Duck, cooked, skin not eaten vs Duck, roasted, skin eatenUSDA FDC 2706138 vs 2706139
A controlled equal-weight comparison of two exact USDA survey-food records, with source descriptions, representative amounts, listed formulation inputs, and interpretation limits kept visible.
Equal-weight evidence
Nutrition per 100 grams
Both columns use the same mass, which compares nutrient concentration without allowing the two representative source amounts to change the denominator. “A +” or “B +” reports the arithmetic difference between the cited values; it is not a quality score.
| Nutrient | A · Duck, cooked, skin not eaten | B · Duck, roasted, skin eaten | Difference |
|---|---|---|---|
| CaloriesMaterial under the publication gate | 200 kcal | 336 kcal | B +136 kcal |
| ProteinMaterial under the publication gate | 23.4 g | 18.9 g | A +4.5 g |
| Total fatMaterial under the publication gate | 11.2 g | 28.2 g | B +17 g |
| Carbohydrate | 0 g | 0 g | Same reported value |
| Dietary fiber | 0 g | 0 g | Same reported value |
| Total sugars | 0 g | 0 g | Same reported value |
| Sodium | 227 mg | 221 mg | A +6 mg |
| Saturated fatMaterial under the publication gate | 3.9 g | 9.6 g | B +5.7 g |
| Potassium | 251 mg | 203 mg | A +48 mg |
| Calcium | 12 mg | 11 mg | A +1 mg |
| Iron | 2.7 mg | 2.7 mg | Same reported value |
Values come from the two cited USDA FoodData Central Survey Foods records. Feast does not average records, impute missing values, or infer a nutrient from the food name.
Calories difference
136 kcal per 100 g
Duck, roasted, skin eaten reports 136 kcal more calories per 100 grams than Duck, cooked, skin not eaten (336 versus 200 kcal).
The absolute difference is 40% of the larger reported value. That percentage describes these two records only; it is not an estimate of biological effect or expected variation.
Protein difference
4.5 g per 100 g
Duck, cooked, skin not eaten reports 4.5 g more protein per 100 grams than Duck, roasted, skin eaten (23.4 versus 18.9 g).
The absolute difference is 19% of the larger reported value. That percentage describes these two records only; it is not an estimate of biological effect or expected variation.
Total fat difference
17 g per 100 g
Duck, roasted, skin eaten reports 17 g more total fat per 100 grams than Duck, cooked, skin not eaten (28.2 versus 11.2 g).
The absolute difference is 60% of the larger reported value. That percentage describes these two records only; it is not an estimate of biological effect or expected variation.
What differs in the source descriptions
The exact names carry information the numbers cannot.
USDA assigns both records to “Turkey, duck, other poultry,” but the topic label is broad. Record A is “Duck, cooked, skin not eaten.” Record B is “Duck, roasted, skin eaten.” Feast paired them because their contrast-neutral food identities are an exact match, not merely because they share a generic category.
The controlled name parser found 2 interpretable distinctions. These labels organize wording already present in the names; they do not establish that a distinction caused a nutrient difference.
| Description dimension | Record A | Record B |
|---|---|---|
| Preparation wording | No matching controlled term in the source name | baked, roasted, or broiled |
| Skin wording | skin not eaten | skin eaten |
Even a clean name contrast can carry hidden companions. A fried record may also include coating; a restaurant record can use a different recipe; a reduced-sodium description can differ in more than salt; and a source amount can reflect a different reporting convention. That is why this page reports the complete pair rather than treating one word as an intervention.
Portion context
A representative source amount is not a shared “serving.”
Record A uses Typical reported amount at 85 g. Record B uses Typical reported amount at 85 g. Those labels and gram weights come from their respective records. They are useful for scaling the chosen profile, but they are not recommendations and need not describe equal amounts.
The representative amounts do not cross Feast’s material portion-difference threshold of both 15 grams and 20%. They are still separate source measures, so the 100-gram table remains the appropriate place for an equal-weight comparison.
A cup, piece, slice, sandwich, fillet, or fluid ounce can have different gram weights across records. A matching label is not enough: check the listed grams. If accuracy matters more than convenience, weighing the food or using a current package label can be more specific than a generic survey-food measure.
Record A measures
- 1/2 duck (yield after cooking, bone and skin removed)220 g
- 1 leg (drumstick and thigh) (yield after cooking, bone and skin removed)75 g
- 1 wing (yield after cooking, bone and skin removed)40 g
- 1 breast95 g
- 1 cup, cooked, diced140 g
- Quantity not specified85 g
Record B measures
- 1/2 duck (yield after cooking, bone removed)380 g
- 1 leg (drumstick and thigh) (yield after cooking, bone removed)90 g
- 1 wing (yield after cooking, bone removed)50 g
- 1 breast120 g
- 1 cup, cooked, diced140 g
- Quantity not specified85 g
USDA source formulation evidence
The listed inputs provide another way to audit the match.
