196 recipes, read so you don't have to.
Every recipe is a different answer to the same few questions. Tell us the cookie you are after and we will hand you the closest one anyone has published, plus what to change to suit you.
Each one is measured across all 196 recipes and then converted into this recipe's own amounts. Every edit costs something. The cost is on the card.
Make it more set in the middle
–
More added liquid went with cookies that came out more set in the middle, across 912 versions. Flour protein plus available water plus agitation builds an elastic network that resists the bite.
Costs you: Nothing else moved that we can measure — but this lever sits on a real trade, gooey middle or cooked through, and we cannot put a number on that side of it.
measured across many versions · 912 versions
Make it chewier · flatter and wider · sweeter
–
More sugar went with cookies that came out flatter and wider, across 830 versions. Sugar is hygroscopic in the dough too. It binds water and denies it to the flour, suppressing gluten.
Costs you: Also makes it crisper at the edge.
measured across many versions · 830 versions
Make it more tender · more crumbly
–
More butter went with cookies that came out more tender, across 799 versions. Butter is 16-18% water. Cold or creamed butter locks that water in fat crystals where the flour cannot reach it; melted butter frees it, so gluten forms before baking.
Costs you: Nothing else moved that we can measure — but this lever sits on a real trade, chewy or cakey, and we cannot put a number on that side of it.
measured across many versions · 799 versions
Make it crisper at the edge · less sweet
–
More salt went with cookies that came out less sweet, across 21 versions.
Costs you: Nothing else moved that we can measure. We only detect 7 of the 60 changes cooks describe, so read that as unmeasured rather than free.
thin evidence — treat as a hint · 21 versions
The five above are the ones we would point you to. This is every recipe we measured, if you would rather look yourself.
Every dot is one recipe, placed by what it is made of — not by its title. The clusters are the cookies above; the space between them is everything else people have cooked.
These share the mechanisms and the vocabulary you have just read, each one measured the same way against its own set of recipes.
The useful half of any measurement is knowing where it stops.
We can say one recipe reads further along an axis than another. We cannot put a number on it, and we do not pretend to.
Where we can name something that measurably changes the result, we can say which way it pushes. We still cannot tell you where your own shortbread will land, because that also depends on your kitchen and your equipment. here is what we can tell you when it goes wrong.
There are far fewer published cakey recipes to choose between, so that pick is provisional. We say so on the card rather than hiding it.
A few publishers load their reviews from somewhere we cannot read. Their star ratings count toward how trusted a recipe is; their comments could not tell us anything about what cooks changed.