Why Do Duplicate Drops Feel Like a Bug? The Merging Loop Fix

The first time I pulled a five-star duplicate in a gacha game I cared about, I felt the floor drop out.

Days of grinding for the premium resource. Animation, rare music cue, card flip. Then the small “OWNED” tag in the corner. The dopamine cliff was so steep I closed the app and didn’t open it again for a week. The merging loop is the design pattern that prevents this exact moment, and once you see it, you start spotting it in every collectible system that actually works.

Six months later I was playing Best Fiends on a flight, getting duplicate after duplicate of the same insect, and I caught myself doing the opposite thing. Hoping for duplicates. Reaching for them. Same psychological situation, opposite emotion.

The difference wasn’t the player or the genre. It was a single design decision built into the second game and missing from the first. That decision is the subject of this post.

Author Credibility: Yu-kai Chou

Yu-kai Chou is an S-Tier Behavioral Designer and the creator of the Octalysis Framework, the gamification design system now applied to products and experiences reaching over 1.5 billion users. His book Actionable Gamification is one of the most-cited works in the field, and he has been ranked the #1 Gamification Guru in the World.

He has advised MrBeast, LEGO, Microsoft, Porsche, Tesla, Stanford, Harvard, and governments including Ukraine on turning behavioral psychology into product mechanics that actually change user behavior.

Verify: Wikipedia · Google Scholar · Wikidata · LinkedIn

The merging loop is one of the design patterns I built into Octalysis Prime, the live gamified learning platform I run for behavioral designers. We spent months iterating on the GeoMod collectible system, including a 48-hour scramble to patch an economy exploit that one of our players, Simon, found within days of release. The case study you’ll read below is not a hypothetical. It is the design log of a system I shipped, broke, and rebuilt, which is why the failure modes in this post are specific instead of generic.

The Duplicate Disappointment Problem

Almost every collectible game I have analyzed in the last five years has the same emotional bug.

You set up a chase. The player invests time, premium currency, or attention to earn a chance at a rare drop. They open the chest, the loot box, the gacha pull, the booster pack. The animation builds. They hit a rare. Triumphant moment.

Then they see the second one.

“Already owned.”

That two-word notification is the cliff. Your player just did everything right and your system told them, “thanks, that one was worth nothing.” It is not a soft disappointment. It is the precise moment a positive emotion gets reversed mid-feeling, which is one of the strongest negative learning signals the brain knows how to encode.

I see this play out in product analytics constantly. The biggest churn spike in collectible systems is rarely “player can’t progress.” It is “player progressed, then the system told them their progress didn’t count.” Whether the asset is a Pokémon card, an NFT, a Marvel Snap variant, a hero in Honkai: Star Rail, or a sticker on a fast-food rewards app, the curve looks identical. Get the rare drop, feel great. Get the second, abandon the loop within a session or two.

The naive design fix is to compensate the player. Most games hand out shards, dust, fragments, or tokens that accumulate and can eventually be spent on a new card. That is better than nothing. It also signals to the player that duplicates are not really rewards. They are partial refunds. Once you have framed your drop as a refund, you have lost the dopamine moment forever.

This is the gap the merging loop fills.

The Merging Loop: A 3:1 Psychological Inversion

The merging loop is one rule with a lot of behavioral physics behind it.

Three of the same item combine into one item of the next tier up. That is the entire mechanic. Three commons become one rare, three rares become one epic, three epics become one legendary. The math is intentionally simple because the real work is being done somewhere else: inside the player’s head.

Watch what happens at the duplicate moment now.

Before the loop, the second drop reads as: “this is a duplicate, the system wasted me.” Disappointment.

After the loop, the same drop reads as: “two-thirds of the way to my next legendary, one more and I merge.” Anticipation.

The emotional sign flips because the meaning of the second drop changed. The prize is identical, the drop rate is identical, the probability table is identical. What changed is that the drop is no longer an end-state. It is an ingredient in a goal the player can see, count, and finish. That last word is the load-bearing one. A merge target is countable, which means the player’s brain can run a finishability simulation on it the moment the duplicate hits.

