A daily visual precision routine.
Set aside about ten minutes for two Pixactly sets and a short review. The purpose is to leave with one observation and one tested correction. The timings below are a suggested schedule; finish at a comfortable pace rather than rushing a round to meet the clock.
Minute 1: fix the setup and choose a question
Note your device, browser zoom, and input method. Keep the window size steady for both sets. Choose a question you can answer from the game: “Do I keep drawing widths too large?” is easier to evaluate than “Am I good at visual design?” If you have no previous notes, use the first set to collect directions without committing to a correction.
Minutes 2–4: play five rounds and record each miss
Draw a full rectangle each round. During feedback, write width as too wide, too narrow, or exact, and height as too tall, too short, or exact. Add the pixel errors if you want a numeric log. Do this before moving on: the final summary does not show each round's signed direction.
After the fifth round, record the total and both average axis errors. If recording takes longer than three minutes, extend the session or stop after this set. The game does not penalize time in its score.
Minutes 5–6: turn the notes into one decision
| Observation | Next-set experiment | What to check |
|---|---|---|
| Width usually too wide | Choose a slightly nearer horizontal endpoint | Width direction and absolute error |
| Height usually too short | Picture a slightly taller vertical distance | Height direction and absolute error |
| Directions alternate | Choose the endpoint before moving | Whether misses become less scattered |
| No clear pattern | Repeat without a size correction | Whether the same issue appears again |
A large average width error identifies an axis to inspect, not a direction to move. Use the notes from the individual rounds to choose between a size correction and a consistency exercise.
Minutes 7–9: play the correction set
Keep the same setup and apply only the chosen change. Record directions again before leaving each round. Leave the other axis alone as much as possible, then compare its error too. A smaller width miss bought by a much larger height miss is a tradeoff to notice.
Targets are newly generated, so the two sets are not identical tests. Keep an improved score as a useful observation and look for the same pattern in later sessions before treating the correction as reliable.
Minute 10: save a note you can use tomorrow
Copy this template into your own notes and fill it from your results:
- Setup: date, device, zoom, input method.
- First set: total; average width/height off; directions noted during rounds.
- One change: the specific adjustment you tried.
- Second set: the same values and whether the other axis got worse.
- Next session: repeat, revise, or collect more observations.
For example, an illustrative note might say: “Mouse, same window and zoom. Four widths overshot. Tried a nearer endpoint. Width error fell, height rose. Repeat before keeping the change.” This preserves more useful context than a total score alone. The practice log guide shows how to calculate signed bias when you also record the numbers.