Communication games are hard because the speaker sees something the listener cannot, and words drop detail on the way over. Describing a bomb module shows that gap plainly: the Defuser sees the panel, the Expert holds the rules, and one vague word costs a strike. A fixed description order and shared names close most of it.
What communication games reveal about seeing and saying
A communication game splits a task so that one person holds the picture and another holds the means to act on it. What it reveals is that seeing is fast and complete, while saying is slow and partial. The speaker always knows more than the words carry and rarely notices what got left out.
Most of us never test this in daily life. A friend points at the thing, or glances at your screen, or asks a follow-up later. A good communication game removes those escape routes on purpose, so the only thing crossing between two people is speech.
Bomb Appétit is built around exactly that split. One player, the Defuser, sees a ticking bomb covered in puzzle panels called modules. Everyone else is an Expert holding the manual, and the game never shows the bomb to the Experts. Nobody sees both halves. In an online room the Experts get the timer, the strike count and how many modules are solved, and nothing else about the bomb. If you want the full list of what sits on the Defuser's screen, the post on the Defuser's seat and what it shows goes through it piece by piece.
What makes the bomb a sharper test than a classroom exercise is the price of a bad sentence. A wrong cut or a wrong press is a strike. The first strike makes the countdown run 1.25 times as fast, the second makes it 1.5 times as fast, and by default the third ends the round. So the team feels every lost detail as time, which is the point. People stop assuming they were understood once misunderstanding has a sound effect.
The game also shows who carries which half of the work. The Defuser has to turn a picture into a list of facts. The Expert has to turn a page of rules into one action. Both jobs are translation, and both fail in the same place: the moment one person assumes the other already knows what they know.
The gap between seeing a module and saying it
The gap is everything a Defuser takes in at a glance but never says aloud: how many parts there are, which end counts as first, the exact colors, and which detail the rule actually checks. The Expert can only search the manual with the words that arrived, so every missing fact becomes a question or a guess.
Take the Wires module. Its manual page describes a panel with three to six wires, counted from the top starting at one. The Expert finds the list for that number of wires, then reads it from the top and stops at the first line that is true. The lines ask about things like how many wires of a color there are, what color the last wire is, or whether the serial number ends in an odd or even digit. A Defuser who says "a bunch of wires, mostly red" has given the Expert none of that. The count is missing, the order is missing and the last wire is missing.
The Glyph Keypad shows a different version of the same gap. It has four keys, each with a glyph: an outer shape with a mark inside, such as a circle with a dot. Only one column in the manual holds all four glyphs, and the Expert has to find it. A Defuser who sees the keys knows instantly which ones look alike. The Expert, reading a column of small drawings, cannot tell "the round one with the thing in it" apart from three other round ones.
The root problem is that the Defuser does not know which detail matters. The rules live on the Expert's side, and on a fresh rule seed even a veteran cannot guess them. So the Defuser has two bad choices: say too little and force questions, or say everything in a random order and bury the important part. The way out is a habit that says everything in the same order every time, which the next section lays out.
There is also a gap in the other direction. An Expert who reads "cut the last blue wire" knows what that phrase means on the page. The Defuser hears it with the panel in front of them and may count from the wrong end. Each person checks the words against their own picture, and both pictures feel obvious from the inside.
The description habit: shape, then color, then position
The description habit is a fixed order for every module: first its shape and how many parts it has, then the colors or markings on those parts, then where each part sits, counted from a starting point you name. The order lets the Expert open the right page before the details arrive and makes a missing piece easy to hear.
Shape first. Say what kind of panel it is and how many pieces it holds. "Wires, five of them" sends the Expert to the Wires page and to the five-wire list in one breath. "Four keys with symbols" sends them to the Glyph Keypad. "One big button" means the Big Button page, whose rules ask about things like the button's color and the word printed on it.
Color and markings second. Give the colors in a fixed order and use plain names. Bomb Appétit draws each module color with a letter beside it, so a wire also shows R, B, Y, W or K for red, blue, yellow, white or black. That letter is there so a face never relies on color alone, and it helps when a screen washes colors out. For glyphs, use the manual's own naming pattern: the outer shape, then the mark inside.
Position last, from a named start. Say where you are counting from. The Wires page counts from the top. The Labyrinth page counts columns from the left and rows from the top. A sentence like "the third wire from the top is white" survives any amount of noise; "the middle one is white" does not, because five wires have a middle and four do not.
Here are six example descriptions, each written twice. They are invented for this post and say nothing about which wire to cut or which key to press, since that depends on the rule seed.
| Module | Vague version (example) | Precise version (example) |
|---|---|---|
| Wires | "Some red ones and a blue one, I think." | "Wires, five. From the top: red, red, blue, white, black." |
| Wires | "The bottom one is dark." | "The fifth wire from the top is black. Letter K." |
| Glyph Keypad | "One looks like a little house with a line." | "Top left key: an arch, with a bar across the inside." |
| Glyph Keypad | "There's a circle thing." | "Bottom right key: a circle with a ring inside. A ring, not a dot." |
| Blinker light | "It's flashing kind of fast." | "Short, long, short. Pause. Long. Then a long dark gap, and it starts over." |
| Dial Alignment lights | "Some of the lights are on." | "Top row, left to right: lit, dark, dark, lit, lit, dark. Bottom row coming next." |
Every precise version follows the same order. It names the module or the part, gives the count, then lists details from a stated starting point. None of them is clever. They are just complete.
