I missed the wave of PIXEL in mid-March. An old brother in the group took a screenshot - for the same target, he bought in at the bottom and multiplied it by nearly three times. I didn't follow. That night, I closed the trading software and stared at the ceiling for a full forty minutes, feeling sour.
After feeling sour, I was unwilling to accept it, so I opened my computer to figure out what fundamental factors were supporting such a rise. I hadn't looked into it, but once I did, it tore open all the old wounds from my six years of user acquisition backend work in Web2.
When I was at a big company, my daily job was to help the company spend money. Who did we spend it on? Google, Meta, TikTok. For a mobile game to survive, user acquisition costs must account for at least 30% to 50% of revenue. We internally referred to this money as "the share money".
Anyone who has driven a taxi or knows taxi drivers understands this term. No matter if you picked up a passenger today or not, no matter if it’s pouring rain or an empty city, you owe the company a fixed fee every day upon opening your eyes. If you can't pay, you lose your car and license. What game companies spend on UA budgets for advertising platforms is essentially this fee—platforms make guaranteed profits, while studios bear their own risks.
As a true passenger sitting in the back seat (player), you paid the full fare, but the destination of the fees has nothing to do with you. The money earned by the advertising platform is invisible to you.
Flipping to 2:30 AM, I found that entire chapter in the Pixels white paper (Ecosystem Flywheel & Data Loop). By the time I reached section five, my fingers were stiff.
The original text of the white paper states: "Our models retrain nightly." (The models retrain every night.)
Such an inconspicuous line of commentary has hardly been deconstructed across the internet. But I have written backend recommendation algorithms for ad placements at big companies, so I know exactly what this phrase means.
The traditional Facebook advertising system also employs machine learning, but its model iteration cycle is measured in "weeks" or even "months" because the data between advertisers is isolated, and the platform will not feed Company A's conversion data to Company B's model. Each studio is fumbling around in their own little black box.
But the Stacked engine is completely different. The white paper clearly states that the Pixels Events API will record all purchases, tasks, transactions, and extraction behaviors of all games and players, forming a primary dataset of cross-game behaviors. Then, this dataset is fed back into the model for retraining every night.
Translated into plain language: Under the traditional fee model, your taxi company (advertising platform) has no idea who is sitting in the back seat, where they want to go, or whether they are willing to tip. It only collects your fees and randomly assigns passengers to you. The Stacked of Pixels is like a smart dispatch center connected across the city; it not only knows each passenger's commuting times, spending habits, and historical ride trajectories, but also refines the complete data of all drivers and passengers every night, making each order dispatched the next morning more precise than the day before.
This implies an extremely terrifying fact: Traditional advertising becomes increasingly expensive, as good users experience fatigue from repeated exposure, and CPM (cost per thousand impressions) rises year by year. However, Stacked's nightly retraining cycle continuously enhances its precision in placements. The white paper states, "Leakage to extractors falls; real players get higher quality incentives"—the waste flowing to exploitative players is continuously decreasing, while the quality of incentives received by real players is continuously increasing.
So that 131% RORS (Return on Reward Spend) is not a static number. It is a compound curve that refreshes automatically every 24 hours. That's what sends chills down my spine.
The "zero-prepayment API" of Stacked has recently undergone public testing. What does this mean? Game studios integrating with Stacked do not need to fork out a large UA budget upfront as if paying a fee. The system only automatically dispenses rewards after players complete specific high-value actions (like clearing dungeons or leveling up to key levels). No spending occurs without triggering.
It's been six years since I assisted my boss in a large company in spending tens of millions of budget on Facebook's Ads Manager daily, always following the logic of "pay first, see results later, and take the blame if results are poor." Stacked has completely reversed this logic: see the real player behaviors first, and then settle based on the results. The fees are reset to zero.
Where did the exorbitant tolls that were previously intercepted by the advertising platform go? The white paper is clear: "instead of Facebook or TikTok ads"—they directly convert into on-chain UA Credits, targeted to real contributors filtered by AI. Every effective game interaction you have, every action that retains for over seven days, helps the entire ecosystem save the fees that would have otherwise gone to Silicon Valley giants. The money saved is the extra rewards in your wallet.
Missing out on the March market is indeed frustrating. But after feeling frustrated, I realized one thing: I'm not watching the price fluctuations of a farming game; I'm observing an advertising distribution system that quietly gets smarter every night, slowly dismantling the fee system that has sustained the gaming industry for twenty years into scrap.
Logical disconnection. This is common sense.