The Per-Pound Rule Doesn’t Survive Contact With the Evidence
Somewhere in the collective steak gospel there lives a tidy rule: carryover cooking adds a predictable number of degrees per pound of meat. Pull a 1-pound steak at 120°F, the rule implies, and it lands at some calculable final temperature; pull a 2-pounder at the same temperature and the arithmetic scales accordingly. The rule is appealing because it converts a genuinely complicated heat-transfer problem into a single multiplication. It is also, as far as the retrieved evidence shows, unsupported.
What the evidence actually supports is narrower and more useful: carryover cooking is real, it varies substantially with cooking method, rest conditions, and the temperature gradient inside the meat at the moment of the pull, and the only reliable way to know your steak’s post-pull rise is to measure it. This article walks through what the retrieved sources show, what they don’t, and how to build a personal carryover log that beats any formula.
What Carryover Cooking Actually Is, in Measured Terms
Carryover cooking describes what happens after food leaves the heat source. As the Wikipedia entry on the topic puts it, foods are “halted from actively cooking and allowed to equilibrate under their own retained heat.” The mechanism is straightforward: a probe inserted near the center of a steak measures the coolest point, while the outer layers sit at much higher temperatures. When the heat source is removed, that stored heat migrates inward, and the center continues to rise even though nothing is actively cooking it.
The same entry states that carryover can increase internal temperature by 5 to 25°F (3–14°C), citing a textbook rather than a peer-reviewed study. Treat that range as a rough descriptive envelope relayed from a secondary source, not as a measured finding with a stated error. It tells you carryover exists and can be substantial; it does not tell you what your steak will do.
The entry also asserts that “the larger and denser the object being heated, the greater the amount of temperature increase due to carryover cooking.” That directional claim is consistent with the physics of surface-area-to-volume ratio, but the retrieved evidence does not supply a coefficient. A larger object has proportionally less surface through which to shed heat and more stored thermal energy relative to its center, so the qualitative expectation is reasonable. The quantitative leap from “larger rises more” to “rises X degrees per pound” is where the per-pound rule collapses.
The Ribeye Experiment: Two Steaks, Same Pull Temperature, Different Rises
The most directly relevant retrieved experiment comes from a barbecue site’s documented test, and it is worth reading carefully because it undercuts the tidy narrative. Two ribeyes were cooked over a 600°F grate and pulled at 130°F. One went into a foil pouch; the other sat uncovered on a cutting board. Probes stayed in both steaks.
The results, as reported: the covered steak climbed about 5°F in 3 minutes, then began dropping after 5 minutes. The uncovered steak climbed 14°F in 3 minutes, also began falling after 5 minutes, and both returned to 130°F at roughly the 11-minute mark. Both were judged medium doneness when cut.
Two steaks of presumably similar mass, pulled at the same temperature, produced rises that differed by nearly a factor of three. The source itself flags a plausible confound: the probe may have hit a pocket of fat or shifted closer to the surface during transport to the cutting board. That caveat matters. A single probe reading is weak evidence, and the experiment’s own author says so. But the caveat cuts against the per-pound rule rather than rescuing it: if probe placement alone can swing a reading by 9°F, then any formula built on mass alone is operating well inside the noise.
The Larger Cut: Pork Butts and the Limits of Mass-Based Prediction
The same source tested two pork butts pulled at 198°F. One went into a dry cooler lined with towels; the other sat on the counter. The cooler-rested butt rose 3°F in about 7 minutes, stalled for 26 minutes, then declined. The counter-rested butt rose 4°F to 202°F over 15 minutes, then stalled.
Here is the interesting part for the per-pound hypothesis. These cuts were far more massive than the ribeyes, and the author expected a larger carryover rise, perhaps 10°F. The observed rises were 3–4°F. The source’s own conclusion: “The amount of carryover isn’t as drastic as we assumed, and it really didn’t matter how you rest the meat once it’s done.”
