Behind the Play
The game loop was already there. It was hiding inside one of the most ordinary frustrations there is: looking for something you know you own.
“What if boring things could be fun?”
The idea arrived while digging through a box of LEGO bricks, trying to find one exact piece. Weeks later I caught myself doing the same thing at the bottom of a sock drawer, hunting the sock that actually matched.
Different objects.
Same interaction.
The searching is the annoying part. Repetitive, slightly humiliating, something you want to be over. Nobody has ever hoped the search took longer.
That instinct isn’t new for me. It had shown up before in ideas with nothing to do with searching — chore cards that pair a household task with a silly constraint, a grocery run rebuilt as a scavenger hunt. The recurring question is the same: where is the playful possibility hiding inside something mundane?
The game loop was already there.
Searching already has the structure of a simple game loop. Not because it’s compelling by nature — it usually isn’t — but because of one specific combination that happens to produce tension for free.
“I know what I’m looking for. I know it’s here. I just don’t know where.”
The player knows the target exists and knows exactly what they need. Nothing about the goal is ambiguous.
They have no idea where it is. That gap between knowing what and not knowing where is the only thing generating tension.
The target exists, and you can picture it.
An information gap opens. Only searching closes it.
Every object you move might be the one.
The gap narrows a little. Back around the loop.
loop back to anticipation
The gap closes instantly. That’s the whole payoff.
The moment players start doubting whether the target is even in there, anticipation curdles into frustration and the round feels unfair. That’s also why losing has to end with a reveal: showing where it was still resolves the information gap, so a lost round ends in “it was right there” rather than in doubt.
“I didn’t have to invent a game loop. I had to notice the one hiding inside an everyday frustration.”
The action stays simple. What your mind is doing changes.
Every game below is the same verb: look through a pile and find the thing. What differs is which demand carries the round — spotting, remembering, recognising, inferring, predicting.
- SPOTvisual search
- REMEMBERwhere have I already looked?
- RECOGNISEdoes this match the target?
- INFERwhat does this new information tell me?
- PREDICTdo I already know the answer?
Brick Bin
Rummage a physics pile of loose bricks for one exact piece.
- visual search
- where have I already looked?
- does this match the target?
Difficulty is similarity, not just quantity: same colour / wrong shape, same shape / wrong colour. Look-alikes are slow to reject, so every brick you shove aside is a small “maybe it’s behind this one”.
Junk Drawer
The household drawer of cables, keys, coins and dead pens.
- visual search
- where have I already looked?
- does this match the target?
Everything is familiar, so recognition is instant — the difficulty is clutter and occlusion. Objects bury each other, and moving one exposes information you didn’t have. Keeping track of where you’ve already been is half the game.
Haystack
The same needle, every single round.
- visual search
- where have I already looked?
- does this match the target?
Maximum certainty about what, maximum uncertainty about where. Nothing new to learn — only signal against rising noise: density, camouflage, occlusion. Lose the round and the reveal shows exactly where it lay: “it was right there.”
Letter Box
Dig letters out of the pile; each one fills a slot in a hidden word.
- visual search
- does this match the target?
- what does this new information tell me?
- do I already know the answer?
It starts as visual search and quietly turns into inference. Every letter found shrinks the space of possible words, until the thought flips from “find the next letter” to “wait — I think I know it.” Guess Word lets you act on that prediction; being wrong costs time.
Shape Box
Plain solids in flat colours — no props, no story.
- visual search
- does this match the target?
A target is a shape and a colour at once, so the pile fills with same-shape/other-colour and same-colour/other-shape decoys. With nothing to recognise, it isolates the pure feature-matching part of searching. Fast, clean, almost reflexive.
Sock Basket
Coming soonA developing experiment, not a finished game.
- visual search
- where have I already looked?
- does this match the target?
There is no unique target object — you’re shown one sock and must find its partner, so the answer is a relationship. That turns one search into repeated pairwise comparison, and the fabric behaviour is still being tuned. It stays in the workshop until it feels right.
