I was a hardware engineer before I was a reviewer. That is the whole reason this site publishes measurements instead of impressions — I know exactly which corners get cut, because I was in the room when they were argued about.
How a device gets scored
01
Box-fresh baseline
Capacity, brightness and thermals recorded before anything is configured.
02
Daily driver
It becomes my actual device. Nothing is simulated, nothing is a benchmark loop.
03
Fade and stress
The baseline is re-run. This is where marketing claims usually come apart.
04
Open it up
Repairability is scored with a screwdriver, never from a spec sheet.
What's on the bench
Published so you can reproduce a result, or tell me where the method is wrong.