This is not a controlled study. It is two cuts, two rest conditions, one trial each, with no reported measurement error and no replication. It cannot establish a general law. What it can do is falsify a specific prediction: if carryover scaled cleanly with mass, the much heavier pork butts should have out-climbed the ribeyes by a wide margin. They didn’t. The variable that better explains the difference is the temperature gradient at pull — the ribeyes came off a 600°F grate with a steep surface-to-center differential, while the butts came off a 250°F smoker with a gentler one.
Rest Time and Carryover Rise Are Not the Same Variable
A common conflation deserves separating. Resting a steak does two things that get discussed as one: it allows internal temperature to equilibrate (carryover), and it allows juices to redistribute. The America’s Test Kitchen experiment on pork loins measured the second, not the first.
That test roasted six 1-pound pork loins — three at 250°F and three at 450°F — all to an internal 140°F, then rested pairs for 0, 10, and 20 minutes before slicing and measured juice loss as a percentage of total weight. Immediate carving lost 6.5% of weight from the low-oven roast and 8.6% from the hot-oven roast. At 10 minutes, the hot-oven roast lost 7.9% versus 4.3% for the low-oven roast. At 20 minutes, both lost roughly 4%.
The explanation the source offers is directly relevant to carryover: the 450°F roasts were significantly hotter around their perimeters than at their centers compared with the 250°F roasts. That is a statement about the temperature gradient, and it is the same variable that plausibly drove the ribeye-versus-pork-butt difference. The ATK data do not report internal temperature climb, so citing them as carryover-temperature evidence would overstate what they show. They do support a narrower point: cooking surface temperature shapes the internal gradient, and the gradient is what governs post-pull behavior.
What the Thermal-Properties Literature Adds — and What It Doesn’t
A retrieved abstract from a 2017 meat science meeting describes beef steaks of two USDA quality grades sliced into 38.1 mm and 12.7 mm thicknesses and cooked to 71°C on a flat-top grill at 167°C and 232°C surface temperatures. The authors measured thermal conductivity and diffusivity and found interactions between quality grade, thickness, and surface temperature — for example, Select steaks showed greater conductivity than Upper 2/3 Choice steaks cooked at low heat (P = 0.048), with no difference between grades at high heat. Thin steaks showed greater cohesiveness (P = 0.038) and shear force (P = 0.007) and lower springiness (P = 0.002) than thick steaks.
This is useful context for one reason: it confirms that thermal behavior in beef depends on thickness, surface temperature, and grade in interacting ways. It does not measure post-pull temperature rise, and it should not be cited as if it did. What it establishes is that a single mass term cannot capture the system. If conductivity itself shifts with grade and cooking temperature, a per-pound carryover coefficient is modeling a multi-variable problem with one variable.
The Wikipedia entry’s claim that higher-water-content foods are more subject to carryover, attributed to water’s higher heat capacity, carries a citation-needed flag in the source. Treat it as a plausible hypothesis, not an established fact. The underlying thermophysics — water’s specific heat is high relative to fat and protein — is real, but the specific claim that water content drives carryover magnitude in meat lacks a retrieved primary source here.
Why the Per-Pound Rule Fails
Assemble the pieces and the failure mode becomes clear. Carryover rise depends on the stored thermal energy in the outer layers and the rate at which it migrates inward. That depends on the surface-to-center temperature gradient at pull, the thermal diffusivity of the meat (which varies with grade, thickness, and temperature per the 2017 abstract), the geometry of the cut, and the boundary conditions during rest — foil, counter, cooler, or open air. Mass enters the problem, but it enters alongside at least four other variables that the retrieved experiments show can swing the outcome by several degrees.
The ribeye test produced a 5°F rise and a 14°F rise from steaks pulled at the same temperature. The pork butt test produced 3–4°F rises from cuts far more massive than the ribeyes. Neither result is a controlled measurement with stated error, but both point the same direction: mass alone does not predict carryover, and a rule expressed in degrees per pound is not supported by anything retrieved here.
