Skills / break-game-theory
Exploit Catalog
Copy page
Strategies that break app mechanics, grouped by what the mechanic does. Use the sections that apply to the players and rewards in your Phase 1 map. Each row is something real users do; the signal says what to look for in the code or rules; the fix is the usual mechanism change.
Use fake accounts, example.com emails, and test payment modes for every reproduction.
Every reward given per account is a reward given per identity-cost. Find out what a new account costs.
Strategy Who does it Signal in code or rules Fix pattern Email aliases (name+1@, dots in Gmail, catch-all domains) Anyone after a signup bonus Uniqueness on raw email string Normalize emails; but treat as a speed bump, not a wall Disposable email services Trial cyclers No domain checks Grant the valuable part later (on payment, on usage), not on signup Same card, new account Trial cyclers Card fingerprint not checked across accounts One trial per payment fingerprint Self-referral Credit farmers Referrer and referee can be the same person Pay referrer on referee's payment or real activation, not signup Sybil accounts to vote, review, or rank Brigades, sellers, creators Votes weighted equally regardless of account age or history Weight by account history; cap influence of new accounts Account sharing (one seat, many people) Teams, agencies, families Pricing per seat, no concurrency check Concurrent session limit, or price on a unit sharing doesn't break (workspace, usage) Account selling (aged or high-reputation accounts) Spammers, sellers Reputation transfers with the account Reputation decays without activity; re-verify on big changes Smurfing (expert on a fresh account) Competitive players Matchmaking starts new accounts at average Fast rating convergence; place by early performance
Strategy Who does it Signal in code or rules Fix pattern Farming signup credits Multi-accounters Credits granted at signup, worth more than identity cost Smaller upfront, more on activation; credits that can't be transferred or resold Converting free credits to cash or resale Arbitrageurs Credits buy something resellable (API calls, gift cards, compute) Restrict what free credits buy; expire them Hoarding then dumping Rational savers Credits never expire, rewards accrue indefinitely Expiry or decay; caps on balance Double redemption via concurrent requests Script users Check-then-write without lock or unique constraint Unique constraint, idempotency key, transaction Refund after reward Anyone Reward granted on purchase, not clawed back on refund Vest rewards past the refund window; claw back on refund Reward for action with no value (empty posts, junk uploads) Reward farmers Reward on count, not quality Reward on downstream signal (someone else used it) Milestone gaming (exactly hit the threshold, then stop) Rational users Step-function rewards Smooth rewards, or rewards that require sustained behavior
Strategy Who does it Signal in code or rules Fix pattern Trial cycling Price-sensitive users Trial per account, cheap accounts Trial per payment fingerprint or org domain; limit trial value, not length Cancel to trigger win-back discount, resubscribe Rational subscribers Win-back offer to anyone who cancels Offer once per customer lifetime; don't make cancellation the cheapest path to a discount Downgrade timing (use annual features, downgrade before renewal) Power users Features stay after downgrade until period end with exports Fine if intended; make sure data access matches plan Plan arbitrage (buy the plan whose limit is easiest to stretch) Teams Limits on one axis, value on another Limit on the axis that tracks value Quota splitting across accounts Heavy users Per-account quota, cheap accounts Per-org or per-payment quota Burst at quota reset Script users Fixed windows reset at midnight Sliding windows or token buckets Exceeding quota via parallel requests Script users Quota checked before work, counted after Reserve quota atomically before work Resale of your API through a proxy Arbitrageurs Unmetered or flat-rate plan with programmatic access Usage-based tier above a fair-use line; ToS plus detection
Strategy Who does it Signal in code or rules Fix pattern Retaliation reviews Sellers, hosts, buyers One side sees the other's review before writing theirs Simultaneous reveal (both publish together or after a window; Airbnb has done this since 2014) Rating inflation (everyone gives 5 stars to avoid retaliation or awkwardness) Everyone Ratings all cluster at max Relative or multi-dimensional ratings; private feedback channel Review swaps and rings Sellers, creators Reviewers who only review each other Graph detection; weight reviews by reviewer history Review-for-reward Sellers Incentivized reviews look like organic ones Disclose or disallow; weight verified purchases Bury bad reviews with new listings Sellers Reputation tied to listing, not seller Reputation on the seller Karma farming in low-effort spaces Status seekers Karma from any upvote anywhere Karma scoped to where it was earned Mass-reporting to remove a competitor Griefers, rivals Auto-action after N reports Weight reporters by accuracy history; human review before removal
Strategy Who does it Signal in code or rules Fix pattern Vote brigading Communities, fans Raw vote counts, no velocity or source checks Weight by voter diversity and history; dampen sudden coordinated bursts Engagement bait (outrage, cliffhangers, "comment YES") Creators Ranking on raw engagement Rank on downstream satisfaction (saves, return visits), not just reactions Sandbagging (stay in a lower bracket to win it) Competitive players Brackets by self-reported or early stats Brackets by recent performance Time-of-day gaming (post when competition is low) Creators Ranking favors recency in absolute terms Normalize for traffic; fine if harmless Leaderboard squatting (one big early lead nobody can catch) Early users All-time leaderboards Seasonal or rolling windows Self-dealing in "trending" Platform insiders, big accounts Trending based on a few big accounts' actions Count distinct actors, not volume
Strategy Who does it Signal in code or rules Fix pattern Take it off-platform after the first match Buyers and sellers Fee per transaction, contact info shareable Make the platform worth the fee (payments protection, reputation, scheduling), not just block contact info Shill bidding or fake demand Sellers Seller can bid via another account Link accounts; hide bidder identity patterns Bait-and-switch listings Sellers Listing can change after purchase or booking Snapshot listing at transaction Cherry-picking (accept only the best jobs, decline the rest) Supply side Supply sees full job details before accepting Show less before accepting, or reward acceptance rate Adverse selection (only the worst risks opt in) Users of insurance-like features (guarantees, protection) Opt-in protection priced flat Price by risk, make it default, or bundle Moral hazard (careless because someone else pays) Protected users Full coverage, no deductible Partial coverage; something at stake Cold-start fakery (seed fake supply to attract demand) The company Fake listings or profiles Seed with real subsidized supply instead; fake supply is a trust bomb
Strategy Who does it Signal in code or rules Fix pattern Sniping (bid in the last second) Bidders Hard close Soft close: extend on late bids (the Amazon vs eBay auction comparison in Roth & Ockenfels, 2002, shows hard closes drive late bidding) Waiting for the inevitable discount Customers Predictable sales or end-of-quarter deals Fewer, less predictable discounts; reward early commitment Front-running a visible queue Fast or automated users Order visible before it executes Batch, randomize, or hide until execution Racing the drop (bots on limited releases) Resellers First-come-first-served Raffle or lottery; verified-human queue Streak protection hoarding Engaged users Streak freezes accumulate without limit Cap freezes Fake activity to keep a streak Engaged users Streak counts any trivial action Streak counts the meaningful action
Payoff isn't money: spite, rivalry, or fun. The test is the ratio of damage to effort.
