Skip to content
Doefin
All articles

Reading a hashrate chart

A change in estimated network hashrate can reflect the observation window. The estimate and its uncertainty need to be read together.

Educational article1 min readReviewed

Concept / Observation windows
In this article 3 sections

How network hashrate is estimated

Bitcoin network hashrate is estimated from difficulty and block timing. It is not a direct count of Bitcoin mining machines.

A difficulty-implied series assumes ten-minute blocks. A 144-block estimate uses recent block timing and can move more sharply. Neither is a live measurement. Check the method, time window and units before comparing charts.

Doefin contracts reference mining difficulty at a specified block. A change in a hashrate estimate alone does not determine the contract outcome.

Why two chart windows can differ

Blocks arrive irregularly. Suppose two estimates use different windows ending at the same time. A run of faster blocks has more influence on the shorter window. The longer window includes more earlier blocks, so the same run has less influence.

The short-window estimate can therefore rise while the longer-window estimate changes little. That difference alone does not establish that more mining machines have joined the network. The longer window smooths recent variation, but is less responsive to a recent change.

Different timestamps or difficulty inputs can also account for a gap between charts. A 144-block window is not always 24 hours, because those blocks can arrive faster or slower. Matching the end time and units helps separate a difference in the calculation from a difference in the period being observed.

What an error band can tell you

The centre line gives one estimated hashrate. An error band shows uncertainty around that estimate under a set of assumptions. A wider band means the chart leaves more uncertainty about the level of hashrate, even if the line itself looks precise.

The method needs to explain what uncertainty the band includes. For example, a band describing variation from irregular block timing does not establish that the difficulty input or observation time is correct. Its width alone does not show that the estimate is accurate.

This article does not report measured band widths or identify a recent network movement as noise.

Continue from here

Check the method

Read the forecast methodology

Check the inputs, assumptions and limits behind the sample difficulty forecast.

Related reading

Testing a difficulty-growth assumption

Higher difficulty reduces expected BTC production from the same fleet. With other assumptions unchanged, less revenue remains to recover the initial cost.

All articles