ScaleRecipe

Technique · July 4, 2025 · 8 min read

Why doubling a cake doesn't double the bake

The cube-root rule, and the geometry of how heat actually moves

A reader once wrote in to ask why their doubled chocolate cake had a raw centre when the recipe's timing said 35 minutes and they'd given it 70. They'd done the scaling math right (2× ingredients, 2× time) and the cake had still failed.

The answer is geometric. Cooking time doesn't scale with quantity; it scales with how far the heat needs to travel, which depends on the shape of the food. And for a cake baked in a pan, that distance grows much more slowly than the volume.

The cube root in plain language

Imagine a cake batter as a cube — easier to draw than a cake, and the math is the same either way. A cube of side 1 has volume 1. Double the volume, and you need a cube of side ∛2 ≈ 1.26. Heat reaches the centre by penetrating from the surface inward; the new cube is only 26% wider, so the heat only has 26% further to travel.

The rule for thick three-dimensional bakes — cakes, roasts, casseroles — is therefore:

new_time ≈ old_time × ∛(new_quantity / old_quantity)

For a 2× recipe: new time is roughly 1.26 × old — call it 25% longer. For a 4× recipe: roughly 1.59 × old, or 60% longer. Not double, ever.

Why liquid scaling is different

For soups, stocks, simmered sauces, the cube-root rule doesn't apply, because heat doesn't penetrate by conduction the way it does in a cake. Convection currents in a liquid mix the heat throughout the volume; the bulk heats up almost as quickly as the volume increases. So for liquids:

new_time ≈ old_time × (new_quantity / old_quantity)

Roughly linear. A doubled pot of soup takes about twice as long to come back to a simmer and reduce.

And sheet bakes

Cookies, sheet brownies, focaccia spread thin in a half-sheet — these scale with area, not volume. The depth doesn't change much when you double the recipe; you spread it across more surface, on more pans. So:

new_time ≈ old_time × √(new_quantity / old_quantity)

In practice, sheet bakes barely change. A doubled batch of cookies is a doubled number of sheet pans, not a doubled bake time.

Three rules, one converter

ScaleRecipe's cooking-time converter gives you all three estimates side by side. Pick the rule that matches your dish's geometry and use it as a starting point — never as the definitive answer. Cooking time depends on more than scaling: oven recovery, pan conductivity, starting temperature, and how full the oven is all matter.

Why the scaler doesn't do this for you

We deliberately left cooking time out of the recipe scaler. The scaler scales ingredients exactly because the math is unambiguous; cooking time is a geometric estimate that depends on which rule applies, and an over-confident time prediction causes more bad bakes than scaling ever does. The cooking-time converter gives you the rule and asks you to match it to your dish.

How to actually verify doneness

Time is unreliable. Doneness checks aren't. Buy a digital instant-read thermometer — it pays for itself the first time it tells you a 6-pound roast really does need another 25 minutes. For baked goods, a wooden skewer in the centre that comes out clean (or with a few damp crumbs for moist cakes) is the gold standard. For bread, an internal temperature of 88–98 °C (190–210 °F) tells you when the loaf is done regardless of how long it's been in the oven.