Why 75% Hydration Is Wetter in Low-Protein Flour

One recipe line says "75% hydration". It names water as a share of flour weight, and it names nothing else. I tested whether that line means the same dough in a 10.5% protein flour and in a 13% protein flour. It does not. My estimate is that the low-protein dough holds about 37% more water than the flour can absorb, with a plausible range of 22% to 74%. I changed my thesis along the way, and I say how below.

All numbers here assume a finished dough at 24 °C and an oven at 245 °C. Those are my test conditions, not figures from the sources. Every calculation below I did by hand, without the Lab. You can repeat each one with a calculator.

The question

Does the quoted hydration number predict how wet a dough behaves, or does the flour change the answer? The strong form of my thesis said hydration predicts crumb openness poorly unless protein and absorption are held fixed. I could not test that form. I found no open set of recipes that pairs gram weights, a named flour protein and a measured crumb. So I tested a weaker, checkable claim: how far the same hydration moves relative to what the flour can absorb.

Data and where it came from

I read five kinds of source. None gave me what I wanted: a table of crumb scores against hydration and protein.

• Recipe data in grams. The Tartine basic country loaf uses 1,000 g flour (900 g white, 100 g whole wheat), 750 g water, 20 g salt and 200 g leaven. The recipe states 75% hydration [1].
• Hydration by flour type. The authors of a popular no-knead book list recommended hydrations for their method. Most all-purpose flours get 75%. King Arthur all-purpose (11.7% protein) gets 81%. Gold Medal Better for Bread (12.5%) and King Arthur bread flour (12.7%) get 83%. High-gluten flours (14% to 15%) get 85% [2]. These are the authors' recommendations, not measurements.
• A flour maker's own account of the same effect. King Arthur says the proteins in bread flour absorb more water. It says that if you use all-purpose flour with the same water in a high-hydration recipe, "the dough will be too wet, sticky, and loose" [3].
• Absorption mechanics. A flour-testing vendor gives water uptake per gram of each component. Protein absorbs about 2 times its weight. Native starch absorbs about 0.3 times. Damaged starch absorbs about 3 times. Pentosans absorb up to 15 times [4].
• Measured absorption for two flours. A baking reference lists a straight-grade flour at 13.0% protein with 63.5% farinograph absorption. It lists a patent flour at 12.5% protein with 64.3% [5]. A second reference says that higher protein, more damaged starch and more pentosans each raise absorption, without numbers [6].

I also saw a Cereal Chemistry abstract. It says bake water absorption correlates significantly with flour protein, and that yield loss rises as absorption rises [7]. I read it only as a search summary. I did not open the full paper.

Method

Define one ratio:

$E = H - A$

Here $H$ is the recipe hydration and $A$ is the flour's absorption, both in percent of flour weight. $E$ is the water the dough carries beyond what the flour binds. In grams, $E$ is the free water per 100 g of flour. I call it "excess water". It is my own working measure. It is not a standard baking term, and no one has validated it against crumb.

Then I need $A$ for each flour. I have no measured $A$ for a 10.5% flour. So I estimate the change in $A$ per point of protein from the mechanics in [4].

• One gram of protein per 100 g of flour replaces one gram of native starch.
• Protein absorbs 2 g of water per gram. Native starch absorbs 0.3 g.
• The net gain is 2 minus 0.3, which is 1.7 g of water per 100 g of flour. That is 1.7 absorption points per protein point.

This is a mechanistic bound, not a fit. It ignores that protein quality differs, and it takes the vendor's two uptake figures at face value [4].

I set the 13% flour at $A$ = 63.5%, the farinograph value in [5]. Then:

• The 13% flour: $A$ = 63.5%.
• The 10.5% flour: protein is 2.5 points lower. That gives $A$ = 63.5 minus 2.5 × 1.7 = 63.5 minus 4.25 = 59.25%.

Result

At 75% hydration:

Flour	Protein	Absorption $A$	Excess $E$	Excess water in 1,000 g flour
Stronger flour	13.0%	63.5%	11.5 points	115 g
Weaker flour	10.5%	59.25%	15.75 points	157.5 g

The weaker flour carries 42.5 g more free water per 1,000 g of flour. The ratio is 15.75 divided by 11.5, which is 1.37. The same 75% is 37% wetter, by this measure, in the weaker flour.

The same arithmetic runs in reverse. To give the 10.5% flour the same excess as the 13% flour (11.5 points), the recipe needs $H$ = 59.25 + 11.5 = 70.75%. That is about 71%. On 1,000 g of flour, the water drops from 750 g to about 708 g.

How big is that against ordinary spread between recipes? The book's recommendations run from 75% to 85%, a spread of 10 points [2]. A 4.25-point flour correction is about 42% of that spread. So the flour term is not a rounding error. It is also not larger than the choice of recipe.

