How We Test Laptop Battery Life: Proposed Reproducible Protocol
A proposed laptop battery-life protocol that controls brightness, refresh rate, network, applications, power mode, battery health, repetitions, and raw logs.
Status disclosure: Laptop Advices has not completed the battery tests described here. The current repository contains no raw run logs, so no review may claim “our battery test” or publish an original runtime. This page is the publication gate for future measurements.
Why one battery number is insufficient
Display brightness and technology, refresh rate, wireless conditions, application mix, video codec, background services, battery health, and firmware can change runtime. A video loop, web workload, and real mixed work answer different questions. Results are comparable only when the test conditions are disclosed.
Proposed test matrix
| Profile | Purpose | Controlled items |
|---|---|---|
| Local video | Decode/display efficiency | Fixed file, player/version, brightness, radios |
| Web workload | Browser/network use | Page set, browser/version, cache policy, access point |
| Office mix | Practical productivity | Fixed script of documents, calls, browsing |
| Heavy work | Short high-load endurance | Named app/project, output, thermal mode |
Required preparation
Record exact model and configuration, OS build, firmware, drivers, battery cycle/health estimate, design/full-charge capacity, display resolution, and active power mode. Calibrate display brightness with a meter to a declared target rather than using an arbitrary percentage. Fix refresh rate, keyboard lighting, audio, and adaptive features unless the profile explicitly evaluates them.
Use the supplied charger, charge to the declared start point, allow thermal stabilization, disconnect power, and verify no update or sync job has started unexpectedly.
Proposed run procedure
- Start synchronized workload and power logging.
- Keep network and room conditions stable and record interruptions.
- Use one declared endpoint, such as controlled shutdown or a fixed remaining percentage.
- Save logs immediately with run identifiers.
- Recharge, cool to baseline, and repeat.
- Publish each valid run, variation, average/median choice, and discarded-run reason.
Do not extrapolate a short percentage drop into an all-day claim. Do not mix an OLED configuration with an IPS result or transfer a larger-battery model’s runtime to a smaller one.
What future result pages must publish
- raw timestamped battery and workload logs;
- exact configuration and software versions;
- brightness in measured units and refresh rate;
- battery health/capacity information;
- network, audio, keyboard-lighting, and power-mode settings;
- run count, variation, and anomalies;
- the difference between test workload and likely reader use.
Readers improving their own runtime should use the battery-life troubleshooting guide. Buyers should evaluate laptop claims using the remote-work guide or college guide without treating a manufacturer’s controlled claim as personal runtime.
Conclusion
A reproducible battery result is a documented workload plus raw evidence, not a single marketing hour count. Until the project stores that evidence, Laptop Advices will publish methodology and sourced specifications only.
Frequently asked questions
Why measure brightness instead of setting 50%?
Percentage curves differ by panel and laptop. The same slider position can produce very different luminance and power use.
Should Wi-Fi remain on?
Yes for a web or office profile and no for a controlled local-video profile, provided the choice is disclosed. Different profiles answer different questions.
Can manufacturer battery claims be compared directly?
Only when workloads and conditions match, which they often do not. Use the claim as a documented vendor result, not an independent measurement.
