Cooking Times by Weight, Not Minutes Per Side: How Roasting Actually Works

Take one six-pound chicken. Roast it at 325F and it needs somewhere between 2 hours 10 minutes and 2 hours 35 minutes. Roast the exact same bird at 425F and it is done in 1 hour 20 minutes to 1 hour 35 minutes. Same chicken, same oven, same cook. The time nearly halves. That is the whole problem with recipe timings in one example: a number like ninety minutes only means something once you also know the weight and the temperature it was measured at.

I built a small reference site called CookClock to fix this for myself, and the more I dug into the underlying data the more I realised how badly the common shorthands mislead people. Timing ranges on the site are derived from the USDA Food Safety and Inspection Service roasting timetables and cross-checked against published guidance, and the safe internal temperatures come straight from the FSIS safe minimum internal temperature chart. It covers thirty cuts across chicken, turkey, beef, pork, lamb and seafood. None of the data is mine. The useful part was putting it in a form where you can enter your actual weight and your actual oven temperature.

Here is why minutes-per-side fails as a rule. It is a rule about surface, and roasting is a problem about volume. Heat enters a roast from the outside and has to travel inward. Double the weight of a joint and you have not doubled the distance to the middle, but you have not left it unchanged either. A thin cut is basically all surface, so a per-side timing works fine for a steak or a fillet. A whole bird, a leg of lamb or a pork shoulder is mostly interior, and the interior does not care how many times you turned it over.

Minutes per pound is the honest compromise. It is not physics, it is a planning estimate that happens to be close enough over the range of weights people actually cook. On the whole chicken page, that same bird works out at 22 to 26 minutes per pound at 325F, 20 to 23 at 350F, 17 to 20 at 375F, 15 to 18 at 400F and 13 to 16 at 425F. Notice those are ranges, not points. Ovens run hot or cold, thermostats cycle, and a fan oven behaves differently from a conventional one. Anyone quoting a single exact minute figure for a roast is quoting a number with a false precision.

There are three things that quietly break the estimate, and all three are worth knowing before you plan a dinner around a clock. The first is starting temperature. Every timing chart assumes properly thawed, fridge-cold meat. A bird that is still partially frozen in the middle can take up to fifty percent longer, which is the single most common reason a Christmas roast runs late. The second is stuffing: if there is stuffing inside the cavity, add another fifteen to thirty minutes so the centre of the stuffing itself reaches temperature, because the stuffing is the coldest part of the whole assembly and it is sitting in the slowest-heating place in the oven. The third is trussing. A tightly tied bird is a denser, rounder object with less exposed surface, so it cooks slower than a loose or untied one.

So the clock estimates and the thermometer decides. That is not a slogan, it is the actual structure of the food safety guidance. The FSIS numbers are simple enough to memorise. Poultry, whether whole, in parts or ground, goes to 165F with no rest required. Ground meats other than poultry, so beef, pork, lamb and veal mince, go to 160F with no rest. Whole cuts of fresh beef, veal, lamb and pork, meaning steaks, chops and roasts, go to 145F with a three-minute rest. Fully cooked ham being reheated goes to 140F.

That three-minute rest is the part almost everybody gets wrong, and it is the most interesting thing in the whole chart. It is not a serving suggestion. It is part of the safety standard. The reason is that after the meat comes out of the oven, the temperature holds or keeps climbing, and that holding time counts toward safety. Time at temperature does the same work that a higher peak temperature would. A cut held at 145F for three minutes is as safe as one taken hotter and served immediately, but it is a much better piece of meat.

Carryover cooking is the mechanism behind that. When you pull a roast out, the outside is far hotter than the centre. Heat moves from hot to cold, so the outside keeps feeding heat inward for several minutes after the oven is out of the picture. The joint is still cooking on the bench. How much it climbs depends on how big and dense it is and how hot the surface got, which is exactly why a thin fillet barely moves and a large rib roast can drift up noticeably. This is why pulling a big roast at the exact target temperature reliably overshoots it.

The practical version: for large, dense cuts, take the meat out a little before your target and let carryover finish the job. For thin cuts and poultry pieces, there is much less carryover to work with, so cook to the number. And because you cannot know how much carryover a particular piece of meat has without measuring it, the thermometer stays in charge either way.

Resting does a second job that has nothing to do with temperature. During cooking, muscle fibres contract and squeeze moisture toward the centre and toward the surface. Carve immediately and a lot of that moisture leaves on the board rather than staying in the slice. Ten minutes or so for a whole chicken, proportionally longer for a large roast, and the difference on the plate is obvious. The meat has not gone cold. It has redistributed.

None of this needs a gadget beyond a probe thermometer, and it is the single cheapest upgrade to home cooking there is. Put it in the thickest part of the thigh for a bird, not touching bone, because bone conducts heat and gives you a reading that is not about the meat. Check in more than one place. If two spots disagree by a lot, believe the lower one.

If you want the numbers rather than the theory, I put the whole thing online as cooking times for 30 cuts, calculated per pound at your actual oven temperature, with the FSIS safe temperature chart alongside it so you can check the clock against the thermometer in the same place. It is free, there is no signup, and it exists mostly because I got tired of doing the arithmetic on my phone with greasy hands.

Building small reference tools like this is a habit at this point. Most of the ones I have made are collected on my tools page and they follow the same logic as this article: take a public dataset that is genuinely authoritative, stop making people do arithmetic in their heads, and be honest about the ranges. Cooking times are a planning estimate. The thermometer is the answer.