The part a flour swap can't explain
Same texture group, still a different cookie
Chewy is chewy is chewy, except it isn't
Hold the texture group fixed and compare like for like. The
13 gluten-free recipes that landed in the same chewy
group as the 381 wheat ones still read as less chewy
on our chew measure than the wheat recipes sitting right next to them in
that same bucket. Same classifier, same group, still a gap. Whatever is
causing that is not something our texture classification is catching,
which means it's coming from somewhere in the rest of the formula, not
from being sorted into the wrong bucket.
The flour axis sees the split, not what's inside it
We track flour protein as a lever because more of it builds more
gluten, and gluten is most of what makes a bite pull rather than crumble
(see what actually makes cookies chewy).
That axis now draws a clean line: all 34 gluten-free
recipes read at its floor value, while the 566 wheat
recipes read higher, most at the standard all-purpose level and a
handful higher still for bread flour. What it can't do is tell almond
flour from oat flour from rice flour from a store-bought blend. Every
gluten-free recipe here reads the identical number, whatever's actually
in the bag.
