The fluorescent lights in Meridian Interactive’s basement data center hummed at precisely sixty hertz—a frequency Aidan Voss had learned to hate during his first week as a junior data analyst. The sound burrowed into his skull like a mechanical mosquito, threading through the server racks that stretched in endless parallel rows toward a darkness the ceiling fixtures never quite reached.
He checked his watch. Half past eleven. The custodial staff had finished their rounds forty minutes ago, and the security guard on the night shift wouldn’t bother walking the subbasement until two. That left him approximately one hundred and fifty minutes of unsupervised access to the Lucky Stars retention database.
“Lucky Stars,” he muttered, scrolling through the user activity logs on his terminal. The name was a joke. The only stars involved were the ones users saw when they drained their bank accounts chasing jackpots that the algorithm had already predetermined they would never hit.
Aidan had been hired three months ago as part of Meridian’s diversity initiative—a phrase the HR director had delivered with a practiced smile that never quite reached her eyes. He’d worn his only suit to the interview, a charcoal-gray off-the-rack number his mother had insisted on paying for with money she didn’t have. “You’re going to work in Ashwick City,” she’d said, pressing the garment bag into his hands. “The people there, they notice things like fabric quality.”
She wasn’t wrong. On his first day, a senior developer named Thorne had glanced at Aidan’s cuffs and said, “Polyester blend. Smart choice. Easy to clean.” The words had been delivered without malice, which somehow made them worse. Thorne genuinely believed he was being kind.
Now Aidan sat alone in the subbasement, tracing his finger across the screen where a graph plotted user retention against time. The curve was a gentle downward slope at first, almost flat, suggesting a healthy product with predictable churn. Then, at approximately the seventy-two-hour mark, it dropped off a cliff.
The death zone.
Users played for three days. They won enough virtual chips to feel clever, lost enough to feel challenged, and then—somewhere between the fourth login and the first real-money purchase prompt—they simply stopped. The app sat on their phones like a forgotten lottery ticket, its push notifications increasingly desperate, until finally the user deleted it and moved on to the next free-to-play distraction.
The problem, Aidan understood, wasn’t the game. The game was beautiful. Meridian’s designers had created a slot machine experience so tactile, so responsive to touch, that users reported feeling genuine physiological arousal when the reels aligned. The problem was the moment right before the purchase decision—that tiny synapse of hesitation where the rational brain asserted itself and asked whether spending nine dollars and ninety-nine cents on a pile of glowing pixels was actually worth it.
He pulled up the psychometric profiles Meridian had purchased from a data broker in Lindsfarne. The files contained predictive models of user behavior based on thousands of variables: time of day, session length, tap velocity, even the micro-hesitations between button presses that revealed cognitive load. Aidan had spent his undergraduate years at Ashwick State University studying behavioral economics under Professor Emilia Kade, who had once described consumer choice as “a bridge built on toothpicks and crossed in fog.” He understood, perhaps better than anyone else at Meridian, how little separated a paying customer from a churning statistic.
The insight came to him at twelve minutes past midnight, while reviewing a single user’s session recordings. User ID 884721, female, sixty-four years old, location: Garrick Parish. She had played Lucky Stars for two hours and forty-seven minutes on her first day, reaching a near-miss jackpot on spin number eighty-three. The near-miss was, of course, entirely illusory—the reels didn’t actually “almost” land; they were programmed to display specific combinations that mimicked near-wins because decades of gambling research proved those moments were more addictive than actual victories.
User 884721 had experienced eight near-misses in her first session. Each one accelerated her tap rate by approximately fourteen percent, a classic dopamine-seeking behavior. Then, at spin number one hundred and twelve, she had paused. The cursor hovered over the “Buy More Chips” button for exactly four point three seconds. Then she closed the app and never returned.
Four point three seconds.
Aidan leaned back in his chair, staring at the ceiling tiles with their water stains and dead insects. Four point three seconds was not an eternity. It was barely a breath. But somewhere in that window, User 884721 had remembered something—a rent payment, a grandchild’s birthday gift, the quiet dignity of a fixed income—and chosen to walk away.
