Fail on purpose

Here is what I hope you get out of this site, and it is not the recipes.

Every number on every page here is the survivor of something that went wrong. The recipes are the record of the failures, not a set of instructions handed down from somewhere. That is the part worth taking.

A while ago I made the burger and salted the meat before it went into the bath. It seemed obvious. Salt seasons, so season it, then cook it. What came out was springy and faintly pink in a way that read as cured rather than cooked, because that is exactly what had happened: salt sealed against ground beef at bath temperature for an hour does not season it, it cures it. The recipe on this site says to season after the bath, and it says why. I only know that because I ruined a burger.

The bread was worse, because it failed slowly. A loaf that looks right, sounds right when you knock it, and is wet through the middle when you finally cut it two hours later. I had scaled the recipe up for a bigger pot using the number the original recipe told me to use. The number was wrong for my oven, and the reason turned out to be simple once I stopped being annoyed about it: a bigger loaf has to drive more water out through the same crust in the same time. Scale is not linear. The recipe now carries two sizes with their own weights instead of a multiplier, less psyllium, and it tells you to bake to an internal temperature rather than to the clock. Three corrections, all of them paid for with a bad loaf.

The coffee is nothing but this. A flash brew at 203 degrees that should have worked and came out thin, when 205 landed sweet. A Yeti brew at grind 3.1 that finished in 2:45 and tasted sharp, when 3.0 gave it the time it needed. An AeroPress XL recipe I was confident about at 22 grams to 320, superseded by 28 to 350 after I actually drank both. None of those are insights. They are just what happened after I got it wrong and paid attention.

The reason failing teaches more than succeeding is not inspirational. It is mechanical. When something works you learn one fact: this combination worked. You do not know which part was responsible, or how much room you had, or what would have happened a step in either direction. When something fails, and you only changed one thing, you learn exactly where the edge is. Ten good loaves in a row tell you less than one that came out wet, provided you can say what you did differently.

Which is the only real rule: change one thing. If I had cut the psyllium and dropped the scale and shortened the bake all in the same attempt, I would have had a better loaf and no idea why. That is a worse outcome than the wet one. You cannot build on an accident.

So use the recipes here. They work, and they are written down carefully precisely because they cost something to get right. But the useful thing is not that I found the number. It is that the number came from somewhere, and yours might be different, and the way to find out is to change one thing and pay attention to what happens next.

Cook it wrong once on purpose. You will learn more in that hour than in the five times it came out fine.