Substituting one flour for another, honestly
There is no conversion table, and the reason is that flours differ along four axes at once rather than on a single scale of strength.
The question this answers
I have the wrong flour. What actually changes if I use what I have, and when does it stop working?
Four axes, not one scale
Flours are usually ranked on a single line from weak to strong, and that line hides the variables that decide whether a substitution works.\n\nPROTEIN QUANTITY. How much gluten-forming protein is present. Real, and the least complete of the four on its own.\n\nPROTEIN COMPOSITION. Breadmaking quality depends on the type of the gluten proteins as well as the amount — the ratio between the two families and the character of the larger one. Two flours at the same percentage can behave differently, which is why a protein figure on a bag is a clue rather than a specification.\n\nSTARCH DAMAGE. Milling ruptures some starch granules, and damaged granules take up several times their own weight in water where intact ones take a fraction of theirs. Because damage is a milling variable, absorption tracks the mill as much as the wheat.\n\nWHAT ELSE IS IN IT. Bran and germ in a wholemeal, the pentosans of a rye, the coarse particles of a semolina. These change water uptake and dough behaviour by routes that have nothing to do with protein at all.\n\nA substitution succeeds when the flour you have matches the recipe on the axis that recipe was relying on. That is the whole method.
Which axis was the recipe using?
BREAD leans on protein quantity and composition together, because the network has to be strong enough to hold gas through a long fermentation. Plain flour makes bread; it makes a slightly tighter loaf with less tolerance of a long ferment. That is a difference in result, not a failure.\n\nCAKES AND PASTRY lean on the opposite — the LESS network the better, so a lower-protein flour gives a more tender crumb. Bread flour in a cake produces something chewier and tougher. Whether that is wrong depends on what you wanted.\n\nANYTHING WITH A CHEMICAL LEAVENER BUILT IN is a different kind of swap entirely. Self-raising flour is plain flour plus raising agent and usually salt, so substituting it in either direction changes the leavening as well as the flour, and adding baking powder on top of self-raising flour is how cakes come to taste soapy and collapse.\n\nGLUTEN-FREE FLOURS are not on this scale at all. They rebuild structure from starch and hydrocolloids rather than from protein, so no wheat flour ratio transfers.
Hydration is the adjustment people forget
Almost every flour swap changes how much water the dough wants, and almost every recipe states water as a fixed number.\n\nWholemeal and rye take up more, because bran and pentosans bind water the gluten would otherwise have had. A finely milled flour with more damaged starch takes up more than a coarsely milled one of the same wheat. Semolina takes up water slowly rather than in a different quantity, so a dough that seems stiff at mixing may be slack twenty minutes later.\n\nThe practical response is to hold back some of the water and add it if the dough wants it, rather than to look for a percentage. BakeHQ publishes no substitution ratios for hydration, and that is not caution — the correct figure genuinely varies between two bags of the same nominal flour from different mills.
Where a substitution stops being a variation and becomes a different product
SPELT FOR BREAD WHEAT works, and the dough reaches its peak sooner and passes it sooner. Mix less and watch it rather than the clock.\n\nRYE FOR WHEAT does not work as a straight swap at any meaningful proportion. Rye does not build a comparable gluten network and holds gas by a different mechanism entirely, so a high-rye dough is a different craft rather than a wheat dough with a substitution in it.\n\nSEMOLINA FOR FLOUR changes particle size, water uptake and colour together. Finely milled durum is much closer; coarse semolina is not the same ingredient.\n\nNUT AND COCONUT FLOURS are not flours in the structural sense. They have no gluten to form, and coconut flour in particular absorbs water on a scale that makes one-for-one substitution meaningless.\n\nThe honest test is whether the thing the recipe depended on still exists in the replacement. Where it does not, you are not substituting — you are making something else, which is a fine thing to do deliberately and a poor thing to discover at the oven.
Two claims worth not believing
“HIGHER PROTEIN IS ALWAYS STRONGER.” Quantity contributes and composition contributes independently, and a high-protein flour with unfavourable composition can underperform a lower-protein one. The number on the bag is one of four things you would need to know.\n\n“BREAD FLOUR CANNOT BE USED FOR CAKES.” It can. The result is chewier and tougher, which some recipes want on purpose. Nothing fails; something different is produced.
Related
Covers
The concept the guide reasons from: flours differ along four axes at once, and a substitution succeeds when the replacement matches on the axis the recipe was using.
The commonest swap asked about in either direction, and the one where the honest answer is that something different is produced rather than that something fails.
The swap that changes the leavening as well as the flour, and the route by which cakes come to taste soapy.
Where the substitution works and needs watching rather than recalculating.
The boundary case: at any meaningful proportion rye is a different craft rather than a wheat dough with a substitution in it.
Where particle size rather than protein is what makes the swap fail.