Strategy Who does it Signal in code or rules Fix pattern Spam invites or mentions Trolls No rate limit or consent for being added Rate limits; consent before adding Reserve and abandon (hold inventory, slots, usernames) Rivals, squatters Free holds with no expiry Short holds, deposits Vandalism of shared spaces Trolls Anyone can edit, no history or revert History and one-click revert; trust levels Report abuse Rivals Auto-actions on reports See Reputation Cost dumping (trigger expensive work on someone else's bill) Malicious users Actions that cost the target money (sending SMS, running jobs) Charge the actor, or require target opt-in
Strategy Who does it Signal in code or rules Fix pattern Win trading (take turns winning) Competitive players Same pairs matched repeatedly Randomize matching; detect repeated pairs Price coordination between sellers Sellers Public pricing plus repeated interaction Usually a legal issue, not a product one; flag it Group buying a solo deal Groups Per-account discount on shareable goods Per-household or per-payment limits Referral rings Groups A refers B refers C refers A Graph checks on referral trees
"When a measure becomes a target, it ceases to be a good measure." Run this on the team's own metrics, not just users.
Metric How it goes up while the goal goes down Better pairing Daily active users Notifications and streaks that pull people in for nothing DAU paired with task completion or retention Time in app Slower flows, infinite feeds Time to complete the job Signups Lower friction, lower intent, more junk Activated accounts Tickets closed Closing without solving Reopen rate Messages sent Spammy prompts to message Replies received Conversion on a funnel step Dark patterns, pre-checked boxes Refund rate, churn after conversion Lines of code or PRs merged Smaller, noisier changes Cycle time with defect rate NPS Asking only happy users Response rate and who was asked
The company is in the game too. When the company profits from the user losing, users eventually notice, and trust is the thing that breaks.
Strategy Signal Why it backfires Hard-to-cancel flows Cancellation needs a call or six screens Chargebacks, regulator attention, reviews Auto-renew without reminder No reminder before annual renewal Refund requests, churn at next renewal Drip pricing (fees revealed at checkout) Price on listing ≠ price at payment Abandoned carts, trust Credits that silently expire Expiry not shown in the balance Users feel robbed; they stop buying credits Rankings that favor paying sellers without disclosure Paid placement looks organic Buyers stop trusting rankings Data lock-in (no export) No export, or export in unusable formats Users don't commit in the first place
Name the game in the report when it helps the team see the shape.
Game Shape Shows up as Prisoner's dilemma Each does better defecting, both do worse if both defect Rating inflation, review retaliation, price wars Tragedy of the commons Shared resource, private benefit, shared cost Unlimited free tiers, shared quotas, notification spam Free rider Benefit without contributing Account sharing, team plans where one person pays Adverse selection Hidden information drives the wrong people in Flat-priced protection, "unlimited" plans attracting the heaviest users Moral hazard Hidden action once protected Full guarantees, no deductible Sybil attack One actor, many identities Votes, referrals, reviews, trials Sniping / last-mover advantage Fixed deadline rewards waiting Auctions, limited offers, end-of-period bonuses Chicken-and-egg Each side waits for the other Marketplaces at launch, network-effect features Cobra effect Paying for the symptom grows the cause (from the story of a bounty on cobras that led to cobra breeding) Bounties per bug found, rewards per item reported Goodhart's law The metric drifts from the goal once targeted Any KPI tied to a reward
Ordered from most to least preferred. Higher ones make honesty the best strategy; lower ones police it.
Change the payoff. Reward the outcome you want, not a proxy for it (pay on activation, not signup; on usage, not upload count).Change the timing. Vest, delay, or claw back rewards past the window where an exploit is reversible (refund windows, trial ends). Soft-close deadlines.Change the information. Simultaneous reveal, hidden bidder identities, hiding the exact threshold, not showing supply the full job before they accept.Change the unit. Price and limit on something sharing or splitting doesn't break (workspace, usage, payment fingerprint).Cap and decay. Caps on balances, expiry on credits, seasonal leaderboards, decaying reputation.Raise identity cost proportionally. Ask for more verification only where the reward is worth more; never make every honest user pay for a rare abuser.Detect and act. Graph checks, velocity limits, reviewer weighting. Needed sometimes, but it's an arms race; it should back up the mechanism, not replace it.Terms of service. A rule only in the ToS is a rule only honest users follow. Back it with one of the above.