The Variable He Couldn't Control
title: "The Prediction Problem" wordCount: 2501
Patricia Chen's pen stopped mid-signature on the term sheet, her eyes locked on the market projection slide I'd just presented, and she said the word that would haunt me for weeks: "Impossible."
The conference room went quiet. Morning light slanted through floor-to-ceiling windows, turning the polished table into a mirror that reflected twelve faces—eleven investors and me. James Whitmore leaned back in his chair at the head of the table, fingers steepled, watching.
"I'm sorry?" I kept my voice level. My left hand found the burn scar on my right palm, thumb tracing the puckered skin.
Patricia set down her pen with deliberate care. "These projections. The smartphone market penetration rates, the specific quarters when adoption curves accelerate, the exact price points where consumer resistance breaks." She tapped the slide with one manicured nail. "You're predicting a 340% growth rate in Q3 2009, then a plateau, then another spike in early 2010 when—and I'm quoting your deck here—'a major manufacturer releases a device that fundamentally changes user expectations.'"
"That's the thesis, yes."
"Mr. Chen." She smiled, but her eyes stayed cold. "I've been in this business for fifteen years. I've seen a thousand pitches from Stanford kids who think they've cracked the code. But this?" She gestured at the screen. "This isn't analysis. This is a script."
Whitmore shifted forward. "Patricia, I think—"
"No, James. Let me finish." She turned back to me. "Either you have insider information from Apple, Samsung, and every other major player in mobile—which would make this meeting very short and very unpleasant—or you're making educated guesses and dressing them up as certainty to impress us."
My Stanford hoodie felt too warm. I'd worn it three days straight, and now I could smell the stale coffee and anxiety baked into the fabric. "Here's the thing—these aren't guesses. I've run the numbers on consumer behavior patterns, manufacturing cost curves, network infrastructure buildout timelines. The data points to—"
"The data." Patricia laughed, sharp and humorless. "I lost eight million dollars last year betting on 'the data.' A mobile gaming company that had every indicator pointing toward success. Solid team, proven market, perfect timing." She paused. "You were an intern at the VC firm that passed on that deal. Three months later, the company folded."
The room temperature dropped ten degrees.
"I don't—" I started.
"Greylock Ventures. Summer 2007. You wrote a memo recommending they decline investment in MobilePlay Interactive." Her voice was surgical now, cutting through my defenses. "That memo cited concerns about user retention and monetization that wouldn't become apparent for another six months. How did you know?"
Because I'd watched MobilePlay burn through forty million in funding before collapsing in 2008, taking Patricia's reputation with it. Because I'd lived through this already. Because I'd come back to fix it.
I couldn't say any of that.
"I did my research," I said. "Talked to developers, looked at early user data—"
"You were twenty-two years old." Patricia stood. "James, I appreciate the introduction, but I'm not signing anything today. Mr. Chen, I suggest you think very carefully about what you're really selling here, because if it's information rather than innovation, there are people who will want to know where you're getting it."
She walked out. Two other investors followed.
Whitmore waited until the door closed. "Well. That was unfortunate."
"She thinks I'm committing securities fraud."
"She thinks you're too accurate." He pulled out his phone, typed something. "Which, to be fair, you are. I've been in this room for a hundred pitches, Marcus. Most founders hedge. They say 'we believe' or 'market indicators suggest.' You said 'will happen' fourteen times in twenty minutes."
"Because I'm confident in the analysis."
"Or because you know something." He looked up from his phone. "I'm not accusing you. I'm observing. And I'm telling you that confidence without visible uncertainty makes people nervous. It makes them ask questions." He stood, buttoned his jacket. "I'm still interested. But you need to decide how much you're willing to reveal about your methodology, because right now it looks like magic, and nobody trusts magic in this business."
He left me alone in the conference room with my laptop and my projections and the sick feeling that success was going to be harder than failure.
The parking garage smelled like oil and exhaust. I was halfway to my car when David stepped out from behind a concrete pillar, hands in his pockets, looking like he'd been waiting.
"That went well," he said.
"You were listening?"
"I was in the lobby. Saw Patricia storm out." He fell into step beside me. "She's got a reputation. Smart, ruthless, holds grudges like they're investment assets."
"Great."