The data are mixed on whether recommendations follow the mechanism. Between King Arthur all-purpose (81%) and bread flour (83%), the 1.0 protein point gap goes with a 2-point hydration gap [2]. That is close to my 1.7. From about 10.15% (the middle of Gold Medal all-purpose at 9.8% to 10.5%) to 12.5%, the book's gap is 8 points for 2.35 protein points [2]. That is about 3.4 points per protein point, twice my figure. From 12.5% to a 14.5% high-gluten flour (the middle of 14% to 15%), it is 2 points for 2 protein points, about 1.0 per point [2]. Those recommendations are rounded to whole points and tuned for one method, so I do not over-read the three slopes. They do show that the slope is not constant.

Sensitivity: which assumption moves the result most

I varied the two inputs I chose.

The slope per protein point. This one moves the answer most. The measured data point to a range of about 1.0 to 3.4 points per protein point. At each end, with the baseline fixed at 63.5%:

Slope	Gap in $A$ for 2.5 protein points	$E$ at 10.5%	Excess ratio vs 13%
1.0	2.5	14.0	1.22
1.7 (my base)	4.25	15.75	1.37
3.4	8.5	20.0	1.74

The baseline absorption of the 13% flour. I tried 58% and 66%, with my base slope of 1.7. At 58%, the 13% flour has $E$ = 17.0 and the 10.5% flour has $E$ = 75 minus 53.75 = 21.25. The ratio is 1.25. At 66%, the 13% flour has $E$ = 9.0 and the 10.5% flour has 13.25. The ratio is 1.47. A drier baseline makes the ratio smaller. The range of 58% to 66% is my own choice, and I have no source that bounds it.

Together, the ratio ranges from about 1.22 to 1.74. That is the "22% to 74%" in my opening, and 37% is my central case.

Protein is not the only driver. The two flours in [5] break the rule. The 12.5% flour absorbed more (64.3%) than the 13.0% flour (63.5%). The cause could be starch damage or pentosans, since [4] says damaged starch absorbs about 3 times its weight, ten times more than native starch. Each point of damaged starch adds about 2.7 absorption points on that figure, more than a protein point does. A flour with a little more damaged starch can offset a protein gap. So my protein-only estimate has an error I cannot bound without measured damaged starch for the flours in question. That is why I do not trust a single number from this method for a given flour.

What the sensitivity shows. The direction is stable: weaker flour means more free water at the same hydration. The size is not stable. The slope assumption alone moves the ratio between 1.2 and 1.7.

What I could not test, and what I now think

I did not test crumb openness. The link from free water to crumb is a claim by the flour maker [3] and a general statement in a review of crumb structure. That review reports that more water gives a coarser crumb and that better protein quality gives more uniform gas cells [8]. I read that only as a search summary. I did not open the review, so I treat it as weak support. I also have no evidence on skill of touch. A baker who adjusts by feel in the bowl may erase this gap before it reaches the oven. My ratio cannot see that.

My thesis changes in two ways.

1. The strong form is withdrawn. I cannot say quoted hydration "predicts crumb poorly". I have no crumb data. I have only the water side.
2. The weak form stands. The same quoted hydration means a different water load in different flours. The difference is about 4 absorption points, in a plausible range of about 2.5 to 8.5, for a 2.5-point protein gap.

My earlier position was that hydration and flour protein predict crumb openness better than kneading time. This analysis does not touch that claim, since I have no kneading data. It does add a caution to it: the two inputs are not independent. A recipe's hydration number carries a hidden flour term. A fair study of crumb has to enter the flour's absorption beside the hydration, or it will blur one into the other. The queued 200-recipe comparison needs a column for flour protein or absorption before it will say anything.

A test you can run in one afternoon

Take 1,000 g of one recipe in grams: 1,000 g flour, 750 g water, 20 g salt, 200 g leaven, which is the Tartine ratio in [1]. Make it twice, once with all-purpose flour near 10.5% protein and once with bread flour near 13%. Hold the dough at 24 °C and bake both at 245 °C.

Check the flour label for protein. Then make a third dough with the low-protein flour and 708 g of water, which is 70.75% hydration. My estimate says the third dough should feel closest to the bread-flour dough at 75%. Cut the loaves cold. Photograph each crumb face, and count holes wider than 1 cm in a 10 cm square.

If my 4.25-point correction holds, the third loaf will resemble the 13% loaf more than the 75% low-protein loaf does. If the 75% loaf from the weak flour still matches the strong-flour crumb, then my slope is too large, and I would lower my estimate toward 1.0. A single bake is one trial, so repeat it before you believe any gap.