SIM13 · 2³ 全因子 DOE / 8 次覆盖全部组合SIM13 · 2³ FULL-FACTORIAL DOE / ALL COMBINATIONS IN 8 RUNS
三个旋钮八发炮弹，找出谁真正管用——全因子 DOEThree knobs, eight shots—which factor actually moves the target?
臂长、皮筋张力、发射角都能调，凭手感一次只动一个旋钮，很容易把最弱的因素当成答案。把三个因素各取高低两档，八发覆盖全部组合；落点和主效应会把真正的控制因子排出次序。Arm length, band tension, and launch angle all look adjustable. Tuning one knob at a time can make the weakest factor look important. Two levels per factor cover every combination in eight shots, letting the landing data rank the real drivers.
🎬 看弹道，也看落点：两者使用同一个射程公式🎬 Follow the trajectory and landing mark: both use the same range formula
DOE · 2³
0 / 8
—
—
炮弹 / 实测落点Projectile / observed landing
1–5 m 尺度1–5 m range scale
主效应Main effect
0.80 m1.00 m
2.03.0
35°45°
💡 跑完看数据：八发把三个主效应排出次序💡 Run it, then read the numbers: eight shots rank all three main effects
落点与主效应均来自 stats()Landing ranges and effects both come from stats()| Run | L (m) | T | θ | 射程 R (m)Range R (m) |
|---|
皮筋张力主效应Band tension effect—
臂长主效应Arm length effect—
发射角主效应Launch angle effect—
🎯 现场规则🎯 Shop-floor rule
别把实验次数花在重复猜测上；用组合设计，把主效应和交互作用一次看清。Do not spend trials repeating guesses. Design the combinations so main effects and interactions become visible together.
⚠ 常见误区⚠ Common mistakes
✗ 一次只动一个旋钮（OFAT）✗ Change one knob at a time (OFAT)
✓ 臂长和张力相乘起作用；只做单因素试验看不到交互作用。✓ Arm length and tension multiply together; one-factor trials hide their interaction.
✗ 多打几发取平均就行✗ Just take more shots and average them
✓ 不设计水平组合，打一百发也分不清谁在起作用。✓ Without planned factor combinations, a hundred shots still cannot isolate the drivers.