ROVIX CREATIVE INTELLIGENCE · 2026
A study of playable ad strategy across 58 mobile titles
Contents
This is a study of when that decision is correct. It is not an argument against it. It answers a narrower and more useful question: given a specific game’s economy and media plan, should it advertise with a mechanic it doesn’t contain and if so, which one?
Estimates throughout are approximate and directional.
THE FINDING
Minutes, not weeks
Trained on your game
Built for iteration
Find the loop your game can own.
We’ll work through your playable strategy - including when not to borrow.
FIRST, DEFINE THE FANTASY
Core fantasy is not core gameplay. A 4X survival game’s core gameplay is tapping buildings and waiting out timers. Its core fantasy is arriving in a frozen world, gathering what you can, and building walls around people who depend on you. The mechanics are what players do; the fantasy is why they stay. Borrowed loops compress the fantasy, not the mechanics.
FAMILY 1 · ~134.1B IMPRESSIONS
Movement loops, for games about growing something
Your game is about growing something over time - resources, an army, your own power. The borrowed loop is that same growth, sped up, with the waiting taken out
FAMILY 2 · ~459.2B IMPRESSIONS
Puzzle loops, for games about solving something
Your game is about solving a problem that sits in front of you: someone in danger, a place that needs fixing, a board that needs clearing.
FAMILY 1 · MOVEMENT LOOPS
Match the borrowed loop to the progression players already want to watch accelerate.
Resources → construction
Gathering materials, then building
Arcade idle
Joystick around an isometric map. Auto-collect trees and rocks, carry a stack to a dashed outline, then unlock the next area.
Unit mass → army scaling
Growing a force until it wins
Gate runner
A lane with +5, ×3 and −2 gates. Steer the crowd, grow or shrink the force, and end in a clash against a larger army.
Comparative power
Picking fights you can win
Free-roam gauntlet
Joystick around a map of numbered enemies. Absorb the smaller, avoid the larger, and watch your own number climb.
FAMILY 2 · PUZZLE LOOPS
When the core fantasy resolves in one screen, give the player the problem immediately.
01 · EXTRACTION
Save someone
A character is in danger - lava, a flood, a collapsing building. Pull the right pin and they live. Pull the wrong one and they don't. Getting it wrong is the point: it turns someone watching an ad into someone who wants another go.
02 · RESTORATION
Fix something
A ruined room, a damaged garden, a wrecked house. Pick what to repair first. The before-and-after is the payoff.
03 · ARRANGEMENT
Tidy something
Sort items into crates. Drop blocks to clear a grid. Order the pieces until the board resolves.
THE RULE
One fantasy, one loop. Two fantasies, two creatives.
Last War: Survival is the clearest case. Its lane-squad combat is showable, so it ships that as a native gate runner at roughly 5.6 billion impressions. Its base-building layer is timer-gated and can't be shown, so it borrows arcade idle for that at roughly 1.6 billion. Both work, both are large, and they sell different halves of the same game.
Puzzles & Survival does the same thing, running two entirely different borrowed loops side by side: a gate runner for its troop-battle side and an arcade idle loop for its base.
WHEN BORROWING FAILS
The same production template can produce radically different outcomes. These cases show what happens when the borrowed loop and the game’s actual core fantasy fall out of alignment.
Homescapes
Renovation · ~40M
Fishdom
Narrative · ~84M
Gardenscapes
Restoration · ~130M
Township
Accumulation · ~6.9B
Negative results deserve to be recorded.
WHERE THESE LOOPS ACTUALLY LIVE
Where these loops actually live.
None of these loops are inventions. Arcade idle is a commercial genre with its own successful games - My Perfect Hotel, Pizza Ready, Eatventure, Burger Please. For them the loop is simply the game: the same joystick, proximity collection, deposit-to-build and blueprint outlines. The gate runner has the same story through its counting-gate runner genre.
The puzzle loops have a longer lineage. Pin-pull extraction puzzles have been running in mobile advertising since roughly 2019, most famously for hero-collector and 4X titles that contain nothing resembling them. That they are still running, at scale, six years later is itself a finding - this is not a trend, it is infrastructure.
So the accurate description is not that midcore games invented fake mechanics. It is that live casual genres’ loops travel to games whose real core fantasy can’t be shown and stay there for years, because they keep working.
MARKET SHIFT
The shift already underway in the West.
For most of the period this data covers, borrowed loops were a Chinese and Asian studio practice. The Western and European 4X and strategy titles we examined ran no playables at all - including several with exactly the core fantasy structure that suits one.
That is changing now.
Playrix, Western-operated, is already deep in the format — four titles running borrowed arcade idle creatives, and restoration puzzles at billions of impressions. Supercell runs native playables at tens of billions across two titles. Western studios that have committed to the format are not experimenting at the margins; they are running it at the top of their creative mix.
And in the window this study covers, a Western studio approached us to build an arcade idle playable modeled directly on a top-grossing Asian 4X title: a category that, until recently, would not have considered the format at all. We can’t name them, but the brief is the point: they had watched Chinese studios in their genre nail this, and wanted the same capability.
Find the loop your game can own.
We’ll work through your playable strategy - including when not to borrow.
DECIDING THIS FOR YOUR OWN GAME
Deciding this for your own game
Everything above resolves into three questions, asked in order. Most of the expensive mistakes in this dataset come from skipping one.
NO
Playables aren’t your channel yet - this is a media-mix question, not a creative one.
YES
Can a player see your core fantasy in 30 seconds?
YES
Ship a native slice
NO
Borrow a loop that compresses your fantasy
1. Are you buying SDK inventory at meaningful volume?
If not, that is the first thing to revisit. Across every genre in this study, the category leaders are on this inventory — and the studios that have committed to it are running playables at the top of their creative mix, not as an experiment at the edges. Nine genres, nothing in common between them, same conclusion. If your media plan doesn’t touch SDK networks, the question isn’t which loop to build. It’s whether you’re leaving a channel untested that your closest competitors have already scaled.
2. Can a player see your core fantasy in thirty seconds?
If yes - we’d like to talk to you. That is the cheapest and most reliable playable to build, and it is also where fidelity matters most. A native slice only works if it genuinely looks like the game: Clash of Clans’ base, Coin Master’s slot machine, Last Day on Earth’s actual environment art. A rough approximation of your game is a worse ad than no ad, because the install arrives expecting something you don’t have.
Why fidelity matters
We train models on your own build so the playable inherits your art direction, lighting, and UI language - then hand you the platform to generate variants from it yourself.
3. If not, name your fantasy before you name your loop.
Do players stay because something of theirs is growing, or because there’s a problem in front of them worth solving? Growing points to a movement loop. Solving points to a puzzle loop. The most common expensive mistake in this dataset: a narrative game building a movement loop because that is what the category leader runs.
Then check whether you have a second fantasy. Last War passes the thirty-second test on one half of its game and borrows for the other, running both at billions of impressions. If you’ve only ever tested creatives for the showable half, you may be advertising half your game.
Be prepared to be wrong several times.
The studios that win are able to test enough variants that being wrong doesn’t hurt.