Why I eliminated everything instead of swapping one thing

Every piece of sensible advice says to start small. Swap one thing. Replace the soda, keep everything else, let the habit settle, then change the next thing.

I did the opposite. On day one I removed everything I suspected wasn't helping me and kept only the short list of things I was reasonably confident about.

That advice is good advice, and I want to be clear about that before I explain why I ignored it.

Why incremental substitution is usually right

It's built for adherence. If you change one thing at a time, the change is survivable, it doesn't blow up your household, and you're not white-knuckling a diet you'll abandon in three weeks. For someone who wants to eat better and isn't running measurements, it's straightforwardly the better approach.

It's also gentler on the part of this that isn't about food at all. Aggressive restriction has a way of becoming its own problem, and I'd rather say that now than bury it at the bottom.

What changes if you're measuring

I wasn't trying to eat better. I was trying to answer a question, and I had a continuous glucose monitor on my arm telling me what my body did after every meal.

Here's the problem with changing one thing inside a complicated diet: everything else is still moving. If I'd swapped one item and watched my glucose, I'd have been reading a number shaped by a dozen other things I hadn't touched. A good day proves nothing. A bad day proves nothing. I'd have spent months collecting readings I couldn't interpret.

Strip down to a small, known set of foods and the picture changes completely. Now the readings mean something, because there's almost nothing else in the frame.

That's not a diet decision. It's an experimental design decision, and it's why the two approaches aren't really competing — they're answering different questions. Incremental substitution optimizes for sticking with it. Elimination optimizes for information.

I wanted information, quickly, because the number I'd found was heading in a direction my family history made me take seriously.

The part that makes it safe, and the part most people skip

Elimination on its own isn't a method. It's just restriction with better branding.

What makes it a method is that removal is provisional. Everything I took out was taken out to be tested later — one item at a time, under controlled conditions, with a threshold decided in advance and a rule for when a test doesn't count.

I'm in that phase now. Cheese is back. Greek yogurt is back. Sour cream is back. Each one went through the same process: test on a clean day, confirm on a second clean day, and if anything else was going on that could muddy the reading, throw the test out and run it again.

Some of those came back and surprised me. I'd seen spikes after meals that contained them — but those meals contained a lot of other things too. When I finally tested them properly, on their own, they passed.

I can't tell you why they passed. The original observations were confounded, so maybe those foods were never the problem. But my fasting insulin was a very different number when I removed them than when I retested, and a body that has changed that much may simply handle them differently now. Those two explanations sit on top of each other and I have no way to separate them.

What I can say is narrower: they don't spike me now. That is not the same sentence as they never did.

That's the whole argument for doing it this way. Not that the removal was correct. That the removal was testable, and I went back and checked.

Who this isn't for

If you're not measuring, don't do this. Without data coming back, elimination gives you all of the difficulty and none of the information — you'd be restricting on a guess and calling it rigor.

If you take insulin, a sulfonylurea, metformin or any other glucose-lowering medication, this needs your prescriber before it needs anything else. Cutting carbohydrates sharply while on those can drop your blood sugar too far, and adjusting a dose to match is a clinical decision, not something to work out from a monitor and a blog post.

If you have any history of disordered eating, this is the wrong tool, and I'd say that even if you are measuring. A protocol that starts with "remove almost everything" is not a safe thing to hand that version of yourself.

And if you don't have a written plan for putting things back before you take them out, you don't have a protocol. You have a diet that's going to get narrower.

What I'd say instead

Not "eliminate everything." Rather: match the method to the question.

If you want to eat better and feel better, change one thing and let it stick. That's the right answer and it's the one most people need.

If you're trying to find out what specifically your body does with specific foods, and you have a way to see the answer, then a clean baseline is worth more than a comfortable one — as long as you've already decided how things get to come back.