Forecast methodology
How to assess a difficulty forecast
Learn why block pace informs an estimate, then assess the calculation, range and evidence behind it.
How block timing informs a forecast
Bitcoin adjusts difficulty every 2,016 blocks to target about ten minutes per block. Within its protocol limits, the adjustment scales current difficulty by the target duration divided by the measured duration. Faster blocks point towards a higher next difficulty; slower blocks point towards a lower one.
A simple forecast estimates the full period’s duration by extending the block pace observed so far. It uses that estimate in the adjustment rule to project difficulty. This needs current difficulty, the target block and the observed blocks’ heights and timestamps.
Blocks arrive irregularly, so the pace can change as more arrive. Extending the observed pace assumes it will continue through the remaining blocks. The projected difficulty and calendar date can therefore change; neither is a measured future value.
What makes the calculation reproducible?
A published method must identify the input block heights and timestamps, capture time, target adjustment height and formula. It must also state protocol constraints, the block intervals counted, rounding rules and method version. These details let another reader repeat the calculation.
A projected percentage change, an observed difficulty level and a probability implied by a market answer different questions. Report them separately.
What supports a forecast range?
A range needs its own evidence. Explain its calculation and the period used to fit the method. Test it on a separate period that was not used for fitting, retaining each forecast’s original publication time. Report both range width and the share of outcomes inside the ranges.
The ranges in this website’s examples use made-up values. They have not been fitted to data or tested against a stated confidence level. They show neither how often future outcomes will fall inside a range nor the probability that difficulty will change direction.
What evidence is available here?
This page explains how to assess a forecast method. It does not specify a verified Doefin forecasting algorithm or provide measured accuracy results. The historical report and API contain captured network observations; the synthetic report demonstrates error arithmetic.
Measured accuracy needs original timestamped predictions, comparable lead times, outcomes and corrections. The forecast archive has no published predictions. Its records do not establish independent publication times or prove that hosted files cannot change.