What if, Aidan thought, the window could be narrowed? What if the near-miss experience could be calibrated to each user’s psychological profile, triggered at exactly the moment their defenses were weakest, their dopamine receptors most receptive? What if the purchase prompt didn’t appear as a choice but as an inevitability, the natural conclusion of a sequence so seamless the rational brain never had time to engage?
He began coding at twelve forty-seven in the morning. His fingers moved across the keyboard with the fluency of a second language, calling up machine learning libraries and behavioral tracking functions he’d been explicitly told were for internal analytics only. The algorithm took shape in fragments: a variable called rewardGap that measured the distance between expected and actual outcomes, a function called tiltDetection that identified moments of emotional vulnerability based on tap pressure and session duration, a subroutine called microLoss that triggered minuscule virtual chip deductions—so small users wouldn’t consciously register them—that cumulatively pushed them toward the purchase threshold.
By three in the morning, he had a working prototype. He called it Erebus, after the Greek primordial deity of darkness, because naming things made them real and naming them honestly felt like the least he could do.
He ran Erebus against the last six months of Lucky Stars user data in simulation mode. The results appeared on his screen in clean, unemotional columns of numbers.
Projected revenue increase: thirty-seven percent.
Projected first-purchase conversion among previously churning demographics: forty-two percent.
Projected average revenue per daily active user: up from $4.83 to $11.27.
He should have felt triumph. Instead, watching the simulated users—represented as anonymous dots on a scatter plot—cluster ever more tightly around the purchase nodes, he felt something closer to vertigo. Each dot was a person. Each person had a User 884721 inside them, a moment of hesitation, a bridge between impulse and consequence that he was systematically dismantling.
Aidan saved the Erebus file to his private directory, encrypted behind a password only he knew. Then he walked to the break room, poured himself a cup of stale coffee, and stared at his reflection in the darkened window.
The face staring back was still the face from Lowmarsh—the cheekbones that had hollowed during the winter his father lost his factory job, the slight asymmetry of the jaw that marked him as someone whose parents couldn’t afford orthodontics. But the eyes were different now. They had acquired something during those three months at Meridian, something he recognized but couldn’t name.
He thought about his mother, working double shifts at the Garrick Parish laundromat, her hands perpetually cracked from industrial detergent. He thought about the eviction notice that had arrived last week for the family in Unit 14B, the one with the three children who sometimes played football in the hallway. And he thought about the Meridian executives on the thirty-seventh floor, who spoke of “conversion funnels” and “lifetime value optimization” as though the words were abstractions, as though the money that flowed upward from millions of small, desperate purchases had no origin point, no human cost.
At six forty-five in the morning, Aidan returned to his terminal and prepared a presentation for his supervisor. The title slide read: “Erebus: A Predictive Retention and Monetization Framework for Casual Gaming Applications.” He used the company’s approved template, muted colors and clean sans-serif fonts. He included case studies, confidence intervals, a sensitivity analysis that demonstrated Erebus could be tuned to different regulatory environments.
He did not include the simulation that showed, with ninety-four percent confidence, that Erebus would increase problem gambling behaviors among users in the bottom income quartile. He had run that simulation at four in the morning, then deleted the results.
The meeting with his supervisor, a vice president named Calder Rourke, took place at nine. Rourke was a heavyset man with the puffy skin of a reformed drinker and the dead-eyed stare of someone who had long ago stopped distinguishing between right and profitable. He listened to Aidan’s presentation without interruption, his fingers steepled beneath his chin.
When Aidan finished, Rourke was silent for a long moment. Then he said, “You built this on company time?”
“On my own time, sir. Nights and weekends.”
Rourke nodded slowly. “Good. That avoids some intellectual property complications. Not all of them, but enough.” He leaned forward, his chair creaking under his weight. “Here’s what’s going to happen. You’re going to hand over all the code to my team. You’re going to sign a document stating you developed Erebus as a work-for-hire. And you’re going to receive a promotion to Senior Data Architect, effective immediately, with a forty percent salary increase and eligibility for the executive bonus pool.”
