Playtesting: The Art of Shutting Up and Watching
You cannot feel your own game. You know where every secret is, what every button does, and what you meant — which makes you the one person on earth who can't experience it fresh. This is the curse of knowledge, and the only cure is watching someone who doesn't have it. Playtesting is the highest-leverage skill most beginners never practice.
The one rule that matters most: shut up and watch. The instant you explain, hint, or reach for the controller, you've contaminated the test. Every place the player gets confused, stuck, or frustrated is data you desperately need — and your instinct to rescue them destroys it. Sit on your hands. Let them struggle. Write down where.
What you're actually looking for:
- ▸Where do they hesitate? Hesitation = unclear communication. Your level failed to teach something.
- ▸Where do they die, and do they understand why? A death they understand is fair. A death that produces "what? that's bullshit" is a design bug, not a player skill issue.
- ▸What do they do that you never intended? This is gold — players reveal the actual affordances of your game, not the ones in your head.
- ▸When do they get bored? The exact timestamp their attention drifts is where your loop has a dead zone.
The protocol:
1. Say nothing except "play this and think out loud."
2. Watch their hands and face, not the screen.
3. Note behavior, not opinions. What they do is true; what they say they want is often wrong (the "faster horses" problem).
4. Five testers find ~80% of usability problems. You don't need a lab — you need five people and the discipline to stay silent.
The mindset shift: Every confusion is a gift, not a criticism. The playtester isn't failing your game; your game is failing to communicate, and they just showed you exactly where.
Game to Study
Every studio, from Nintendo to solo itch.io devs
Think of the last time you watched someone else play something you made (or even play a hard game). What did you feel the urge to explain? That urge points at a spot where the game should have explained itself.
Unity Tip
Add a dead-simple analytics logger: write a line to a text file on every death, level-complete, and quit, with a timestamp and position. After a friend plays, you'll have a map of exactly where they died and where they gave up — playtest data even when you can't watch live.
You hand your platformer to a friend. Within 30 seconds they're stuck on something you find completely obvious, and you feel a strong urge to just tell them what to do. Explain why telling them is the *worst* thing you could do right now, and what you should do instead — and what their being stuck actually proves.