What it is. You put too many simultaneous choices in front of the player at once. They can’t evaluate them, so they freeze, stall, or pick at random — and your carefully-tuned depth reads as noise.

Why it hurts. A choice the player can’t reason about isn’t a choice — it’s a tax. The first minutes decide whether they keep playing; a wall of unexplained options at the door reads as “this game is homework.” Depth that arrives before the vocabulary to parse it is indistinguishable from chaos.

The smell

Cognitive overload at the point of decision. Options outnumber the player’s ability to compare them, or arrive before they have the concepts to weigh a single one.

How it happens

• Everything unlocked day one. The full skill tree, the whole shop, all twelve units — visible before any of them mean anything.
• Flat, unranked lists. Twenty items with no default, no recommendation, no sort — every one demands equal scrutiny.
• Undifferentiated options. Choices that all look and cost about the same, so nothing anchors the eye (see antipattern-fake-choice — paralysis and fake-choice are cousins: too many that matter vs. many that don’t).
• Naked numbers. +3 STR / +2 DEX / +4 INT with no idea what a point buys — the player can’t convert stats into a decision.
• No off-ramp. No “recommended,” no “auto,” no way to defer. The only exit is to commit blind.

The tell (check your OWN design)

• Playtesters hover — the cursor drifts across the menu for 10+ seconds with no click.
• Players ignore the menu entirely: skip the shop, dump points into whatever’s leftmost, close the screen unread.
• You catch yourself writing a tooltip to explain why the screen is fine rather than fixing the screen.
• New players and veterans open the same screen and the veteran is faster by an order of magnitude — the layout only works once you already know the answer.
• You can’t name the one thing the screen wants the player to decide. If there are five co-equal decisions, it’s five screens.

The fix

Reduce what’s live at once, not what exists total. Depth stays; exposure staggers.

Staggered reveal is system-onboarding‘s core job — pace introductions so each choice lands with vocabulary already in hand. The curated-slice move is a whole design lever: see pattern-meaningful-choice for making the 3 you do offer distinct, and system-inventory-and-ui for laying them out legibly. Keep the presentation calm — pattern-readability and pattern-restraint-and-negative-space are what let a big menu breathe.

Twist seams

• Full complexity, but rationed (twist: unlock cadence). anchor-slay-the-spire shows only 3 cards per reward and one relic at a time — the whole 300-card space exists, you just never face it at once. Compare a naked “here’s the collection, build a deck” screen: same depth, opposite feel.
• Many options, but one obvious default (twist: guided freedom). anchor-civilization‘s advisor and recommended build let a new player click “yes” through a turn while a veteran micromanages the same menu — the screen serves both by ranking, not hiding.
• Overwhelming board, but read one glance at a time (twist: staged legibility). anchor-into-the-breach hands you a fully-solved-information puzzle every turn, yet telegraphs each threat so cleanly you parse it in seconds — paralysis dissolves under system-telegraphs, not fewer pieces.

Seen in…

• Well-managed: anchor-balatro and anchor-slay-the-spire — draft of 3–5, never the whole space; the deck deepens without the screen deepening. anchor-hades gates boons behind rooms so the build assembles one legible pick at a time. anchor-reigns is the extreme cure — every choice is two options, swipe left or right. anchor-vampire-survivors level-up offers a draft of three — a small slice of the full weapon/passive pool, never the whole roster.
• Where it bites: deep 4X and grand strategy (genre-4x, anchor-civilization without its advisors) open the tech tree and city menu at once and lose players in the first hour. Character creators that front-load a dozen stat sliders before the player knows the game. RPG skill trees (system-skill-trees) shown fully at level 1. Any genre-deckbuilder that skips the draft and hands over the full collection cold.

Verify / guard

Frontmatter points at docs/VERIFICATION.md. This is a design-review gate, not an assert: before handoff, open every choice screen and ask “how many decisions is this asking for, and does the player have the words for each?” Watch a first-time playtest for hover-stalls and ignored menus. Adjacent failure modes to check in the same pass: antipattern-fake-choice (options that don’t matter), antipattern-feature-soup (too many systems, upstream cause), antipattern-endless-tutorial (the over-correction — staggering so hard the player never gets to choose at all).