Measurement pages
Reading the latency test, the benchmark and the stutter scan honestly.
Each article reads one measurement: what its numbers mean, and what they do not.
- The benchmark, frame times from your own game: Reading the benchmark.
- The latency test, its colours and what a good result looks like: Reading the latency test. Why it asks for twelve seconds of real mouse movement: Why the test asks for 12 seconds of mouse movement.
- The stutter scan, and why it says so when it finds no cause: What the stutter scan looks for.
- What w0a itself costs, with the engines off and on: What w0a's engines cost while a game runs.
It measures frame times while you play your own game, not a synthetic scene. What the percentiles mean and why the average is the least useful number on the page.
What w0a's engines cost while a game runsWith the engines on and a fake game running, w0a used 0.15% of one logical processor on average. Measured on one PC; not measured in a real game.
Reading the latency testWhat the colours mean, what a good result looks like, and — the part most tools get wrong — what this number is not.
Why the test asks for 12 seconds of mouse movementNot an arbitrary number. Here is what 12 seconds buys that 3 does not, and why it needs your real movement rather than a synthetic one.
What the stutter scan looks forIt looks for the causes that leave a trace, tells you which ones it found, and does not invent a cause when it finds none.