Every noodle plant knows the measurement. Weigh the dry noodle, cook it, dry down what stayed in the water, and express it as a percentage. It is one of the two or three numbers the category is judged on, and it is the one a customer sees without measuring anything, because cloudy water on a domestic stove reads as a cheap product.
What is less widely agreed is what the number is telling you. Read literally, cooking loss says starch left the noodle. Read correctly, it says the structure around the starch was not continuous enough to keep it there. That distinction is not academic. It decides whether the next trial adds something to the recipe or changes something in the process, and those two projects have very different costs.
What actually leaves
Cooking heats water inside the noodle above the gelatinisation temperature of the starch. Granules take up water and swell. Chains that were ordered inside the granule become mobile and diffuse outward. A 2024 review of starch in wheat based noodles and pasta describes leached amylopectin attaching to the noodle surface, and starch chains diffusing outward during gelatinisation.
So the leachate is starch polymer, mostly amylose with amylopectin behind it, together with soluble protein and low molecular weight solubles that were never bound to anything. None of that is unusual. Every cooked noodle loses some. The question is only how much, and what decided it.
The barrier, and what weakens it
In a wheat noodle the answer is the gluten network. The same 2024 review describes gluten as a barrier around the starch granules, and states that diluting gluten content by introducing exogenous starch into the noodle leads to a diminished barrier effect surrounding the starch. Where the network is strong the granule swells into resistance. Where it is diluted or discontinuous, the granule swells into space, and what swells into space eventually diffuses into water.
Two consequences follow that are worth holding on to.
The first is that starch and protein are competing for room. The review describes the expansion of starch giving rise to a steric hindrance on the amalgamation of gluten. The granule swelling and the protein network forming are not independent events happening in the same noodle. They are two things trying to occupy the same volume.
The second is that amylose content is a brake on that competition. Higher amylose constrains starch expansion, which reduces the space the starch occupies during cooking and leaves greater expanse for the protein to establish a robust framework. This is why a raw material specification that looks like a nutrition line, amylose percentage, turns out to be a structural specification.
One flour, three drying routes, three different noodles
The experiment that settles the argument
If cooking loss were about what is in the recipe, the same flour would give the same result. It does not.
A 2025 study of gluten free noodles made from pregelatinised rice flour compared noodles containing 15 percent pregelatinised rice flour prepared by three processing routes, then measured the turbidity of the cooking water at 675 nanometres, the standard optical proxy for what has leached out. Hot air drying gave 1.172. Freeze drying gave 0.765. Twin screw extruder drying gave 0.507. A wheat control measured 0.160.
The formulations used the same inclusion level. The highest cooking-water turbidity was more than twice the lowest.
The authors attribute the difference to what each route did to the starch. They describe extrusion forming numerous hydrogen bonds through starch gelatinisation, creating a three dimensional network that retains structural integrity during cooking, and hot air drying quickly removing moisture while partially damaging the crystalline structure of the starch, which compromises binding capacity. Freeze drying preserves the starch structure but does not build viscosity or gel forming properties, so it lands between the two.
That is the whole argument in one experiment. The ingredient did not change. The matrix did, and the matrix is what the water sees.
Why gluten free is harder, stated honestly
The same study is also the clearest available answer to a question customers ask constantly, which is how close a gluten free noodle can get.
The best of the three gluten free routes measured 0.507 against a wheat control of 0.160. Roughly three times the cooking-water turbidity, from the route specifically chosen to build structure. The gap narrowed a great deal and it did not close.
Texture behaved the same way. The wheat control measured 84.6 newtons for hardness and 0.312 newtons for extensibility. The three gluten free noodles were all harder, at 94.01, 117.3 and 100.8 newtons, and all far less extensible, at 0.133 to 0.158 newtons. A noodle can be made firm without gluten. Making it firm and extensible at the same time is the part that is difficult, because extensibility is what a continuous protein network provides and there is no protein network to provide it.
Anyone promising a gluten free noodle indistinguishable from wheat is either working from a much better dataset than this one or is not being careful.
The raw material window
Where there is no gluten, the starch has to be both the filler and the frame, and then its own properties decide almost everything.
A 2022 study of four rice varieties made into noodles found cooking loss ranging from 14.33 to 25.50 percent. Across those varieties, amylose content correlated negatively with cooking loss at r equals minus 0.81, and flour solubility correlated with cooking loss at r equals 0.99. The paper cites a 20 to 25 percent amylose range from earlier work, alongside a different published range, and states that selection depends on the intended noodle product.
Amylose is a useful raw material specification, but the appropriate range needs to be established for the product and process.
Wheat is more forgiving but not immune. A 2025 comparison of noodles from four whole wheat flours reported cooking loss between 5.22 and 8.34 percent, all below the 10 percent that study treats as the acceptability threshold, with farinographic water absorption between 77.07 and 79.40 percent for whole wheat against roughly 58 percent for refined. The differences between the flours were attributed to fat content interfering with gluten formation, with the highest fat flour producing the lowest dough strength.
Which is the plant floor version of the same point. The flour changed, so the water absorption changed, so the network that formed at that water level changed, so the bite changed. Nobody altered the recipe.
The tradeoff nobody escapes
The most useful figures in this whole area come from a 2022 study of six commercial instant noodles, three fried and three not, because it measures the trade the category actually makes.
The fried products had lower cooking loss, 9.60 to 11.08 percent against 12.40 to 14.93 percent for the non fried. They cooked faster, 159 to 238 seconds against 261 to 382 seconds. Their water absorption was lower, 126.34 to 139.16 percent against 175.65 to 205.02 percent. Their structure is described as honeycomb like, porous and less uniform, which is what lets water in quickly.
And their fat content was 12.86 to 17.49 percent, against 1.95 to 2.72 percent for the non fried.
The fried products in this comparison combined lower cooking loss and shorter cooking times with higher fat content. Their different formulations and processing histories mean that frying alone cannot be assigned all of those differences. There is no version of this where a formulator optimises one of them in isolation. The porosity that lets water in in three minutes is the porosity that held the oil.
Four measured properties of commercial instant noodles
What this changes about how a project is run
If cooking loss is a property of the matrix rather than a property of the recipe, three things follow for anyone with the number on a specification sheet.
Measure the matrix, not only the outcome. Water absorption, amylose content and flour solubility explain more of the variation between batches than anything added later can correct.
Expect movement elsewhere. A change that holds solids in the noodle will also change how fast it rehydrates, how much oil it takes in frying and how it bites. Those are not side effects. They are the same structure seen from four directions.
Set the dose deliberately. More is not monotonically better in this category, and past a point additional inclusion increases leaching rather than reducing it.
This is why a noodle problem is rarely answered by a single ingredient. It is answered by a system chosen against the process that is actually running, which is a different exercise from choosing a material against a specification.
If you are working on cooking loss, bite or oil uptake in a noodle, tell us what you are making and what is going wrong.
