Is the gold rush over? A look into cryptocurrency mining

Interest in cryptocurrency mining seems to rise and fall with the market. Between March and April last year, when Ethereum’s token saw its sudden rise from around $10 to over $300, it triggered a gold rush. Ether miners were making unbelievable returns, paying off their hardware in a few months or less and pocketing huge profits.

What made Ether mining even more attractive to the tech-savvy was that it didn’t require specialised hardware. Trying to mine Bitcoin without an application-specific integrated circuit (ASIC) specifically designed for the purpose is a fool’s errand. ASIC-based miners are incredibly expensive, and have no use outside of mining.

Ethereum was a different animal, designed to resist ASIC mining and favour the use of graphics processors. A new wave of digital prospectors started buying up graphics cards, tweaking them in ways the usually gaming-focused hardware makers could not have anticipated—minimizing power draw, while maximizing memory bandwidth.

The rush caused a global shortage of high-end graphics cards, driving up the prices and angering gaming enthusiasts who had no interest in mining. Now a bit more than a year later, as the craze has started to subside, stock is more readily available and prices are returning to sane levels.

As with any gold rush, those who got in early enough made good money. Those who got in later—like me—either lost money or are still hoping that prices will rise enough as to pay off the original investment.

Ethereum’s proof-of-work algorithm

What many people rushing into ether mining did not realize was that Ethereum is transitioning from a proof-of-work system, to proof-of-stake.

Blockchains use proof-of-work and proof-of-stake algorithms to decentralize the processing and securing of transactions.

Proof-of-work is the more technical term for what is commonly called “mining” in the context of cryptos. In the simplest terms, that is when you get a computer to perform many millions of calculations, and are rewarded for being first to find the solution to a specific problem.

The basic idea is that finding solutions in a particular proof-of-work algorithm is difficult, requiring that you guess millions of combinations to discover one. Verification of the solutions, on the other hand, must be trivially easy and cheap.

In reality, unless you have massive amounts of computing power under your control, you shouldn’t be mining on your own. Individual miners group together in communities called pools, with each using an algorithm of its choosing to divide rewards amongst participants in the way it deems fair.

Proof-of-stake is a different form of verifying transactions that depends on a validator’s investment in the platform. In general, it requires that validators vote on the next block, with the weight of their vote linked to the amount of currency they lock into a deposit.

Both types of algorithm have benefits and drawbacks, including vulnerabilities that must be considered when developing a blockchain system.

Increasing difficulty

Proof-of-work algorithms typically have a mechanism built in to scale the difficulty of finding solutions based on how many miners are active on the network. Ethereum attempts to adjust its difficulty so that, on average, one block is produced by the network every 12 seconds. (In reality, the average block time is currently between 14 and 15 seconds.)

A feature of the Ethereum proof-of-work algorithm called the directed acyclic graph (DAG) is also constantly decreasing the effective speed with which your hardware finds solutions. The DAG is a file that is generated every 30,000 blocks, and the larger it gets, the slower graphics cards are able to generate solutions to the proof-of-work problem.

The DAG affects all miners on the platform, so it should not impact your relative earnings as much as the network-level difficulty adjustment. However, the DAG does slowly decrease your effective hash rate, and that is worth keeping in mind when consulting sites like WhatToMine.

For example, I operate a rig with six Radeon RX580 graphics cards (4GB variants), which WhatToMine claims can achieve 181.2 megahashes per second. Hashes are solutions to the proof-of-work problem, so that means 181.2 million solutions per second. Under that assumption, Ethereum is the most profitable coin to mine for my rig.

However, thanks to the difficulty adjustments on the Ethereum network my rig now gets below 100MH/s. Adjusting the values in WhatToMine yields an entirely different list of most profitable coins and algorithms to mine. For my rig it would be Monero, or coins based on the Neoscrypt or Equihash algorithms, like Zcash.

Profitability

When you just want to get paid and aren’t too concerned about which algorithm you are mining, so long as it is the most profitable, a multipool system can be useful.

As the mining difficulty on Ethereum increased, we switched our rig over to Nicehash. It switches between the most profitable algorithms automatically and pays you for the work in bitcoin.

There are other options for multipool mining, such as MultiPoolMiner and Awesome Miner, but Nicehash gave us the best result at the time. Unlike systems where you mine coins from a pool directly and automatically convert them to a coin of your choosing, Nicehash allows people to buy mining power from its distributed pool of miners, then pays those miners a share based on the work they contributed.

Just under a year ago I started mining with two rigs—one I put together on my own, and the other we bought from a local company that specializes in mining hardware. Both rigs run with six graphics cards: Radeon RX 580 in the one, GeForce GTX 1080 in the other.

We went in eyes-open with respect to Ethereum’s difficulty increases and potential move to proof-of-stake, deciding to run the rigs as an experiment. Most of us are gamers, so if the mining becomes unprofitable we would be able to use the graphics cards to build sweet gaming machines and write off whatever capital losses we couldn’t recover.

There have been some hiccups along the way, including Nicehash getting hacked. Nicehash has agreed to repay all money lost in installments, and four waves of repayments already made.

Even if both rigs had been mining optimally since July and we somehow sold all our tokens at the historical highs achieved in December, and taking into account the money lost during the Nicehash hack, neither of our rigs would have paid themselves off yet.

Based on current exchange rates, we are still earning more money than we are paying in electricity, but the return on investment period for a mining rig is not nearly as short as it was at the beginning of the Ethereum rush.

The longer it takes to make a return on investment, the riskier it gets. Mining wears hard on components like graphics cards, and if your hardware starts failing before you’ve made your money back the ROI calculation starts looking grim.

Small-time mining with the intent to make immediate obscene profits from cryptocurrencies is therefore not possible at the moment, and hasn’t been for some time.

There are still other reasons to buy a mining rig, but if you are looking for the most financially sound way to buy into cryptocurrencies, you would be better served taking the money you would spend on a mining rig, and buying a spread of the top coins directly instead.

Regardless of which route you choose, remember to thoroughly investigate the coins you are getting into, and make sure you understand all the security implications of how you choose to store your cryptocurrency.