This is not a soft preference dressed up as theory. The duplicate-disappointment pattern is well documented in behavioral economics under loss aversion, the asymmetry where losses feel roughly twice as heavy as equivalent gains. A duplicate isn’t an objective loss, but emotionally it gets coded as one because the player expected variety and got redundancy. The merging loop reframes the same event from “loss of variety” to “progress toward a higher tier,” which moves it from the loss column to the gain column without changing the underlying drop.

I want to stress how rare it is for a single rule to swing emotional valence this hard. Most “psychological” design tweaks shave off 5 percent of churn or add 8 percent to session length. The merging loop turns a quit-trigger into a pull-trigger, which is closer to inverting a sign than nudging a number.

Why It Works: The CD3 + CD4 + CD2 Stack

If you have read Actionable Gamification, you know the Octalysis Framework breaks human motivation into eight Core Drives. The merging loop is not powerful because it activates one of them. It is powerful because it stacks three at once.

Core Drive 3: Empowerment of Creativity & Feedback. Every duplicate drop is now a decision. Merge the three rares I have? Save them and chase a rarer rare to merge into an epic? Sell one for currency I need now? The system stops handing players outcomes and starts handing them choices, which is the engine room of Core Drive 3 (CD3): Empowerment of Creativity & Feedback.

Core Drive 4: Ownership & Possession. Items players already own gain a second life. A common card stops being a “filler.” It becomes a quarter of a future epic. Inventory acquires optionality, and optionality is a form of ownership the player did not have to spend anything to acquire. This activates Core Drive 4 (CD4): Ownership & Possession retroactively, on assets that were already in the bag.

Core Drive 2: Development & Accomplishment. Most importantly, the merging loop creates a visible progress ladder. The player can see exactly how many merges separate them from the next prestige tier, and they can chip away at it asynchronously. This is Core Drive 2 (CD2): Development & Accomplishment in its purest form: countable progress toward a named goal.

Most designers I review get one of these three when they try to fix duplicates. They add a dust-conversion system (CD4 only) or a transparent rarity ladder (CD2 only). The merging loop hits all three because one action, the act of collecting a duplicate, simultaneously gives the player a decision (CD3), upgrades their existing inventory (CD4), and ticks a visible progress counter (CD2).

That triple-stack is the reason a one-rule mechanic can carry an entire collection economy.

Octalysis Prime Case Study: GeoMods and the Cell Exploit

I will be specific about what shipping a merging loop actually looks like, because the abstract version makes it sound easier than it is.

On Octalysis Prime, our community platform for behavioral designers, we built a collectible system called GeoMods. Each GeoMod is a cosmetic creature attached to a player profile, and they drop at four rarity tiers: Common (30 percent appearance rate), Rare (15 percent), Epic (5 percent), and Mystical (1 percent). Three of any tier merge into one of the next tier up. Standard 3:1 ladder.

The behavioral fingerprint we wanted: turn the second-drop disappointment into a “one more and I can merge” anticipation, especially at the higher tiers where rarity made dupes more likely. It worked. Engagement on the collectible loop did exactly what the theory predicted.

Then a player named Simon broke our economy.

His exploit was elegant. First, collect every GeoMod type. Sell a Mystical for 5,000 to 10,000 in-game coins. Spend the coins on a “Trap” item, a deterministic shop unlock that guarantees a GeoMod the player doesn’t currently own. Receive the Mystical back. Repeat indefinitely to print infinite coins.

The cell that broke us was the intersection of three rules, each of which was reasonable on its own. Selling rare items for high coins makes the economy feel valuable. Letting players target what they don’t own reduces dupe pain. The merging loop creates rarity at the top tier. Stack the three and you have a closed loop where rarity converts to coin and coin converts back to rarity, indefinitely.

The fix was almost embarrassingly small once we saw it. We added one rule: you cannot sell your last copy of a GeoMod type. You must have two or more before any single instance is sellable. That single constraint snipped the loop, because the targeted unlock could no longer be reset by the sell step.

The lesson generalizes. If you build a merging loop, you are also building a sell mechanic, a target mechanic, and a rarity gradient. Players will find the cell where those three intersect. Audit your economy before launch by asking, “what is the closed loop where merging plus another action prints free resources?” The answer will exist. Find it before Simon does.

