You want a medium-rare roast, so you pull it out of the oven at 135 degrees, the number every chart says means medium-rare. Ten minutes later, after it rests, you slice in and it's closer to medium. Nothing went wrong with the thermometer. The roast kept cooking after it left the oven, and if you didn't plan for that, your target temperature was never actually your target.
Carryover Cooking Is Real Physics, Not a Rounding Error
Meat doesn't stop cooking the instant it leaves the heat source. The outer layers are hotter than the center, and heat continues migrating inward toward the cooler core for several minutes after the roast is removed, driven by the same temperature gradient that was cooking it in the oven. This is carryover cooking, and it's a well-documented, predictable effect, not kitchen folklore.
The size of the temperature rise depends mainly on two things: the size of the roast and the cooking temperature it came out of. A large roast cooked at a high oven temperature carries a bigger internal-to-external heat gradient, and therefore carries over more, sometimes 10 degrees or more. A small cut, or something cooked at a lower, gentler temperature, might carry over only 2 to 3 degrees.
Why This Trips Up Even Careful Cooks
The frustrating part is that carryover cooking is invisible at the moment you make the pull decision. The thermometer reads exactly what it reads. There's no visual cue that the roast has 8 more degrees of cooking still coming. The only way to account for it is to know it's happening in advance and pull earlier than the final target number.
This is different from most cooking mistakes, which show up as an obvious visual or textural cue you can course-correct on the spot. Carryover error only shows up after you've already cut into the meat, which is exactly when it's too late to fix.
Working Backward From Your Actual Target
The fix is straightforward once you know it's a factor: subtract the expected carryover rise from your desired final temperature, and that's your actual pull temperature. If you want a final 135 degrees for medium-rare and you're roasting a large cut at a high oven temperature with an expected 8 to 10 degree carryover, your pull temperature is closer to 125 to 127 degrees, not 135.
This matters more than it sounds like it should, because the doneness bands for meat are narrow. The difference between medium-rare and medium is roughly 10 degrees. Miss the carryover adjustment on a large roast and you can slide an entire doneness category without the thermometer ever being wrong at the moment you read it.
Resting Time Isn't Optional Either
Carryover cooking and resting time are related but not identical. Resting lets the meat's internal juices, which get pushed toward the center under heat, redistribute evenly through the muscle fibers before you cut into it. Cut a roast immediately after pulling it and a disproportionate amount of juice runs out onto the cutting board instead of staying in the meat. The USDA's guidance on safe minimum cooking temperatures covers the food-safety side of rest time, since carryover cooking is also part of how a roast reaches a safe internal temperature after leaving the heat, not just a texture consideration.
Rest time and carryover cooking happen simultaneously. The carryover temperature rise mostly occurs during the first several minutes of the rest period, which is also when juice redistribution is happening. Skipping the rest to save time undercuts both effects at once.
Cut Size and Shape Change the Number
Carryover isn't a fixed number that applies the same way to a whole beef tenderloin and a bone-in prime rib. A thick, dense roast holds heat longer and carries over more than a thinner cut of the same weight, because the internal-to-external temperature gradient is steeper across more mass. A spatchcocked or butterflied cut, with more surface area relative to its thickness, carries over less because heat has less distance to travel and less mass retaining it.
Bone-in cuts add another variable. Bone conducts heat differently than muscle tissue, which is part of why bone-in roasts sometimes cook and carry over slightly unevenly near the bone compared to the rest of the cut.
A Rough Framework by Cut Size
As a starting point, not a universal rule: small cuts (a chicken breast, a thin steak) carry over 1 to 3 degrees. Medium cuts (a pork loin, a whole chicken) carry over 3 to 6 degrees. Large roasts (prime rib, whole turkey, large beef roasts) carry over 5 to 10 degrees, sometimes more depending on oven temperature.
These ranges are wide because oven temperature, roast shape, and starting temperature all shift the actual number. A precise pull temperature for a specific cut, weight, and target doneness is a genuinely useful thing to calculate rather than estimate, since the cost of guessing wrong is an entire meal.
Reverse Sear and Sous Vide Change the Rules Slightly
Not every cooking method produces the same carryover behavior. A reverse-seared roast, cooked low and slow first and then finished with a hot sear at the end, carries over less after the sear than a traditional high-heat roast, because the low-temperature cook phase already brought the interior close to target before the sear ever starts, and the sear itself is brief enough that it doesn't build up the same deep heat gradient. A sous vide-cooked piece of meat carries over the least of all, often close to zero, because the entire piece equilibrates to the water bath's exact target temperature before it ever gets a quick finishing sear, so there's no meaningful internal gradient left to keep cooking.
This matters because carryover guidance written for a standard oven roast doesn't transfer directly to these other methods. If you're used to pulling a conventional roast 8 to 10 degrees early and you apply the same habit to a sous vide piece, you'll actually undercook it, since there's little to no carryover left to account for.
A Common Mistake: Trusting the Pop-Up Timer Instead
A lot of store-bought turkeys and some roasts come with a built-in pop-up temperature indicator, and these are worth treating with more skepticism than a separate instant-read thermometer. Pop-up timers are calibrated to trigger at a single fixed temperature, usually one that guarantees a well-done result for food safety margin across a wide range of consumer ovens and cooking styles, and by the time it pops, carryover cooking on a large bird can push the final rested temperature well past that already-conservative number. Anyone who has had a noticeably dry Thanksgiving turkey has likely run into exactly this effect: the pop-up timer did its job, but its job wasn't optimized for juiciness, and carryover cooking pushed the result even further past done than the trigger point already was.
Why a Single Universal Carryover Number Doesn't Work
It's tempting to want one simple rule, "just pull everything 10 degrees early," and skip the rest of this. The problem is that a fixed 10-degree rule overcorrects for small cuts, which might only carry over 2 to 3 degrees, leaving them underdone at pull time relative to what you actually wanted, and it can undercorrect for very large roasts cooked at high heat, which can carry over more than 10 degrees. A single flat number trades one kind of error for another depending on what you're cooking, which is exactly why cut size, roast weight, and starting oven temperature all need to factor into the actual pull temperature rather than a rule of thumb that ignores them.
Where a Roasting Calculator Actually Earns Its Keep
The free Meat Roasting Calculator by EvvyTools calculates roasting time, oven temperature, and pull temperature with carryover already factored in, for 20-plus cuts of beef, pork, poultry, and lamb, along with rest time. Instead of memorizing carryover ranges by cut type, you get the actual pull number for your specific roast.
For more on kitchen math that doesn't behave the way straightforward multiplication assumes, Why Doubling a Recipe Breaks More Than Just the Math at EvvyTools covers similar nonlinear effects in scaling and baking, including why leavening agents and pan geometry don't scale the way bulk ingredients do.
ThermoWorks, a company built around instant-read thermometers, publishes detailed carryover cooking data by cut and weight if you want to go deeper on the underlying numbers, and Wikipedia's entry on food safety covers the temperature side of why hitting the right internal number, accounting for carryover, actually matters beyond just texture preference.
The thermometer isn't lying to you when you pull the roast. It's just telling you where the meat is right now, not where it's about to be in ten minutes. Planning for that gap is the difference between hitting your actual target and being ten degrees off every single time.
Top comments (0)