Game Design
Playtesting
Why do games that the team thinks are 'perfect' fail? Because creators are blind to their own design. Playtesting reveals truth: not what you thought you made, but what players actually experience.
- **Nintendo** - families play behind a one-way mirror. If grandma can't figure out the controls, redesign
- **Valve** - weekly playtests with random employees. Left 4 Dead's AI Director tuned through thousands of hours of observation
- **Dead Space** - death heatmaps revealed problem spots. Data-driven horror design
Observation Methods
**Observation** - watching players play without interference. Most valuable data: what players DO, not what they SAY. Body language, hesitation, confusion, delight. Silent observation reveals truth.
**Nintendo's observation labs.** Family members play in a one-way mirror room. Designers watch without interference. If grandma can't figure out the controls, redesign. Miyamoto's 'Wife-o-meter' - if his wife ignores the game, it needs work.
**Observation discipline:** - Shut up (the hardest part) - Note timestamps, not interpretations - Watch the face, not just the screen - Confusion = a design problem, not a player problem
A player is struggling with a puzzle and getting frustrated. You're observing. What to do?
Interview Techniques
**Post-play interviews** - asking questions AFTER observation. Players rationalize and misremember - their explanations are often wrong. Ask about feelings and moments, not solutions. Design is YOUR job.
**'Players are bad designers' rule.** If players say 'add more guns', the real problem might be: combat is boring. Listen to the feelings, ignore the solutions. Their fix is often wrong, their frustration is always real.
**Interview tips:** - Reference specific observed moments - Don't defend your design - 'Tell me more' is the go-to phrase - Silence is okay (let them think)
A playtester says 'The game needs a minimap.' Best response?
Metrics Vs Feedback
**Quantitative (metrics)** - numbers: completion rates, death locations, session length. **Qualitative (feedback)** - feelings, opinions, experiences. Both are needed: metrics show WHAT, feedback explains WHY.
**Dead Space's metrics.** Tracked every death location and weapon usage. Heatmaps showed players dying at a specific corner. Solution: move a light, change the enemy spawn. Data-driven design.
**Balancing metrics and feedback:** - Use metrics to find problems - Use feedback to understand problems - Neither alone is sufficient - Trust behavior > opinions
Metrics show 60% of players quit in Chapter 2. A survey says '95% enjoyed the game.' Trust which?
Playtest Cycles
**Playtest cycles** - a structured schedule of testing throughout development. Early + often > late + once. Each cycle: test → analyze → change → test again. Never ship without a final playtest.
**Valve's playtesting culture.** Every week, random employees playtest. Fresh eyes constantly. Left 4 Dead's AI Director was tuned through thousands of hours of observation. Iteration is the culture, not an event.
**Cycle best practices:** - Test early (prototype), test often - Same issue from 3+ testers = real problem - Fix one issue at a time - Final playtest mandatory before ship
Playtesting once before launch is enough
Playtesting is a CONTINUOUS process. Early prototypes, mid-production, polish phase, pre-launch. Each stage reveals different issues. Late testing finds bugs; early testing prevents disasters
Finding a core loop problem a week before launch = catastrophe. Finding it in a prototype = easy pivot. The cost of fixing grows exponentially with the development stage
One playtester hated the ending. Another loved it. What to do?
Summary
- **Observation** = watch silently, note behavior. What players DO > what they SAY. Don't help, don't explain
- **Interviews** - ask about feelings and moments, not solutions. Players know WHAT felt wrong, not how to fix it
- **Metrics + feedback** together explain truth. **Playtest cycles** throughout development, not just before launch
Related Topics
Playtesting connects to the iteration process:
- Prototyping — What you test first
- Iteration — What you do with playtest results
Вопросы для размышления
- Have you ever watched someone use something you made? What surprised you?
- Why is it so hard not to help a struggling player?
- How would you organize a playtest for your game idea?