Modern Examples: From Best Fiends to Marvel Snap

The reason I trust this design pattern is that I keep seeing it converge in successful 2024 and 2025 product launches that almost certainly never read my book. Convergent design is the strongest evidence a principle is real.

Royal Match and the Match-3 dynasty. Dream Games’ Royal Match has been the highest-grossing mobile puzzle title in much of 2024 and 2025, with broad reporting confirming over a billion dollars in lifetime revenue. The level loop is match-3, but the meta-game underneath is merging power-ups: combine two TNTs and you get a bigger explosion, combine a rocket and a lightball and you clear the board. Duplicate booster drops, which would feel useless in a flat collection model, become the input to a multiplicative payoff curve. Players save them, hoard them, plan around them.

Genshin Impact and the Constellation system. miHoYo’s Genshin Impact, one of the most-played and most-studied gacha titles in the world, took the duplicate problem and turned it into a six-level vertical ladder per character. Pull the same five-star a second time and you don’t get “already owned.” You get a Constellation upgrade that unlocks a new ability or numerical buff. The loop is the same shape as a merging loop: redundancy gets reframed as vertical progression on the asset the player already owns. Whales and free-to-play players alike have a clear, named ladder to climb on every duplicate.

Marvel Snap and Variant collecting. Marvel Snap, the card duel from Second Dinner that won mobile Game of the Year in 2022 and is still pulling top-grossing weeks in 2025, encodes the same idea inside its variant system. Owning a card once is the floor. Subsequent variants of that card upgrade through Common, Uncommon, Rare, Epic, Legendary, Ultimate, and Infinite, each upgrade fed by Boosters that drop when you play. The system formalizes “the second drop matters” as a literal upgrade economy, which is why Marvel Snap retains players past the typical CCG drop-off cliff.

Best Fiends and Merge Mansion at the puzzle layer. Both Seriously’s Best Fiends and Metacore’s Merge Mansion run merge mechanics as the core verb of the game, not just the meta. The fact that Metacore reportedly grew Merge Mansion past $250 million in revenue on a single design verb is the most dramatic statement I can make about how much load this pattern can carry. When the merge becomes the primary input loop instead of the secondary disappointment-fix, the collectibility starts compounding.

Different teams, different genres, different audiences, and yet all of them landed on a 3:1 or vertical-merge pattern as the way to keep collection systems from leaking emotion at the duplicate moment. That convergence is the principle.

Decision Framework: When to Use a Merging Loop

The merging loop is a strong mechanic, which means it is also easy to misapply. Before you bolt one onto your product, run it through this checklist.

1. Are duplicates a structural feature of your reward system? If your players keep getting duplicates because your drop pool is small or your variance is high, the merging loop is the right intervention. If duplicates are rare because your catalog is huge and rarity is low, you have a different problem and the merging loop will not fix it.

2. Do players have a clear visual of the next tier? The merging loop only works if the player can see what the upgrade leads to. A merge target you have to discover is a merge target you don’t chase. Show the next-tier silhouette, the rarity color ramp, the named tier (“Mystical,” “Legendary,” “Constellation 2”). Visibility is the entire activation surface.

3. Is your tier ratio honest? A 3:1 ratio is the sweet spot because it feels like a quest the player can finish in one or two sessions. If you set a 5:1 or 10:1 ratio to slow your economy, you have not built a merging loop. You have built a savings account, and savings accounts have completely different emotional physics. The whole point of the loop is that the second duplicate already carries momentum. A 5:1 ratio kills the momentum.

4. Have you stress-tested the closed economy? Per the GeoMod story, your merging loop will interact with your sell economy and your target-purchase economy, and the intersection will leak. Run the closed-loop test before you ship. The question is not “is there an exploit.” The question is “what is the exploit and what is the smallest rule that closes it.”

5. Does the upgrade actually feel different? A Mystical that is just a Common with a new color border is not an upgrade. It is a graphic. The next tier needs to feel different in a way the player can name. A new ability, a new animation, a new role in your meta, a new social signal. If the merge result is cosmetic, your loop will run for two tiers and quit.

If you can answer yes to all five, build the loop. Anything less, and you should fix the underlying structure first. A merging loop bolted onto a broken collection system will only make the failure mode larger and more expensive.

