gain exposure to gains, but instead simply buying is.
Hypothesis:
For gaining exposure to a cryptocurrency such as Bitcoin, for the average person with non-subsidized electricity, it is more efficient to outright purchase the currency than to mine — even with the latest and greatest miner.
Methodology:
Experimental Case — Mining:
Edit: Thanks to reader feedback from Full Name, I have updated this analysis to include the Antminer S9 — I’ve preserved the S7 analysis as a ‘budget’ miner, and put the S9 in as a ‘high-end’ miner.
- I imported daily average bitcoin price and network difficulty for the year 2017 from Blockchain.info.
- I looked up what the most economical miner, and the most advance miner available on Jan 1, 2017 was, and their cost, power draw, and mining power.
- For the budget miner, this led me to the Antminer S7 (19th Gen) which had a cost of about $440, the APW3 Power Supply at $140, and estimated shipping of $40 combined for a total upfront capital expenditure of $620.
The mining power is 4.73 TH/s, and has a power draw of 1210 W at the wall (within 3% of the advertised 0.25 W/GH). - For the high-end miner, this led me to the Antminer S9, which had a cost of about $2100, the APW3 Power Supply at $140, and estimated shipping of $40 combined for a total upfront capital expenditure of $2290.
The mining power is 14TH/s, and has a power draw of 1468W at the wall (within 7% of the advertised 0.098 W/GH). - I ran the CoinWarz mining calculator for each of the network difficulties over the course of 2017 in order to determine how much BTC would be mined on every given day (on average).
Note: As network difficulty changes only every two weeks, I had to calculate only once every two weeks. - Summing the amount of BTC mined over the year, I multiplied this ‘mined total’ by the cost of BTC at the end of the year to measure gross revenue.
- I calculated the daily power cost based on the wattage draw at the wall of the miner, and an array of electricity prices: $0.01/kWh, $0.05/kWh, $0.10/kWh, $0.15/kWh, and $0.20/kWh.
- I summed the amount of initial startup (the miner) and the daily cost of electricity to determine the overall expenditure over the year. Subtracting this from the gross revenue gave me the profit or returns.
Control Case — Buying Bitcoin:
- In order to match the upfront investment of a miner, in the control scenario, I simulated the purchase of BTC on January 1st, 2017 in the amount of $620 — the same investment as in the mining scenario.
- For each of the energy costs modeled, I then took the daily cost to power a miner, and simulated the purchase of BTC in that amount every day for the year. This is essentially a year-long dollar cost averaging.
- In this scenario, over the course of the year my expenditure is equal to the mining scenario and on the same schedule so I don’t have to consider time-value of money and do discounting.
- Summing the amount of BTC purchased upfront, and over the course of the year provides a ‘purchased total’ which, when multiplied by the cost of BTC at the end of the year provides a gross revenue.
- I summed the amount of initial investment and the daily purchase of BTC to determine the overall expenditure over the year. Subtracting this from the gross revenue gave me the total returns.
Note: These expenses — by design — exactly equal the expenses in the mining case. - To see how it holds up in a down market, I ran the same calculation while inverting the BTC prices over the year. This ‘reverse’ calculation is actually only impacting of two things: The entry price of buying BTC on Jan 1, 2017 in the ‘buy’ scenario, and the exit price in that scenario — everything else has a zero net-sum.
Results
The results for net returns are presented here for each of the five electricity prices, for each of the two market scenarios (bull v. bear), and for the two experimental and control scenarios (budget miner v. buy, and high-end miner v. buy). The full results can be found here.
Budget Miner
Across the bands of $0.15/kWh and $0.20/kWh, we can see that buying is a dominant strategy (using the game theory definition) to mining. We can calculate the electricity cost at which buying is no longer a dominant strategy, and it occurs at $0.1159182/kWh. Below this electricity cost, mining provides a stronger hedge against a bear market then buying. Yet, at this electricity cost in a bull market, the buying strategy still significantly outperforms mining, providing 213% better returns.
If we believe that a bull and bear market have equal probabilities, then the electricity price at which we believe the expected returns for mining equals the expected returns for buying is $0.01092999/kWh. This calculation changes of course depending on your view of the market and the probability of a bull or bear market.
High-End Miner
One thing of particular import with the high-end miner that becomes immediately obvious in the results is how much more energy efficient they are: the high-end S9 miner is just over 2.5x more energy efficient than the budget S7. This means buying outright is no longer a dominant strategy to mining for electricity prices below $0.277774/kWh— which is much more expensive than most people pay today.
However, due to the increased cost of the upfront purchase of the miner, the much greater the opportunity cost for not buying. In fact, even if your electricity costs were $0, the high-end miner would have been unable to outperform the simple buy methodology in a bull market.
Conclusion and Discussion:
Due to its increased power efficiency, the high-end miner provides much more dependable results for mining compared to using a budget-miner. However, in a bull market, the opportunity cost lost due to the much greater up-front capital expense negatively impacts expected returns.
According to a recent EIA report, the average cost of electricity in the US is $0.1284 — so most Americans have a strictly better position buying and holding bitcoin rather than mining with the budget miner, and even then only the ones most pessimistic about the market should consider mining. However given that 57% of Americans have fewer than $1,000 in savings, a $2,100 up-front expense for a high-end miner isn’t viable, while the buying methodology can be followed by anyone. Further, the gains made in a bull market far outweighed the losses in the bear market, with a greater than 15x exposure to gains compared to losses. Thus, if you’re chasing the veritable Black Swan of investing, the buy methodology squarely outperforms mining.
If you are one of those people who has the lucky combination of cheap electricity and pessimism about the market, mining may be viable (though the Kodak miner isn’t likely to be). For everyone else, this is a Kodak moment we’d rather forget.
Use my referral link on Coinbase to claim $10 in free Bitcoin when you buy your first $100 in any cryptocurrency using the link here.
Footnotes:
1. There were no trading fees taken into account on the buy option — these fees would likely amount to less than 0.5% (Gemini is 0.25% and GDAX is 0.3%). However, I also did not take into account pool fees on the mine option — these usually amount to about 1%. Thus this omission of fees is a benefit to the mining analysis.
2. For the budget miner, even at the dominant strategy break even electricity cost of $0.116/kWh, 110 days of the year were spent mining bitcoin at a rate less than the cost of electricity — on these days it would have been strictly better to turn off the miner and buy bitcoin with the money you would have spent on electricity. A mixed approach between mining and buying may be called for, but it also points to the speed of obsolescence of miners — how will your Kodak miner mine profitably for two years, if it only keeps up profitably for 70% of the first year?
