Precision confidence, not identity probability.

What the Accuracy Index is, in one sentence

The Accuracy Index is a percentage score that summarizes how tightly the current V3.1 model can constrain adult stature for the inputs provided, using the displayed interval half-width, bone count, measurement condition, consistency between bones, population context, sex assessment, and model type.

What it is not

The Accuracy Index is not the probability that the calculated stature is the correct stature. It is not the probability that the remains belong to a named individual. It is not a court-admissible identification statistic. Antrometric is a stature estimation instrument; identification requires a complete biological profile and case-specific evidence.

The implemented formula, term by term

• H is the half-width of the displayed model interval in centimeters: SEE_combined · 1.645 for 90 percent or SEE_combined · 1.96 for 95 percent.
• B is the bone-count gain: min(24, (n - 1) · 5.1).
• Q is the average quality penalty: complete 0, minor surface loss 4, reconstructed 9, fragmentary 16.
• C is the consistency penalty: 0 for high consistency, 4 for moderate consistency, and 8 for possible outlier.
• S is the scenario and model penalty: unknown sex subtracts 8, generic population subtracts 4, and a non-multivariate combined single-bone model subtracts 1.
• cap is the maximum score allowed for the scenario. Base cap is 92. Single bone caps at 78, unknown sex at 82, generic at 86, moderate consistency at 88, possible outlier at 84, and any fragmentary measurement at 76. The lowest applicable cap wins.
• clamp(min = 32) is the floor. V3.1 never displays a value below 32 because even a weak adult measurement scenario can still carry limited orientation value.

Worked example A: three complete European male bones

Inputs: European Male, 95 percent interval, femur 47.0 cm, tibia 38.5 cm, humerus 33.0 cm, all complete, high consistency, no diurnal correction.

• H = 5.34 cm from combined SEE 2.73 cm at z = 1.96.
• 3.7H = 19.8
• B = 10.2 for three bones.
• Q = 0
• C = 0
• S = 1 because this is a robust combined single-bone model rather than the validated generic femur-tibia multivariate model.
• cap = 92
• AI = clamp(100 - 19.8 + 10.2 - 0 - 0 - 1, 32, 92) = 89

This score is high because the bones are complete, internally consistent, and tied to a known sex and reference family.

Worked example B: single fragmentary generic tibia, unknown sex

Inputs: Generic reference, Unknown sex, 95 percent interval, one fragmentary tibia at 38.0 cm.

• H = 15.43 cm from total SEE 7.87 cm.
• 3.7H = 57.1
• B = 0
• Q = 16
• C = 0 because there is only one bone.
• S = 13: unknown sex 8, generic population 4, non-multivariate model 1.
• cap = 76 because fragmentary measurement is the lowest applicable cap.
• AI = clamp(100 - 57.1 + 0 - 16 - 0 - 13, 32, 76) = 32

This is intentionally weak. The interval, not the point estimate, is the main reportable information.

Worked example C: possible outlier

Inputs: Thai Male, 95 percent interval, femur complete at 45.0 cm and reconstructed tibia at 25.0 cm. The femur implies 169.93 cm; the tibia implies 141.20 cm.

• H = 11.19 cm
• B = 5.1
• Q = 4.5 as the average of complete 0 and reconstructed 9.
• C = 8 because the tibia reaches 3.28 standardized residuals.
• S = 1
• cap = 84 because possible outlier is the lowest applicable cap.
• AI = 50 after clamping and rounding.

The instrument is doing the correct thing here: it keeps the discordant bone visible, names it in the result panel, and lowers precision confidence.

What the number means in practice

For investigators and police officers

If you need a one-line interpretation: the Accuracy Index tells you how constrained the calculator's stature estimate is for this specific input set. A high score does not identify the person. Compare the stature interval with case information; do not compare the score with a missing-person height.

Why caps and a floor exist

The cap prevents weak scenarios from appearing stronger than they are. Unknown sex, generic population context, fragmentary measurement, single-bone input, or possible outlier behavior each limits the maximum score. The floor of 32 prevents the interface from implying that a weak adult measurement has no information value at all.

References

1. Albanese, J., Tuck, A., Gomes, J., & Cardoso, H. F. V. (2016). An alternative approach for estimating stature from long bones that is not population- or group-specific. Forensic Science International, 259, 59-68.
2. Konigsberg, L. W., Hens, S. M., Jantz, L. M., & Jungers, W. L. (1998). Stature estimation and calibration: Bayesian and maximum likelihood perspectives in physical anthropology. Yearbook of Physical Anthropology, 41, 65-92.
3. Mahakkanukrauh, P., Khanpetch, P., Prasitwattanseree, S., Vichairat, K., & Troy Case, D. (2011). Stature estimation from long bone lengths in a Thai population. Forensic Science International, 210(1-3), 279.e1-279.e7.
4. Scientific Working Group for Forensic Anthropology. (n.d.). Stature estimation. National Institute of Standards and Technology.
5. Trotter, M., & Gleser, G. C. (1958). A re-evaluation of estimation of stature based on measurements of stature taken during life and of long bones after death. American Journal of Physical Anthropology, 16, 79-123.