Recipe Collection

Is it the chicken or the pita?

If you have ever wondered whether the protein on your plate is what's driving your blood sugar, here is the short answer: almost certainly not. I calculated the glycemic load of every dinner in my cookbook — fifty recipes, the same method applied identically to each — and the pattern was so consistent it was almost boring. Every single high-GL dinner was high because of one starch. Not one protein appeared as a meaningful driver anywhere in the fifty. The chicken in Chicken Gyros contributes essentially nothing to the glycemic load. The pita bread contributes 18.3 out of a total GL of 19.1. That is 96% of the number, sitting in the bread.

Some context, because the numbers only mean something if you know where they came from. I am not a dietitian. I am a software developer who crossed from pre-diabetes into diabetes a while back, went looking for recipes that would tell me how much a meal would actually move my blood sugar, and found calories and macros everywhere and glycemic load nowhere. So I built the thing I wanted. That became Glyc: paste in any recipe and it returns the glycemic load of a serving, with the ingredients driving it laid out underneath. The fifty dinners below are what happened when I pointed that engine at a cookbook — one method, applied identically to every recipe, with the arithmetic shown instead of asserted.

I want to be clear that glycemic load is one input among many, and that individual blood sugar response varies considerably from person to person. If you are managing diabetes, pre-diabetes, PCOS, or any other condition, please work with your clinician or dietitian — this is educational, not medical advice. But the arithmetic here is not ambiguous, and understanding it can make the numbers on a label or a calculator a lot less mysterious.

The eight highest-GL dinners, and what's actually driving them

Here is the table I kept staring at. Across these eight dinners, the single biggest contributing ingredient accounts for an average of 85% of the total glycemic load.

Dinner	Total GL	Biggest driver	Driver's GL	Share
Chicken Gyros	19.1	Pita bread	18.3	96%
Maple-Mustard Salmon with Veggies	19.1	Baby multicoloured potatoes	15.2	80%
Greek Chicken Kebabs	19.0	Pita bread	18.3	96%
Skillet Chicken with Orzo and Tomatoes	18.0	Orzo	17.0	94%
Brown Butter Turkey Meatballs with Crispy Sage Potatoes	17.0	Yukon Gold potatoes	13.6	80%
Maple Dijon Salmon and Roasted Sweet Potatoes	16.5	Sweet potato	10.8	65%
Cajun Fish Tacos	14.9	Flour tortillas	11.5	77%
Lemon Pepper Shrimp and Orzo Bowl	14.7	Orzo	13.5	92%

Notice that two of those eight dinners are salmon dishes. The fish is identical in both — a high-quality protein with essentially no carbohydrate. What changes is what sits beside it. The maple syrup adds a small but real contribution in both recipes (GL 3.6 in one, 5.4 in the other), but it is the starch — potatoes in both cases — that sets the ceiling.

The low end: when garlic is the biggest driver

The contrast at the bottom of the list is stark. Garlic Butter Steaks has a measured GL of 0.0 — no meaningful contributor at all. Garlic Butter Salmon comes in at GL 0.1, with garlic itself as the single largest contributor at 0.1. Grilled Skirt Steak Chimichurri is also 0.1, again with garlic carrying the whole number. Beef Tenderloin reverse-seared lands at GL 0.2, with fresh horseradish contributing 0.1. Brisket Flat is GL 0.3, driven entirely by the BBQ rub.

When the biggest driver of an entire dinner is the garlic at 0.1, you are looking at a plate built around protein and non-starchy vegetables. There is no starch to dominate the calculation, so nothing dominates it.

Why it works out this way — the arithmetic, not the opinion

Glycemic load is calculated as: (GI × grams of available carbohydrate) ÷ 100, summed across every ingredient. That means an ingredient contributes to the GL only in proportion to the carbohydrate it actually delivers. Protein contains almost no carbohydrate. Fat contains none. So no matter how much chicken, salmon, steak, or shrimp is on the plate, those ingredients contribute a number close to zero — not because protein is "good" or fat is "safe," but because the formula has nothing to work with.

Starch is the opposite. A serving of orzo, pita, flour tortilla, or potato delivers a substantial amount of carbohydrate, and if that carbohydrate has a moderate-to-high GI, the product of those two numbers dominates everything else on the plate. It is not a nutritional opinion. It is arithmetic.

What makes this non-obvious is that we name meals after the protein. You think "Chicken Gyros." The calculation thinks "pita bread with some chicken on it." You think "Cajun Fish Tacos." The calculation thinks "flour tortillas with seasoned fish." The protein is often the most visually prominent, most expensive, and most culinarily interesting part of the dish — and the least relevant part to the glycemic load number.

What to do with it: the side is the lever

If you want to move the glycemic load of a dinner, the most effective place to look is the starch. One ingredient is doing 80–96% of the work. That means a few things in practice:

• Swapping the protein changes almost nothing. Replacing salmon with chicken in the Maple Dijon recipe would barely register. The sweet potato is GL 10.8 either way.
• Swapping or reducing the starch changes almost everything. The same salmon served alongside roasted broccoli instead of sweet potato would look much more like Garlic Butter Salmon (GL 0.1) than Maple Dijon Salmon (GL 16.5). That comparison is real: Feta, Spinach and Tomato Stuffed Salmon comes in at GL 0.4. The same fish, a different plate, and about forty times the glycemic load.
• Halving the starch roughly halves its contribution. Since that one ingredient is doing most of the work, portioning it down moves the total number more than any other change you could make to the rest of the plate.
• Reputation does not change the math. Sweet potatoes are genuinely nutritious — they carry fiber, vitamins, and a lower GI than white potatoes. They are also a starch, and the calculation does not adjust for reputation. At GL 10.8 per serving, they are still the dominant driver of that dinner. That is not a reason to avoid them; it is a reason to know what they contribute.

None of this is a directive to remove starches from your plate. Glycemic load is one signal among many. Fiber, micronutrients, satiety, enjoyment, and cultural meaning all matter too, and none of them appear in this formula. What the math does is clarify where the leverage actually is, so that if you want to make a change, you are pulling the right lever.

Where to go from here

If you want to run this calculation on your own recipes, the free Glyc calculator at getglyc.com takes any recipe URL and shows the full breakdown — every ingredient, every gram, every step of the math. Disclosure: I built it.

The fifty worked examples behind this post are collected in Dinner by the Numbers, which is free on Kindle through Saturday, September 12 — you can find it here. But if you never pick it up, the point stands on its own: in a dinner, look at the starch first. That is where almost all of the number lives.