The Bitcoin network’s hashrate jumped 31% from 188 EH/s to 248 EH/s in 24 hours last Sunday. Well, a quick look at blockchain data would suggest so.
In reality, Bitcoin’s hashrate did not gain over 50 EH/s in one evening and resides just over 200 EH/s. What in fact occurred is best explained by first unraveling what hashrate – a measurement of the number of computations Bitcoin miners are executing per second – actually is.
Hashrate is a measurement of guesses. Bitcoin mining machines perform one function: guess for a value less than or equal to the current difficulty of the Bitcoin network. Finding this value results in the Bitcoin network rewarding a lucky miner a block (while also securing the blockchain). Fresh bitcoins reside inside each block as a reward for processing the block.
Hashrate is measured in hashes per second, or how many hashes a computer can perform over a one second time interval. As additional miners joined the network, hashes per second scaled into greater units, such as terrahashes per second (TH/s) or exahashes per second (EH/s).
Hashrate itself is best used as a benchmark for interest in mining Bitcoin and gauging Bitcoin’s security. In both cases, up and to the right remains a good signal. More specifically, the metric is most useful when used over a period of time as a moving average.
How is hashrate derived?
Hashrate is a metric derived from other values the network records, rather than its own stand alone measurement. Nodes do not keep information on network hashrate – it would be impossible!
Indeed, hashes per second is a volatile measurement in and of itself. Hashrate depends on numerous conditions surrounding the physical hardware, the firmware operating on the machine and the work shares sent from pools to miners.
Most importantly, miners only submit winning blocks to the network’s nodes for propagation, meaning nodes have no knowledge of hashrate from the start. Nodes collect information about the block’s current difficulty target and what target value the winning block hit.
These two values can be used to derive how long it takes for a block to propagate around the network, and then eventually hashrate. As Blockchain.com writes:
The hashing power is estimated from the number of blocks being mined in the last 24h and the current block difficulty. More specifically, given the average time T between mined blocks and a difficulty D, the estimated hash rate per second H is given by the formula H = 2 32 D / T
Data providers are not wrong in showing a hashrate value so high. It’s an implicit consequence of the fact that hashrate is measured by block time. But data users should be careful of using such values without context.
As readers may know, block time remains dependent on numerous factors such as internet connectivity, node peer connections and block weight. Typically, 144 blocks are produced per day at about 10-minute intervals. But this is not always the case.
The key word for describing block creation and propagation is variance, or how far an expected value strays from the mean. Blocks may come in faster or slower, depending on miners finding the correct value to add the next block to the tip of the chain, leading to a deviation from the 10-minute norm. As blocktime varies, so will hashrate measurements.
In fact, measurements of Bitcoin’s hashrate can vary up to 20% in one direction “just based on normal statistical noise,” Bitcoin KPIs creator typerbole writes. Given such high variance, it's a best practice to use discretion when using hashrate as a metric.