If you want to know how to playtest a board game, the first thing to accept is that most of what people tell you afterward is wrong, and most of what they do during the game is right. Playtesters are kind. They want you to succeed. They will say “that was fun” because it is what you say to a friend who just showed you something they made. Their bodies, their pace, and their choices during the game cannot be polite. That is where the information is.
This article is about reading the table instead of the feedback form. It covers what to watch, what to ask, what to ignore, and how to tell the difference between a problem with your game and a problem with your test.
Why is playtest feedback so unreliable?
Three reasons, and none of them make your testers bad people.
Social pressure. You are in the room. You made the thing. The cost of saying “I was bored” is a moment of awkwardness; the cost of saying “that was great” is nothing. Most people take the free option without even noticing.
Poor self-knowledge. People are bad at explaining why they enjoyed or did not enjoy something. They will often name a surface feature when the real cause was structural. “The art is confusing” sometimes means “I did not understand what I was supposed to be doing,” which is a different fix entirely.
Solution bias. Testers who do notice a problem tend to offer a fix rather than describe the problem. “You should add a trading phase” tells you very little. What did they want to do that they could not? That is the question underneath the suggestion, and it is the only part you can use.
Feedback is a lead, not a verdict. The verdict comes from what happened at the table.
What should you watch for during the game?
Be an observer first and a host second. Keep a notebook with timestamps. Track the following.
Where do people look? At the board, at their own cards, at each other, at their phones. Eyes on phones is the clearest signal a game can send. Eyes on other players’ boards is often a good sign; it means they care what opponents are doing.
When does the table go quiet? Some silence is concentration. Some is disengagement. You can tell the difference by posture. Leaning in and quiet is thinking. Leaning back and quiet is waiting for it to be over.
Who takes the longest turns, and when? Long turns early can mean rules confusion. Long turns late can mean the decisions have become too heavy, or that someone is counting points to see if they can still win. Both tell you something about pacing.
When do people laugh, groan, or react? Mark every one. These are the moments your game is producing. If you have fifty minutes of play and two reactions, you have a problem no rule tweak will fix. If you have twenty reactions and they cluster in the last ten minutes, you know where the good part is and can ask why the rest is not like that.
When does someone stop trying? This is the single most important thing to watch. A player who realizes they cannot win will change how they play. Faster turns, less deliberation, more “whatever.” Note the exact moment. If it happens with a third of the game left, you have a runaway leader issue, and our article on catch-up mechanics is where to go next.
What do people do wrong? Rules mistakes are design data. If three tables forget the same rule, the rule is in the wrong place, phrased badly, or should not exist.
What questions get honest answers?
After the game, you will ask questions. The ones below are designed to route around politeness.
Instead of “Did you have fun?” ask:
- “What was the most interesting decision you made?” If they cannot name one, that is the answer. If they name one, you have learned where the game is working.
- “When did you know whether you were going to win or lose?” This exposes pacing problems with precision. “About halfway” is a red flag.
- “Was there a point where you wanted to do something and the rules didn’t let you?” This gets at the desire underneath feature requests.
- “Which part would you skip if you could?” People will tell you what to cut long before they will tell you what is bad.
- “Would you play again right now?” Not “would you play again.” Right now. The hesitation before the answer is the data.
Ask one person at a time if you can. Group feedback converges on whoever speaks first.
When someone offers a solution, thank them and ask what problem it solves. “A trading phase, interesting. What were you hoping trading would let you do?” Then write down the problem, not the solution.
Designers: the newsletter sends one piece like this every week.
How do you separate a bad test from a bad game?
Sometimes the test fails and the game is fine. Learn to spot those cases so you do not change things that did not need changing.
| Symptom | Might be the game | Might be the test |
|---|---|---|
| Nobody understood the rules | Rules are too complex or badly structured | You taught badly. Try the five-minute teach next time |
| One player dominated | Runaway leader or a broken strategy | One tester had played before, or is simply much stronger |
| Game ran long | Too many turns, or decisions too heavy | First-play slowness, or a tester who deliberates on everything |
| Flat energy throughout | Not enough meaningful decisions or moments | Tired group, wrong time of night, or a table that does not like this genre |
| Everyone loved it | It works | Your friends love you |
The way to control for the test is to run more of them with different people. One session tells you almost nothing. Five sessions with five groups show a pattern. If the same thing goes wrong with strangers as with friends, it is the game.
Why does blind playtesting matter?
A blind playtest is one where you are not in the room. You hand over the prototype and the rules and walk away. The group teaches themselves, plays, and reports back.
This is uncomfortable and extremely valuable. It is the only way to find out whether your rulebook works, because when you are present, you fill every gap without noticing. It also removes most of the social pressure.
You do not need to do this early, when the rules change too fast. But before you consider a design finished, it needs to survive a few tables you never saw.
How do you decide what to change?
A playtest produces a pile of observations. Resist the urge to fix everything.
- Fix the biggest structural issue first. Pacing, runaway leaders, decisions that do not matter. These are the problems that make people not want to play again, and they are the ones that change everything downstream.
- Change one thing at a time when you can. If you change five rules and the next test goes better, you have no idea which change did it.
- Ignore anything that only one person said once. Note it, wait, see if it recurs.
- Do not add. Remove. Most prototypes are too big. When a test goes poorly, designers add a mechanic to compensate. Try cutting one instead. Games like Azul and The Crew are remarkable for how little they contain.
- Keep the moments. When you find a part of the game that reliably produces reactions, protect it. Every change should be checked against “does this still happen?”
What does a good playtest log look like?
Keep it simple enough that you will do it every time. One page per session:
- Date, player count, who had played before.
- Time to teach. Time to play. Time for setup and teardown.
- A timeline of reactions with timestamps, even rough ones.
- The moment, if any, when someone stopped trying.
- Rules mistakes that happened.
- Answers to the five questions above, one line each.
- Your one-sentence read of the session, written before you look at anyone’s feedback form.
That last line is the honesty check. If your read and the feedback disagree, trust your read. They were being nice.
What does this have to do with the rest of us?
Most readers will never design a game. But everyone who hosts one is running a small playtest every time. Which games made the table loud? When did someone check out? Hosts who notice these things pick better games for their group, and groups with a thoughtful host keep meeting. Watch the people, not the box.
FAQ
How many times should you playtest a board game?
Dozens of sessions is typical before a design is finished, across many different groups. One or two sessions with friends is enough to find obvious breaks but not enough to tell you whether the game is good.
What is blind playtesting?
A test where the designer is not present and the group learns the game from the rulebook alone. It is the only reliable way to test whether the rules can be understood without you.
What should you watch for during a playtest?
Where players look, when the table goes quiet, how long turns take, when people react, and especially the moment anyone stops trying to win. Behavior during the game is more honest than feedback afterward.
How do you get honest playtest feedback?
Ask specific questions like “what was the most interesting decision you made” and “when did you know whether you’d win,” talk to testers one at a time, and ask what problem a suggested fix is meant to solve rather than adopting the fix itself.
If you are testing a game of your own, the newsletter is the easiest way to keep up: one piece a week on design, playtesting, and the tables games get played at, plus first notice of open prototype nights when The Ludarium opens in Indianapolis. Subscribe at theludarium.com. The game is the reason. The people are the point.
Get the next one in your inbox.
Free guides, game picks, and hosting ideas from The Ludarium.