A social loop on top of the anticipation loop.
The searching mechanic doesn’t change at all when someone else joins. What changes is that your uncertainty is now shared with a person who might resolve theirs first.
It gives away nothing useful — not the letter, not the position, not the word. It only tells you she is closer than you are. That is enough to make time feel expensive: she knows something you don’t, and the round suddenly has a pace you didn’t set. Each player digs their own pile toward the same hidden word, and the first correct guess ends it.
The race
In the object hunts everyone gets the same pile and the same target on one shared countdown. First to find it takes the round — and you can hear the other person not finding it either, which turns out to be most of the fun.
Letter Box co-op
Playable today alongside VS: one word, shared progress, everyone digging their own pile. The quick searcher feeds letters in; the person who is good at words predicts the answer early. Different strengths, one answer — discovery becomes something you hand to each other instead of race for.
Joining is a short game code or a link — no account, no download. The takeaway wasn’t “add a leaderboard”: another person changes what the activity is. Attention becomes shared, failure becomes a joke, and “there it is!” is worth more out loud than the point it scores.
Almost no tutorial required.
“Borrow the interaction from real life. Change what it feels like.”
Every game deliberately borrows an action people have already practised for years: rummage through a pile, hunt a matching sock, search a haystack for a needle, scan a crowded scene, dig in sand, notice the missing piece. The core verb needs almost no explanation, because everyone has spent real time looking for keys, socks and pieces.
That isn’t a claim of zero onboarding — controls, difficulty and party setup still need signposting. But the part that usually costs the most explanation, “what am I supposed to be doing?”, arrives already understood. The design work moves to the other half of the line: changing what the familiar action feels like.
One constraint, many verbs.
“Search” is not one verb. It’s rummage, scan, dig, match, infer, uncover, collect. These are designed but not yet playable — each one takes the same loop and swaps the verb.
Sand & Gold
DigSweep material away and hope. Displaced sand refills behind you, so where you’ve searched stays readable — and plenty of glints turn out to be nothing.
Bouquet
CollectNot one target but a recipe. Pick the listed flowers and leave with something you assembled.
Missing Piece
InferA jigsaw inverted: the hole comes first. Work out what belongs, then go find it — searching with a prediction you formed yourself.
Find the Toddler
ScanNo pile at all. A crowded, moving scene and a small human who won’t stay put.
Some ideas aren’t games at all — a flashlight, fog, a table that won’t stop shaking. Those are dials that drop on top of any hunt, which is exactly what a well-chosen constraint is supposed to give you.
How this gets designed.
- Observe an everyday frustration.
- Notice the loop already hiding inside it.
- Constrain the design space to that one loop.
- Vary what creates tension and what kind of thinking it asks for.
- Make it playable and tune it by playing.
- Add a social layer.
- Ask what else the same constraint can generate.
Technically it’s a React app with 3D scenes and rigid-body physics in the browser, plus lightweight realtime multiplayer, built through an AI-assisted workflow. The design decisions — difficulty curves, how heavy a pile feels, how much clutter is fun before it’s mean — are human-directed and settled by playing. Plenty of prototypes were thrown away for failing exactly that test.
What Search Party actually is.
A free collection of quick search-and-find games that runs in the browser. Each game gives you one target and one messy 3D pile — bricks, junk-drawer clutter, hay, letter cubes or simple shapes — and you rummage with your finger or mouse until you find it before the clock runs out. The piles are physics-driven, so objects tumble, bury each other and can be shaken loose.
Playing with friends: one player creates a party and shares a short game code or link. No account, no sign-up, no download.
Price and ads: free to play, no ads, no in-app purchases.
Devices: any reasonably modern desktop or mobile browser with WebGL. Built touch-first.
Playable today: Junk Drawer, Brick Bin, Letter Box, Shape Box and Haystack. Sock Basket and everything labelled coming soon is still in the workshop.