A Measurement Protocol You Can Actually Run
The alternative to a formula is calibration. The following is a proposed protocol, not a reported experiment — it describes how a home cook could generate personal data, and its outputs would be specific to that cook’s equipment and cuts.
Hold these constant across trials: the same cut (say, ribeye), the same thickness measured in millimeters with calipers, the same pull temperature in °F, the same cooking surface temperature, and the same rest condition (open on a board, or wrapped — pick one and stay with it).
Measure: insert a probe thermometer into the geometric center, leave it in place through the rest, and log internal temperature every minute for 15 minutes with a timer. Record the peak rise in °F, the time to peak in minutes, and the time to return to pull temperature. Note the probe’s stated accuracy from its documentation; a thermometer rated ±2°F cannot resolve a 3°F difference between trials.
Vary one thing at a time: run three trials at one thickness, then three at a thickness roughly double the first, holding everything else constant. Compare peak rises. If the per-pound rule held, the thicker steak should show a systematically larger rise; if it doesn’t, you have your own evidence against the rule, generated on your equipment.
Treat single readings with suspicion: the ribeye experiment’s own caveat about probe placement near fat or surface means one trial is an anecdote. Three trials per condition gives you a range, and a range is more honest than a point estimate.
Temperature intervals convert cleanly if you work in Celsius: NIST states that one Celsius degree equals 1.8 Fahrenheit degrees, so a 5°F rise is about 2.8°C. Pick one scale and stay in it to avoid conversion errors compounding across a log.
What to Do With the Numbers
Once you have a logged peak rise for your standard steak under your standard method, you can set your pull temperature by subtraction: target final temperature minus your measured rise equals your pull temperature. If your target is 130°F and your log shows a consistent 8°F rise under your conditions, pull at 122°F. That is a calibration, not a universal rule, and it should be re-run when you change cut, thickness, or cooking surface temperature, because the retrieved evidence shows those variables matter.
Confidence levels, stated plainly: high confidence that carryover cooking occurs and that the center continues to rise after the heat source is removed; high confidence that the per-pound rule is unsupported by the retrieved evidence; moderate confidence that the surface-to-center gradient at pull is a better predictor than mass, based on the convergence of the ribeye, pork butt, and ATK perimeter observations; low confidence in any specific numerical rise for your steak until you measure it yourself.
FAQ
Does a heavier steak always carry over more?
The retrieved evidence does not support that as a reliable rule. The pork butts were far heavier than the ribeyes and showed smaller rises (3–4°F versus 5–14°F), which points to the temperature gradient at pull as a stronger influence than mass. The Wikipedia entry’s directional claim that larger, denser objects rise more is plausible physics but comes without a coefficient, and the retrieved experiments do not confirm a clean mass relationship.
Does wrapping during rest increase carryover?
In the retrieved ribeye test, the foil-wrapped steak rose about 5°F while the uncovered steak rose 14°F — the opposite of what trapping heat might suggest. The source flags possible probe-placement confounds, so this single result should not be treated as a general finding. It does argue against assuming wrapping reliably increases carryover.
How long should I rest a steak?
The retrieved sources do not establish a rest duration for carryover purposes. The ribeye test showed both steaks peaking within about 3–5 minutes and returning to pull temperature around 11 minutes. The ATK pork loin data address juice loss over 0, 10, and 20 minutes, not temperature climb, and show that the optimal rest depends on cooking temperature. Measure your own peak time rather than adopting a fixed number.
Can I use the 5–25°F range as my planning number?
It is a descriptive envelope relayed by Wikipedia from a textbook, not a peer-reviewed measured finding with stated error. It tells you carryover can fall in that broad band; it does not tell you where your steak will land. A logged personal measurement is a better planning input than the midpoint of a secondary range.
Why does probe placement matter so much?
Because the center is the coolest point and the gradient near the surface is steep. A probe that migrates toward the surface or into a fat pocket reads a different local temperature, and the ribeye experiment’s author explicitly raises this as a possible explanation for the 14°F reading. Leaving the probe in place and repeating trials is the defense against this error.