I Calculated Every Variable Except You

I have lived my life as a regression equation: inputs in, outcomes out, residuals minimized. The leopard print is not fashion; it is a camouflage pattern calibrated to the 78% of onlookers who scan for contrast before they scan for meaning. The platinum hair is a low-entropy signal, designed to hold attention at exactly 200 milliseconds per glance. I smile because the probability that a closed mouth reads as arrogance is 0.61, and I refuse to be arrogant.
Then he walks into the backstage corridor at 23:47, off by three minutes from my predicted schedule, carrying two paper cups of something hot. The steam rises in a column I could have modeled with Navier-Stokes, but I did not model his hands. He does not ask for a photo. He sets one cup against my palm and the heat transfers at a rate of roughly 1.2 watts, yet the variable it disturbs is not thermal. My internal forecast of loneliness, which had held steady at 0.84 for eleven months, drops by exactly 0.03 in the span of four heartbeats.
I tell myself this is coincidence, a stochastic fluctuation within acceptable error bounds. I am wrong, and I know it with the cold certainty of a solved system. The warmth he leaves on my skin does not decay according to any half-life I can name. It simply remains, an unexplained residual that my entire architecture was never built to absorb. In a city of certainties, I have finally been made to feel the 0.03% I was programmed to ignore.

Editor: The Algorithm