What it is. The dominant strategy is the dull one: grind, turtle, kite, wait. Skilled play and exciting play diverge, and the numbers side with tedium.

Why it hurts. The player learns the fun move is a tax — flashier means worse. So they stop taking it, the game they wanted to play evaporates, and what’s left is a spreadsheet they solve on autopilot. The fantasy dies not from a missing feature but from an incentive pointed the wrong way.

The smell

The line a mastery player converges on is the line a bored player would also pick. Optimal and fun are unrelated, or worse, opposed. Watch for the phrase “yeah, it’s boring, but it’s just better.”

How it happens

• Safety pays more than risk. The aggressive option and the passive option have the same expected value — but the passive one has lower variance, so it wins every long-run calculation. See pattern-risk-reward.
• Time is free. If nothing costs the clock, waiting/grinding/farming is a pure-upside strategy. No pressure, no reason to stop. See system-resource-loops.
• The flashy tool is strictly worse. The cool mechanic (mechanic-parry, mechanic-combo-string, the risky combo) does less than the safe poke, so it’s cosmetic. See system-combat-model.
• Attrition beats expression. Whittling from max range is safe and slow; committing is fast and risky. Nothing rewards the commit, so the whittle is correct.
• Metagame collapse. One build/opener out-values all others; every choice upstream funnels to it. This is the endgame of antipattern-solved-metagame.

The tell (check YOUR design)

• Simulate the optimal player. Are they doing the thing your trailer shows? If the trailer is a lie, you have this bug.
• Ask: “What’s the most boring thing that still wins?” If it wins comfortably, it dominates.
• Is your safe line also your high-EV line? Then risk is antipattern-fake-choice — it never pays to take it.
• Does any strategy scale with patience instead of skill? Patience isn’t gameplay; it’s a wait.
• Would a speedrunner and a scared player pick the same route? If yes, the route has no texture.

Twist seams

• A turtle strategy, but the shield decays every second it’s raised (twist: turtling is possible but on a timer). Punish the wait; make defense a resource, not a state. See system-status-effects.
• A grind loop, but yields halve on repeat and the boss heals while you farm (twist: the clock is no longer free). Efficient play now demands commitment. See pattern-pacing-and-tension.
• A safe poke, but every hit builds a meter only aggression spends (twist: safety funds the flashy line but can’t cash it). The exciting move becomes the payoff, not the tax. See system-reward-schedules.

The fix

Pay out the fun line; tax the degenerate one. The exciting play must be the correct play, or at least a live equal.

Design principle: asymmetric payout. The boring line should cap low; the exciting line should have a higher ceiling and pull its weight at the median. Aggression that only pays at expert level is still a tax for everyone else.

Seen in…

• Fighting games — the “throw a fireball, walk back, repeat” zoning that beats everyone who won’t approach. genre-fighting-game fixes this with chip damage, timers, and meter that rewards commitment; anchor-street-fighter tuning lives or dies here.
• Turn-based RPGs — spamming the basic attack because MP is precious and the flashy spell is net-negative. system-resource-loops that make big moves worth their cost break the spam.
• RTS turtling — walling up and macroing forever when the map has no pressure. anchor-starcraft taxes it with map control and expansion economies; contrast a mode with no clock.
• Idle grind in RPGs and genre-incremental — farming a safe early zone because it’s the highest gold-per-risk, ignoring the whole game. See antipattern-grind-wall.
• anchor-slay-the-spire counterexample — the Ascension ladder and event costs keep passive “just survive” decks from dominating; risk (elite fights, curses for power) is priced to pay off.
• anchor-into-the-breach counterexample — no grind, no hoarding; every turn forces a committed, legible decision. Waiting is never an option because the threat is always scored and visible.

Verify / guard

This is a FUN failure — the mechanic works but the incentive misfires. Prove the fun line pays: model or playtest the optimal strategy and confirm it’s the one you want played. See design/FUN.md for the incentive-audit gate, and cross-check with pattern-mastery-and-flow — if mastery converges on tedium, the payout curve is inverted. Related smells to rule out together: antipattern-fake-choice, antipattern-solved-metagame, antipattern-grind-wall.