How to play 4Stack

The rules take thirty seconds. Not burning all four mistakes in the first minute takes a bit longer.

The rules

You're shown sixteen cards. They belong to four hidden groups of four, all drawn from gaming culture — franchises, hardware, esports history, streamers, memes, studio history, speedrunning jargon. Your job is to find all four groups.

Tap four cards, then press Submit.
Get it right and the group locks in, revealing its category name.
Get it wrong and you lose one of your four mistakes. If you're one card away, the game tells you — but not which one.
Spend all four mistakes and the board solves itself in front of you.
One new puzzle every day at midnight, the same for everyone.

Finishing a puzzle gives you a spoiler-free emoji grid you can share, and keeps your streak alive. Streaks, stats and results are stored in your own browser — there's no account to make.

What the colors mean

Once a group is solved it takes a color, and the colors are a difficulty ranking set when the puzzle is written:

Yellow — the group most people spot immediately. Usually a plain, well-known category.
Green — straightforward if you know the subject, invisible if you don't.
Blue — needs a specific piece of knowledge, or a second look at something you thought you'd placed.
Purple — the trick. Often a wordplay group, a technicality, or a connection that only exists once you know the history.

The useful consequence: the purple group is usually where the cards you're most confident about are hiding. A puzzle is built by picking the hard group first and then choosing obvious-looking categories that steal cards from it.

A worked example

Here's a real board — 4Stack #4, which ran on July 17, 2026. The sixteen cards were:

SNES · N64 · GameCube · Switch · Genesis · Dreamcast · Saturn · 32X · Virtual Boy · Wii U · Ouya · Stadia · Game Boy · PSP · Vita · Steam Deck

Everyone sees the same first move: that's a pile of Nintendo, a pile of Sega, and some handhelds. So you grab Game Boy, PSP, Vita and… Virtual Boy, because the Virtual Boy is a handheld. Submit. Wrong. One mistake gone, and you don't know why.

The reason is that this board has four categories and three of them can claim the Virtual Boy. Work through the overlaps before you touch Submit:

Virtual Boy — a Nintendo console, arguably a handheld, and one of the most famous flops in the industry.
Wii U — a Nintendo home console, and a flop.
32X — Sega hardware, and a flop.
Vita — unambiguously a handheld, and a commercial disappointment by any measure.

Four cards, each of which fits "commercial flops" — but a group only holds four, and Ouya and Stadia are on the board too. Ouya and Stadia are the tell: neither of them fits anywhere else. They aren't Nintendo, they aren't Sega, and neither is a handheld. So the flops group already has two locked members and only two slots left, which means at most two of Virtual Boy, Wii U, 32X and Vita can join them.

Now count the other groups. Sega hardware needs four: Genesis, Dreamcast, Saturn — and 32X is the only fourth available. So 32X is Sega, not a flop. Nintendo home consoles needs four: SNES, N64, GameCube, and either Switch or Wii U. Handhelds needs four: Game Boy, PSP, Steam Deck, and either Vita or Virtual Boy.

That leaves exactly two seats in flops for exactly two of {Wii U or Switch} and {Vita or Virtual Boy}. Switch is not a flop and Vita has a stronger claim to "handheld" than Virtual Boy does — so Wii U and Virtual Boy take the two remaining flop seats, Switch completes Nintendo, and Vita completes handhelds. The whole board resolves without a single guess. See the full breakdown →

Strategy that actually helps

Solve the constrained group first, not the obvious one. Find the cards that fit only one category — like Ouya and Stadia above — and build outward from them. The obvious group is obvious because it's been baited.
Count to four before you submit. If five cards fit your category, you haven't found the category — you've found the trap. Work out which one belongs elsewhere first.
Say the category out loud. If you can't name it in a few words, it probably isn't a real group. "Nintendo home consoles" is a category. "Old systems, sort of" is a hunch.
Look for the second axis. When a board is all people, "is a streamer" tells you nothing — everyone streams now. Sort by where they came from instead. That's the entire trick of puzzle #2: Ninja, Shroud, summit1g and Dyrus were competitive players long before they were famous broadcasters.
Spend your first mistake on purpose. If two arrangements are genuinely tied, submit the one that tells you the most either way. A "one away" response on a four-card guess is real information; a random swap later is not.
Watch for the technicality group. Purple is often pedantic — mods that became games, developers rather than publishers, things named after a person. If three cards share a surface theme and the fourth doesn't quite, the fourth is usually right and your theme is usually wrong.

If you're stuck right now

The hints page is built for exactly this. It gives you today's category nudges in escalating order — a vague pointer first, a named group later — and keeps the answers behind a separate click, so you can take just enough help to get moving. Yesterday's full answers live there too.

Once a puzzle is finished it gets its own answers page with a written breakdown of every group and what the board was doing to you. And the archive lets you play any previous puzzle — archive plays don't affect your streak, so it's a consequence-free place to practise reading traps.

That's everything. Play today's puzzle → · More about 4Stack