I got to school earlier than usual the next morning.
Not deliberately early — I'd slept badly, woken before it was properly light, lay there for a while without going back under, and eventually just got up. The streets were quiet in that not-quite-dawn way, the convenience store on the corner its own little island of light. I went in, bought water, and kept walking.
I'd accepted the card result. It sat in my mind now as something solid — cold, heavy, but stable. Not something I could keep turning over without losing function. The question of how long I'd been in here was one I still couldn't answer, so I'd put it to one side. The only thing I could move forward on was finding Noah.
When I arrived, the classroom was almost empty. I put my bag down and shifted my gaze toward the back of the room.
Last row, window seat — empty.
I looked away, pulled out a notebook, and started going over yesterday's notes. Something to occupy my hands. Students filtered in gradually, voices starting to fill the room. I kept my eyes on the page, listening without looking, and waited.
Then I heard the sound of a chair being pulled out at the back.
I didn't look up immediately. I finished the line I was on, wrote two more, then lifted my gaze naturally toward the back of the room.
Noah was there. Bag on the desk, taking things out, movements the same as they'd always been — composed, unhurried. He looked like someone returning from ordinary absence. As if no interruption had occurred.
I watched him for a moment. The posture, the angle of his head when he glanced down — all of it matched. But I already knew that if this place could construct everything else with that degree of fidelity, the details of a person were no more difficult. Matching the surface wasn't the same as being real.
I looked back at my notebook.
The morning passed without me hearing much of it. I kept returning to the same problem: how to ask a question that a constructed version couldn't answer accurately. If the dream knew everything I held in my mind, it could produce a version of Noah who seemed to answer correctly. There was no clean way through that.
At lunch, Ashly asked if I'd slept okay. I said a bit, not really. She didn't push, just put an extra item on my tray. I noted it and kept walking.
When the first afternoon class ended and Ashly left for the library, the room thinned out. A few people stayed, working quietly. The light came through at an angle.
I stood up and walked to the back of the room.
Noah was still at his desk, a book open in front of him. He looked up when I reached him, without particular surprise, and waited.
"Where have you been." I sat in the chair beside him. No preamble.
"Things came up."
"What kind of things."
"Family," he said, giving nothing more.
I looked at him. "I ran a test."
He didn't ask what test. He just watched me.
"I shuffled a deck of cards. Drew four, face down. Chose four in my mind first — didn't say them out loud. Then I turned the cards over. All four matched. Exact order."
He said nothing for a moment. Then: "You've been under a lot of pressure."
I paused. "What?"
"You were only just discharged," he said, his voice measured. "Mentally and physically, you're not fully recovered yet. In that kind of state, it's normal to over-attribute significance to coincidences. It's a recognised response."
I looked at him. "Four cards in a row isn't a coincidence."
"The probability is low," he said, "but not zero. And when you say you let your mind produce four things 'randomly' — human beings aren't actually random. We have preferences. Research shows that seven is the number people select most often when asked for a random number. Was your first thought a seven?"
I said nothing.
"And," he continued, "the way you shuffled, the way you drew — those aren't perfectly random either. Your physical state at that moment affects things. I'm not saying you deceived yourself. I'm saying there are multiple explanations for that result, and you don't have to land on the most extreme one."
He said all of this evenly. Every sentence had its own internal logic. I sat there absorbing it, feeling something in my thinking loosen slightly — not because I believed him, but because I couldn't immediately break the argument. Seven was one of the most commonly selected random numbers. Shuffling wasn't truly random. Both of those things were true.
But I knew what I'd thought. And I knew the four cards.
"Then what about the other things," I said. "History class. English period. The car at the junction. Ashly's tomato and egg."
"You've started looking," he said. "Once you're looking for this kind of pattern, your brain begins selecting for it. The instances that match feel significant. The ones that don't, you don't hold onto. That's confirmation bias. It doesn't mean the events are actually occurring at a higher rate — it means you're noticing them differently."
Confirmation bias. The same word I'd used on myself.
I was quiet for a moment. "You came to see me at the hospital."
He stopped for just a fraction of a second. "Yes."
"Why."
"I wanted to make sure you were okay," he said, simply, nothing added.
I held his gaze. I was looking for something — a shift, something that didn't quite stay still, a moment where what he was performing and what was underneath came apart. His expression stayed even. Composed in a way that could be real or could be exactly what a composed expression looked like.
"Where do you think I am right now," I said. Directly. No lead-up.
He looked at me for a few seconds, then said: "You're at school. Sitting next to me."
"Noah."
"Elena," he said, dropping his voice slightly. "You need rest. You've just been discharged. You shouldn't be doing this to yourself."
I held that for a moment, then stood up, pushed the chair in.
"Alright," I said. "Thank you for the explanation."
He looked up as I rose, and something crossed the edge of his expression — not quite formed, not said.
I walked out to the corridor and kept moving. The light in the hallway was ordinary, people still flowing past.
I went through the conversation again, sentence by sentence.
He wasn't wrong. The things he'd said each had a basis. Seven being the most chosen random number — true. Shuffling not being truly random — true. Confirmation bias — real. Taken together, those arguments could dismantle what I'd done. Almost entirely.
But there was one thing he hadn't addressed.
In my mind, the third thing hadn't been a number. It was a specific card — Jack of hearts. Not just a face card, but a specific suit. He'd pointed to seven as the most common choice, but he hadn't said anything about the most common card face, and he hadn't said anything about the suit.
That gap was small. Small enough that I could have told myself to let it go.
I didn't let it go.
I filed it somewhere that wouldn't be easily lost — not because it proved anything, but because I wasn't going to pretend I hadn't noticed it.