Anchor Strategy | DimensionsPot

Which measurements to supply and how they affect prediction accuracy and Confidence Scores.

What is an anchor?

An anchor is any body measurement you supply in the request. The API uses anchors to generate predictions for all other dimensions. Every request needs at least one anchor.

The engine accepts any of the 130 output dimensions as an input anchor — including uncommon measurements like knee height or bitragion arc.

The four anchor tiers

The confidence_score on each output dimension reflects which anchor tier your input falls into:

Tier	What you supplied	BONE score	FLESH score
`PRIMARY_RICH`	Height + weight + ≥1 circumference	~87	~80
`PRIMARY_BOTH`	Height + weight	~85	~78
`PRIMARY_ONE`	Height OR weight (not both)	~79	~62
`SECONDARY`	Foot length, knee height, span, etc.	~74	~67
`TERTIARY`	Any other single measurement	~69	~62

Scores decrease by up to −10 points when primary anchors (height/weight) are missing and the engine infers them via dynamic imputation.

PRIMARY_RICH: adding a circumference

Supplying any circumference alongside height and weight unlocks the PRIMARY_RICH tier:

"anchors": {
  "body_height": 1780,
  "body_mass": 82.0,
  "waist_circumference_omphalion": 880.0
}

Supported precision anchors:

hip_circumference
waist_circumference_omphalion
chest_circumference
neck_circumference
wrist_circumference

Adding all five produces the largest accuracy gains for soft-tissue (FLESH) predictions: waist breadth, hip breadth, calf circumference, and upper arm circumference all improve measurably.

Supplied anchors are preserved as-is

Any dimension you supply as an anchor is returned in body_dimensions with type: "MEASURED" and confidence_score: 100 — the engine does not overwrite it with a prediction.

"body_height": {
  "value": 1780,
  "type": "MEASURED",
  "confidence_score": 100,
  ...
}

Which anchor to choose if you only have one

You have	Recommended anchor key	Resulting tier
Height in cm	`body_height` (× 10 for mm)	PRIMARY_ONE
Weight in kg	`body_mass`	PRIMARY_ONE
Shoe size	Derive `foot_length` (EU size − 2) × 6.67 mm	SECONDARY
Wrist circumference	`wrist_circumference`	SECONDARY
Any circumference	Use it directly as the key name	TERTIARY or SECONDARY

Dynamic anchor imputation

If neither body_height nor body_mass is supplied, the Imputation Engine reconstructs them from whatever anchor you provide using a second layer of Ridge Regression models (550 models per gender).

This works but carries a confidence penalty: all output scores decrease by up to 10 points because the primary anchors are estimated rather than measured.

Imperial input

Set input_unit_system: "imperial" and use inches for lengths and pounds for mass:

"input_data": {
  "input_unit_system": "imperial",
  "anchors": {
    "body_height": 70.1,
    "body_mass": 180.8
  }
}

The engine normalizes to metric internally. Output unit is controlled by output_settings.output_format.unit_system.