FNDDS includes input-food rows behind each standardized survey-food record. Record A’s leading listed inputs include Duck, domesticated, meat only, cooked, roasted, Salt, table. Record B’s include Duck, domesticated, meat and skin, cooked, roasted, Salt, table. Ingredient-code overlap is 33%, and normalized listed-input-weight variation is 100%.
Record A · 2 input rows
Duck, cooked, skin not eaten
- Duck, domesticated, meat only, cooked, roastedIngredient code 514299.6% of listed input weight
- Salt, tableIngredient code 20470.4% of listed input weight
Record B · 2 input rows
Duck, roasted, skin eaten
- Duck, domesticated, meat and skin, cooked, roastedIngredient code 514099.6% of listed input weight
- Salt, tableIngredient code 20470.4% of listed input weight
1 shared coded input
The largest shared entries include Salt, table. Shared codes can still carry different listed weights.
1 A-only and 1 B-only coded inputs
A-only entries include Duck, domesticated, meat only, cooked, roasted. B-only entries include Duck, domesticated, meat and skin, cooked, roasted.
Important: these weights belong to USDA’s source formulation. Their percentages are shares of the listed input weight, not a reconstructed plate, final cooked yield, visitor serving, ingredient declaration, or estimate of what is visible in a photo.
Choosing the closer record
Match context first; scale second.
- 01
Read both names literally
Compare cooking method, coating, edible skin or peel, sauce, additions, venue, source form, and food subtype. For this pair, the full A and B descriptions above are the evidence—not shortened search labels.
- 02
Use listed inputs as a check
If the name still leaves ambiguity, compare the leading and unique source inputs. They can reveal that two near names refer to meaningfully different standardized formulations, but they do not reconstruct your dish.
- 03
Choose an amount for that record
After choosing A or B, use its own named measure and gram weight. Do not transfer a cup, slice, or “typical amount” from the other record as if the measures were interchangeable.
- 04
Account for missing details
Brand, current restaurant recipe, cooking oil, drained weight, toppings, substitutions, and unfinished portions may not be captured. A label or weighed recipe can supersede a generic reference when it better matches the actual food.
Limits and responsible interpretation
This is a comparison of records, not a causal study.
The two records come from 2021–2023 FoodData Central Survey Foods data. The nutrient table faithfully compares their published values on the same weight basis. It does not measure two plates prepared at the same time, isolate one ingredient, or estimate the uncertainty around either value.
Several details can change together. Moisture affects concentration per 100 grams. Coating or sauce can add carbohydrate and sodium. Edible skin or fat changes the portion represented. Restaurant, school, frozen, canned, fresh, home-recipe, and ready-to-heat records may use distinct standardized formulations. The source descriptions and input lists help show those differences, but they cannot assign cause.
A higher or lower number is not an overall verdict. Protein, fiber, sodium, energy, fat, micronutrients, preferences, allergies, budget, portion, and an individual’s goals answer different questions. Feast therefore reports directions and magnitudes without applying a universal “healthier” label.
This page is educational and does not diagnose a condition, prescribe a target, or replace advice from a qualified professional. For an allergy, medically directed diet, or strict ingredient limit, use verified ingredient information and guidance appropriate to that situation.
How this page qualified
A page is published only when the pair clears materiality and diversity gates.
Both records must occupy the same USDA survey-food topic and share a close food identity after a controlled vocabulary separates preparation, edible-part, source, claim, crust, and sauce wording. Exact contrast-neutral identities receive priority. Close ingredient or form substitutions can qualify when name overlap remains high and the numerical evidence is material.
For primary nutrients, Feast requires both an absolute and relative difference: at least 20 kcal and 12% for energy; 2 g and 18% for protein or fat; 5 g and 15% for carbohydrate; 1 g and 20% for fiber; 4 g and 20% for sugars; or 100 mg and 20% for sodium. A focused one-nutrient pair also needs a material representative-amount or formulation distinction. Rounding-only pairs are not published.
Unordered FDC IDs define a canonical pair, so reversing A and B cannot create another page. Each source record can appear twice at most, each USDA topic contributes no more than 40 pages, and closely repeated identity families are capped. Those rules produced a selective comparison library rather than every mathematical combination in the corpus.
Read the complete nutrition-page methodology for import, nutrient mapping, portion scaling, Daily Value, and uncertainty rules.