Aidan felt his pulse quicken. “And if I refuse?”
“Then you’ll be terminated for unauthorized use of proprietary user data, which is a firing offense with cause. No severance. No reference. And given the non-compete clause you signed, you won’t work in gaming for three years.” Rourke smiled. It was not an unkind smile. It was the smile of a man explaining the rules of a game he had long ago mastered. “This isn’t a negotiation, Aidan. It’s an opportunity. Take it.”
Aidan signed the document at ten fifteen. By noon, his access credentials had been upgraded, and a secretary was arranging his move to a cubicle on the twenty-second floor, where the windows actually opened and the coffee came in single-serve pods instead of industrial urns.
At two thirty, he called his mother from the privacy of a stairwell.
“I got promoted,” he said.
“Oh, Aidan.” Her voice cracked with pride. “I knew it. I told Mrs. Ellerby just last week, my son is going places.”
He closed his eyes and pressed his forehead against the cold concrete wall. “Yeah,” he said. “Places.”
That evening, he walked through the glass doors of Meridian Interactive’s Ashwick City headquarters and stepped onto the rain-slicked pavement of Varick Street. The financial district gleamed around him, towers of glass and steel reflecting the purple-orange dusk. Investment bankers in thousand-dollar shoes hurried past, shouting into wireless earpieces. A woman in a camel coat walked a dog wearing a rain jacket that cost more than Aidan’s entire wardrobe.
He bought a bottle of whiskey from a shop in the Galleria, a mid-range label he’d seen Rourke drinking at company functions. At the register, the cashier glanced at his Lowmarsh address on his driver’s license and did a double-take. “Long way from home,” she said.
“Not anymore,” Aidan replied.
In his studio apartment that night, he poured himself a glass and opened his laptop. The Erebus file was gone from his private directory, absorbed into Meridian’s proprietary codebase, but he had saved one thing the company didn’t know about: the original simulation results, including the deleted problem-gambling projection.
He didn’t know why he’d kept it. Evidence, perhaps. Insurance. A receipt for his soul, should anyone ever ask to see it.
He drank the whiskey and stared at the numbers until they blurred. Somewhere in the data, User 884721 was still out there—unaware that a stranger had studied her hesitation like a safecracker studying a lock, unaware that the next game she downloaded might not give her four point three seconds to change her mind.
Aidan closed the laptop and went to bed. Tomorrow he would start his new position. Tomorrow he would smile at his new colleagues and learn their coffee orders and pretend he belonged among them.
But tonight, in the darkness of his apartment, with the whiskey warm in his stomach and the rain tapping against the window, he dreamed of the Lowmarsh housing projects. In the dream, he was a child again, walking through the dim stairwell where the lights were always broken. The walls were plastered with eviction notices, hundreds of them, thousands, each one bearing a different name but the same final line: You have thirty days to vacate the premises.
At the top of the stairs, where the roof access door should have been, there was instead a massive slot machine, its reels spinning endlessly. And on each reel, instead of cherries and sevens, there were faces—his mother, his father, the children from Unit 14B, User 884721—spinning and spinning, never quite aligning, never quite stopping.
He woke at three in the morning, gasping.
On his nightstand, his phone glowed with a notification from the Meridian internal messaging system. Calder Rourke had sent a file labeled “ER-001 Production Deployment Schedule.” The first line of the attached document read: Target rollout to Lucky Stars user base: Tuesday, November 14th. Full integration with Fortune’s Loop: Pending successful pilot results.
Fortune’s Loop. The new project everyone on the twenty-second floor had been whispering about. The one targeting the casual senior demographic. The one with the loot-wheel mechanics that skirted so close to gambling regulations that even Meridian’s legal team had expressed concerns.
Aidan stared at the message for a long time. Then he turned off the phone, rolled over, and tried to find his way back to sleep.
But the dream was waiting for him, patient as a debt collector, and morning was still hours away.
No comments yet. Be the first to comment!