How to Use This Tool
Note a percentage, wait, note it again, and enter the minutes between. Longer gaps give better answers, which is the whole point.
The arithmetic
Rate is the change in percentage divided by the time it took. Remaining time is the distance to your target divided by that rate.
A battery going from 80% to 65% in 30 minutes is losing 0.5 percentage points a minute, so reaching 10% takes another 110 minutes — an hour and fifty. Charging from 20% to 55% in 25 minutes is gaining 1.4 points a minute, so 80% arrives in another 17.9 minutes.
Why the built-in estimate is worse
A progress bar or battery indicator usually extrapolates from a very short recent window — the last few seconds. That makes it responsive and extremely noisy, because almost nothing progresses at a constant rate:
- File transfers race through one large file and crawl through ten thousand small ones, because per-file overhead dominates when the files are tiny.
- Batteries drain at whatever the current load demands, and the relationship between the percentage shown and the charge remaining is not linear anyway.
- Installers and builds often have no idea what fraction of the work is done and are reporting steps completed rather than time.
Measuring over ten or thirty minutes averages all that out. It is still an extrapolation and it is a far steadier one.
Where a straight line is most wrong
Linear extrapolation is a good approximation in the middle and a poor one at the ends.
- Charging above 80% slows sharply, so a rate measured from 20% badly underestimates the time to full.
- The last few per cent of a battery often drop faster than the arithmetic suggests, partly because the gauge is recalibrating rather than because the chemistry changed.
- The last percent of a transfer is frequently verification, flushing buffers or cleanup, which the percentage never accounted for.
The practical version: take your two readings inside the region you actually care about. Estimating the time from 80% to 100% works much better if both readings were above 80%.
Reading it honestly
This produces a point estimate from two data points, which is the least information from which an estimate can be made at all. It has no error bars and no knowledge of what happens next.
If a decision depends on it — catching a train, leaving a laptop unattended — take a third reading and see whether the answer moved. A stable estimate across two windows is worth something; one that halves each time you measure is telling you the rate is not steady.