"Here's the thing, though." David stopped walking. "You need someone who can build what you're describing. The predictive algorithms, the data infrastructure, the whole technical stack. I can do that."
I'd heard this pitch before. David had approached me in the original timeline too, right after a different disastrous meeting. I'd turned him down then, built everything myself, and spent six months debugging code that he could have written in six weeks.
"I work alone," I said.
"Yeah, I noticed. How's that going?" He pulled out his phone, showed me a screen. "I've been following your GitHub. The neural network architecture you're using for pattern recognition? It's elegant, but it's not scalable. You're going to hit a wall around ten thousand data points, and then the whole thing collapses."
He was right. I'd hit that wall in March 2009, lost three weeks rebuilding the entire system.
"I can fix it," I said.
"Sure. In six months, after you've burned through your runway and Patricia Chen has convinced half of Sand Hill Road that you're either a fraud or a security risk." He pocketed his phone. "Or you can bring me on, we build it right the first time, and you focus on the business side instead of debugging memory leaks at three in the morning."
"Why do you want this so badly?"
Something flickered across his face. "My brother. Younger brother. He's got acute lymphoblastic leukemia. The treatment's experimental, not covered by insurance. I need money, you need technical expertise. We can help each other."
The honesty caught me off guard. In the original timeline, David had mentioned his brother once, months after we'd started working together, and only because I'd asked why he was working hundred-hour weeks.
"How old is he?"
"Fourteen. Smart kid. Wants to study physics." David's voice stayed level, but his hands were fists in his pockets. "The treatment costs two hundred thousand. I've got forty saved. The rest..." He shrugged. "I'm running out of time."
I thought about Lily. About the mechanic shop and the brake lines and the things I'd done to keep her alive. About the calculations I'd made, the variables I'd optimized, the way I'd treated people like problems to solve.
"I'll think about it," I said.
"Don't think too long. Patricia's not going to stop with walking out of a meeting. She's going to dig, and when she digs, she's going to find things that don't add up." He started walking away, then turned back. "And Marcus? Whatever you're really doing, whatever you actually know—you're not as careful as you think you are."
He left me standing next to my car with my keys in my hand and the feeling that I was being watched from every angle.
My apartment was dark when I got home. I didn't turn on the lights. Just opened my laptop at the kitchen table and started searching.
Patricia Chen. Born 1968. Stanford MBA. Joined Redwood Ventures in 1995. Made partner in 2001. Lost eight million on MobilePlay Interactive in 2007.
I clicked through news articles, blog posts, investor forums. Found a quote from her in TechCrunch: "The mobile gaming sector is poised for explosive growth. MobilePlay represents the future of entertainment."
Published April 2007. The company filed for bankruptcy in November.
I'd known about MobilePlay because I'd watched it happen. I'd been an intern at Greylock that summer, and I'd written that memo because I'd already seen how the story ended. In the original timeline, Greylock had ignored my recommendation and invested anyway, lost three million when the company folded.
This time, they'd listened.
This time, Patricia had lost alone.
My phone buzzed. Lily's name on the screen.
"Hey," I said.
"I had another dream." Her voice was tight. "About the accident. But this time it was different."
"Different how?"
"There was someone else there. Standing on the sidewalk, watching. I couldn't see their face, but they were just... watching. Like they knew what was going to happen." She paused. "Marcus, I know this sounds crazy, but it felt real. Like a memory, not a dream."
My mouth went dry. "What else do you remember?"
"Nothing. Just the feeling that someone was there, and they didn't try to stop it." She laughed, but it sounded forced. "I'm probably just stressed. The mechanic said my brakes were fine, so I'm being paranoid."
"Paranoid is good. Paranoid keeps you alive."
"That's a weird thing to say."
"I just mean—be careful. Pay attention. If something feels wrong, trust that feeling."
"Okay, now you're scaring me. Is everything all right?"
No. Nothing was all right. I'd come back to save her and instead I'd created a timeline where she was having dreams about her own death, where Patricia Chen was hunting me for insider trading, where David's brother was dying and I was the only one who could help but I was too afraid to let anyone close enough to see what I really was.
"Everything's fine," I said. "Just tired. Long day."
"Get some sleep. And Marcus? Thanks for caring. I know you're not great at the whole emotional support thing, but you try, and that matters."
