Educational

A boiled potato eaten hot tests at a glycemic index of about 80. Chill that same potato overnight and it comes back around 56.

Nothing was added. Nothing was removed. The starch simply rearranged itself while it sat in the fridge.

What happens to starch when it cools

Raw starch is packed into tight crystalline granules that digestive enzymes struggle to reach. Cooking with water and heat blows those granules open, which is what makes cooked food digestible in the first place. It's also what makes cooked starch fast.

Cooling reverses part of that. As the temperature drops, some of the loosened starch chains re-associate into a new crystalline structure that amylase can't easily break down. The technical name is retrogradation. The practical result is that a portion of what was fast carbohydrate now behaves like fibre: it passes through the small intestine largely undigested and gets fermented by bacteria in the colon instead.

That fraction is resistant starch. It contributes little to blood glucose, and the bacteria that ferment it produce butyrate, a short-chain fatty acid the cells lining your colon use as fuel.

How much the numbers actually move

The size of the effect depends on the food and how long it sits.

• Potatoes: boiled and eaten hot, GI around 78–82. Boiled, refrigerated overnight, eaten cold: roughly 56. That's the difference between a high and a moderate rating.
• White rice: cooked and chilled 12–24 hours shows glycemic responses roughly 10–20% lower than the same rice fresh.
• Pasta: already moderate at 45–55 hot, and cooling gains a few more points. Pasta salad is a genuinely low-GL format.
• Green bananas and cooked-cooled oats carry resistant starch by default, which is part of why overnight oats behave differently from a hot bowl of the same oats.

These are averages from testing groups, not guarantees for one person on one evening. The direction is consistent; the magnitude varies.

Reheating doesn't undo it

This is the part that surprises people, and it's the part that makes the whole thing usable.

Once retrograded starch has formed, gently reheating the food doesn't fully convert it back. Some studies find reheated cooled potato lands even lower than cold potato, though the effect there is small enough that I wouldn't build a plan on it. What's solid is that yesterday's rice, reheated for tonight, keeps most of the advantage it gained in the fridge.

So the practical version is not "eat everything cold." It's "cook it a day early."

Where this fits into real cooking

Potato salad is the obvious one, and it's a rare case where the traditional preparation is also the lower-GL preparation. Boil, cool completely, dress with oil and vinegar or a yogurt-based dressing. The acid and fat slow things further.

Rice cooked Sunday and used Monday and Tuesday does the same work with no extra effort. Fried rice made from day-old rice is the version most cuisines already insist on, for texture reasons that happen to align.

Pasta salad, cold soba, chilled grain bowls, and lentil salads all sit in this space. So does bread that's been frozen and toasted, which picks up a smaller version of the same effect.

One caution on food safety: cooked rice needs to go into the fridge within an hour or so of cooking, not sit on the counter cooling for three hours. Bacillus cereus is the reason for that rule, and it's unrelated to glycemic load but far more likely to ruin your week.

Other places resistant starch shows up

Cooking and cooling is one of four routes to the same compound, and the others are worth knowing because they don't require planning ahead.

Some foods carry it natively. Green bananas are largely resistant starch, and it converts to regular sugars as they ripen — which is the same mechanism behind the ripeness effect on a banana's glycemic load. Raw oats hold more than cooked ones, so overnight oats and muesli start ahead of porridge. Legumes contain resistant starch whether or not they've been chilled, which is part of why their glycemic index runs so low.

Then there's cooking method. Par-boiling rice, or the traditional practice of cooking rice with a spoon of coconut oil and then refrigerating it, has been shown in small studies to increase the resistant fraction beyond cooling alone. That research is preliminary and the sample sizes are tiny, so I'd file it as interesting rather than settled.

Where the evidence is thinner

Two things get overstated about resistant starch, and both are worth naming.

The first is the size of the fermentation benefit. Butyrate production is real and well documented; the leap from there to specific health outcomes in humans is longer than the supplement marketing suggests.

The second is consistency between people. How much benefit you get depends on which bacteria you're carrying, and gut microbiome composition varies enormously. Two people eating the same cooled rice can ferment it quite differently. If you have a CGM, this is a cheap thing to test on yourself over a few days rather than take on faith.

What it doesn't do

Cooling doesn't turn a high-carbohydrate serving into a low-carbohydrate one. A large portion of cooled rice still delivers a large portion of carbohydrate, and the glycemic load calculation still starts from grams on the plate. Retrogradation shifts the quality of those grams; it doesn't shrink the quantity.

It also doesn't stack indefinitely. A second and third cooling cycle add very little beyond the first.

What it does offer is one of the few adjustments that costs nothing, requires no substitution, and works inside food you already cook. Most of the levers in blood sugar management involve giving something up. This one just involves cooking it the night before.