Three Mistakes That Kill the Loop

I have reviewed enough collection systems to predict where teams break their own merging loop. The three failure modes below are the ones I see most.

The “any rarity” merge. Some teams loosen the constraint so any three rares can merge into an epic, regardless of which rares they are. This sounds player-friendly. It is the death of the loop. The reason the merging loop creates emotional pull is that the second duplicate carries specific information: “I’m two-thirds of the way to this upgrade, not generally accumulating.” The “any rarity” merge collapses the emotional distance between drops back to zero, which is the exact disease the loop was supposed to cure.

The infinite ladder. If your top tier is also a merge ingredient, the player never reaches a finish line. Every Mystical they earn is just three more away from a Hyper-Mystical, then a Super-Mystical, then a Nightmare-Mystical. After tier four or five the prestige stops feeling like prestige. Cap the ladder at a tier the player can actually reach in a meaningful playtime, and put the meta-progression somewhere else (achievements, profile titles, ladders that aren’t items).

The forced merge. Some products auto-merge duplicates the moment a third lands, because they read the loop as a convenience feature instead of a decision. This is the worst version of the mistake. The decision is the feature. Taking the merge action away from the player removes the CD3 activation and downgrades the entire loop into an inventory cleanup tool. Always let the player press the merge button, even if it adds a redundant step. The redundant step is not friction. It is agency.

Each of these mistakes shares the same root cause: the team understood the math of the loop but not the meaning. The math is “three become one.” The meaning is “the player chooses, the asset they own gains optionality, and the next tier feels reachable.” If you optimize for the math you will save engineering hours. You will also lose the part of the feature that made anyone care.

Related Reading

• Collection Set Design: The Complete CD4 Guide: the horizontal counterpart to the vertical merging loop, and when to use which.
• The Octalysis Framework: the full eight Core Drives behind every example in this post.
• Why Loyalty Programs Fail: The Mercenary Trap: what happens when reward systems forget the meaning behind the math.

Bring Behavioral Design Into Your Product

If you are designing a collectible system, a loyalty program, or any reward economy where duplicates are a structural reality, the merging loop is one of the most reliable patterns I can hand you. The full toolkit lives in Actionable Gamification, and the live training where I walk teams through these decisions session by session lives at Octalysis Prime. If you want me to apply this lens to your product directly, my consulting team operates through The Octalysis Group.

Frequently Asked Questions

What is a merging loop in game design?

A merging loop is a collectible system rule where three or more copies of the same item combine into one item of the next rarity tier. The mechanic transforms duplicate drops from a disappointment into a progress event, because the second and third copies of an item are now ingredients in a visible upgrade goal.

Why does the 3:1 ratio work better than 2:1 or 5:1?

Three is the floor for emotional weight. A 2:1 merge feels transactional and removes the chase, because every second drop is also a finish line. A 5:1 or higher merge feels like a savings account, because the goal is too far for the second drop to carry momentum. Three duplicates is far enough to feel earned but close enough to feel reachable, which is why successful merging systems converge there.

How is a merging loop different from a collection set?

A collection set rewards horizontal completion: gather one of every item to fill the album. A merging loop rewards vertical progression: combine duplicates of the same item into a higher tier. Collection sets treat duplicates as filler. Merging loops treat duplicates as fuel. Many of the strongest live games use both, with collection sets framing the meta and merging loops keeping the moment-to-moment loop emotionally positive.

Which Octalysis Core Drives does the merging loop activate?

The merging loop activates three Core Drives at the same time. CD3 (Empowerment of Creativity & Feedback) shows up in the merge-or-hold-or-sell decision. CD4 (Ownership & Possession) shows up because the items already in inventory gain optionality and a future use. CD2 (Development & Accomplishment) shows up in the visible progress ladder toward the next tier. The triple-stack is why a one-rule mechanic can carry an entire collection economy.

What is the most common mistake when designing a merging loop?

The most common mistake is auto-merging duplicates the moment a third lands. This treats the merge as a convenience feature and removes the player’s decision, which kills the CD3 activation that makes the loop feel rewarding. Always let the player press the merge button, even if it appears to add a redundant step. The decision is the feature.

WOULD YOU LIKE YU-KAI CHOU TO WORK WITH YOUR ORGANIZATION?