She hung up before I could respond.
I sat in the dark and thought about the memo I'd written. About Patricia's eight million dollars. About David's brother and Lily's dreams and all the ways my future knowledge was creating ripples I couldn't predict or control.
Whitmore was right. I was too accurate. I'd been so focused on using what I knew to build something successful that I hadn't considered how suspicious that success would look. In the original timeline, I'd failed slowly, made mistakes, learned through trial and error. This time, I was skipping all the steps that made failure look like learning.
I opened a new browser tab and searched for Patricia Chen's investment history. Clicked through portfolio companies, board positions, public statements. Found a pattern: she specialized in mobile technology, had a track record of early bets that paid off, and she was ruthless about due diligence.
If she was digging into my background, she'd find the Greylock memo. She'd find my academic record, my internships, my published papers. She'd find a twenty-three-year-old with no business experience making predictions that shouldn't be possible.
And then she'd start asking questions I couldn't answer.
My laptop chimed. New email. No sender name, just a string of numbers that looked like an IP address.
Subject line: "I know."
I stared at the screen. Didn't click. Didn't move.
The preview text showed one line: "I know what you're doing, and I know how you know."
My hand was shaking when I opened the email. No message body. Just an attachment: IMG_2847.jpg.
I downloaded it.
The photo loaded slowly, top to bottom, like a curtain being pulled back on a stage. Daylight. A street corner. A mechanic shop with a red awning.
And me, standing on the sidewalk across the street, watching the entrance.
Yesterday. I'd been there yesterday, making sure Lily's car was actually being fixed, making sure the mechanic I'd paid extra wasn't cutting corners.
Someone had been watching me watch her.
Someone knew.
My phone buzzed. Text message from an unknown number: "Check your email."
I was already reading it when the second message came through: "We should talk. Tomorrow. 10 AM. The coffee shop on University Ave where you met Sophia. Come alone."
My fingers found the burn scar on my palm, pressed hard enough to hurt. The apartment was too quiet. The darkness felt like it was pressing in from all sides.
I typed back: "Who is this?"
Three dots appeared. Disappeared. Appeared again.
The response came through: "Someone who remembers."
I stared at those two words until they stopped making sense. Someone who remembers. Remembers what? The original timeline? That was impossible. I was the only one who'd come back. The only one who knew.
Unless I wasn't.
My laptop screen flickered. Another email. Same anonymous sender.
Subject: "Proof."
I opened it.
The attachment was a video file. Thirty seconds long. I clicked play.
Security camera footage. Black and white, timestamp in the corner: November 15, 2008. Three weeks from now. A parking garage. Two figures visible.
One was me.
The other was Sophia, and she was holding a gun, and in the video I was on my knees with my hands up, and she was saying something I couldn't hear because there was no audio, and then—
The video cut off.
My phone rang. Unknown number.
I answered.
"Hello, Marcus." The voice was distorted, run through some kind of filter that made it sound mechanical. "I see you've received my messages. I want you to understand something: I'm not your enemy. But I need you to stop what you're doing, because you're changing things that shouldn't be changed, and if you don't stop, people are going to die. People who lived the first time. Do you understand?"
"Who are you?"
"Someone like you. Someone who came back. Someone who's trying to fix what you're breaking." A pause. "The coffee shop. Tomorrow. Ten AM. We'll talk about Lily, and Sophia, and what happens in that parking garage if you don't listen."
The line went dead.
I sat in the dark with my laptop open and my phone in my hand and the knowledge that I wasn't alone anymore, that someone else remembered, that they'd been watching me, and they had proof of a future that hadn't happened yet, and they knew about Lily and Sophia and things I hadn't done but apparently was going to do, and—
My laptop screen went black.
Then it flickered back on.
A single line of text appeared in a new browser window, white letters on black background, typing itself out one character at a time:
"You're not the only one who knows what's coming. But you're the only one stupid enough to think you can change it without consequences. See you tomorrow. Don't be late. And Marcus? Bring the photo of Lily's brake lines. The ones you took before you paid the mechanic to cut them. We need to discuss your definition of 'saving' people."
The window closed.
My apartment door handle turned, slow and deliberate, metal scraping against metal, and I realized I'd forgotten to lock it when I came home, and someone was—