The habit also helps with edgework, the labels on the bomb's casing: the serial number, batteries, indicators and ports. An indicator is a three-letter label beside a light that is either lit or unlit. "One indicator, Y U M, light lit" follows the same pattern as a wire description. Read all of it once at the start, as the manual's intro asks, and the Experts can write it down before any rule needs it.
Why do Experts hear something different from what the Defuser said?
Experts hear something different because they fill every gap with what the page in front of them expects. The words arrive intact, but the listener attaches them to a picture built from the rules rather than from the bomb. "The blue one" means one wire to the Defuser and possibly the answer to a different question to an Expert.
This happens most when the Expert is on the wrong page or the wrong part of the right page. An Expert halfway down the six-wire list who hears "five wires" may keep reading the six-wire lines out of momentum. An Expert scanning the Color Echo tables may pick the right row and the wrong column, because that module picks its column by the number of strikes on the bomb right now, and the strike count changed a minute ago.
It also happens with words that have a hidden reference point. "Second" depends on which end you count from. "Left" depends on whose left, and on the 3D view the Defuser may be looking at the back of the case while the Expert pictures the front. The manual's intro tells Experts to ask which modules are on the bomb, front and back, for exactly this reason: the bomb has more than one face.
The fix sits on both sides. The Defuser can name the reference point every time. The Expert can say back what they heard in the manual's terms before reading out an action, which turns a hidden mismatch into an audible one. That second habit is the subject of a separate post on listening drills that use the repeat-it-back method, and it pairs naturally with the description order above.
The broader lesson reaches past the game. A team that keeps having the same misunderstanding is usually not careless. Its members are checking words against different pictures. The post on how to be a better communicator using lessons from a bomb turns that into a short set of habits for meetings and calls.
Which words fail at the bomb?
The words that fail depend on the speaker's viewpoint or memory: "this one", "that", "the other one", "left" with no starting point, "squiggly", "same as before", and any word that sounds like another. They feel clear to the person saying them. They give the listener nothing to look up.
Pointing words. "This", "that" and "here" only work when both people can see the thing. At a bomb, only one of them can. Replace each one with a count and a position.
Invented pictures. "The little house" or "the one like a fish" works once, between two people who happen to share the image. It breaks the moment a second Expert joins in or the same picture fits two glyphs. The Strange Symbols mission says it directly in its blurb: agree on names for the glyphs. The manual's own pattern, outer shape plus inner mark, gives you names that already match the page.
Shades of color. "Dark blue", "pinkish" and "sort of orange" force the Expert to guess which named color you mean. The wire colors are red, blue, yellow, white and black. Use those words and nothing in between.
Sound-alike words. The Word Grid page warns that many of its words sound alike and tells Experts to ask the Defuser to spell every word. The Word Salad mission's blurb gives the same advice. Spelling feels slow, and it is still faster than a strike.
Ambiguous characters. The serial number is six characters, and the last one is always a digit. The manual's edgework appendix notes that the letters I and O are never used, so a round character is a zero and a straight one is a one. Knowing that saves an argument about whether the Defuser is looking at a letter or a number.
Memory words. "Same as last time" refers to a moment the Expert may have spent on a different page. Repeat the actual facts instead, even if you said them a minute ago.
The communication mistake: describing while the Expert is still reading
The common mistake is talking over the page turn. The Defuser starts listing details while the Expert is still finding the module or reading a rule, so the details land on nobody. The fix is a ready signal: wait until the Expert says they are on the page, then describe once, in order.
The Wires page shows why this matters. The Expert has to find the right list and then read its lines from the top, stopping at the first one that is true. That is a reading task. If the Defuser talks through it, the Expert either loses their place or stops listening, and usually both. The facts then have to be said again, which costs more time than the pause would have.
A ready signal can be tiny. The Expert says "Wires page, go" and the Defuser starts. With several Experts, the manual suggests splitting the pages between them so two people never search for the same thing, and the ready signal comes from whoever owns that module. Everyone else stays quiet until that Expert asks for help.
The same rule runs in reverse. An Expert who starts reading an instruction while the Defuser is still describing has stopped listening to the description. If something important comes at the end, such as the color of the last wire, it gets missed. Let the description finish, then speak.
If you want to try this today, open the standard manual for rule seed 1 on one device and the bomb on another, then pick the First Bite mission from the play page. It has one module and plenty of time, and its own description says what it is for: learning to describe what you see. Run it once talking freely, then again with the shape, color and position order, and listen to how many fewer questions come back.