| Elapsed | Core | vs model | |
|---|---|---|---|
| Nothing logged yet. | |||
Target is the core temperature as it leaves the oven. If you want it higher on the plate, the rest and the sear add to it — set this to the number you want on the probe when you open the door. The rest alone is worth two or three degrees for a thin steak from a gentle oven, and five or more for a thick one or a fan.
Probe the raw steak, type that number, and press Start cook as the steak goes into the oven — that press is the clock's zero.
The core of a thick steak is not a lumped body — at 40 mm the Biot number is near 1 — so it follows the
one-term slab solution: after a dead time it approaches an asymptote exponentially. Evaporation sets that
asymptote well below the oven: a genuinely wet surface sits near the oven air's wet-bulb
temperature, about 38–45 °C for a dry 125 °C oven, which is why a damp steak stalls maddeningly
close to 44 °C. So three parameters get fitted to your readings:
T = T∞ − (T∞ − T₀)·exp(−(t − lag)/τ).
Every reading updates a posterior over those three, not a single best guess. The next check is scheduled for the earliest moment the steak could plausibly already be 2 °C short of target — the 5th percentile — so the risk of sailing past unnoticed stays at 5% whether the fit is sharp or vague. A vague fit simply earns an earlier look. Once that moment falls past the predicted finish there is nothing left to learn, and it tells you to stop opening the oven and coast.
Keep it honest: probe the thermal centre, in the same spot each time, and get the steak back in within about half a minute. Pat the surface dry before it goes in — that alone removes most of the dead time.