Share

Copy link LinkedInXEmail

The share of eligible people withheld from the activity in a holdout experiment, sized with the minimum detectable effect and the power rather than chosen for comfort.

Your value— 0 of 1 held

A parameter is the model's slot; the value in it is yours. It is kept in this browser and written to the model by nothing; share it as a link and it enters the browser that opens it.

• Holdout share % standardbenchmarkexperimentdecision

Share your settings Clear

Questions answered from here— 3

• In what unit is a parameter measured?
• Which evidence designs does a parameter set?
• On what basis does a parameter take its default value?

Compose from here— 11

Holdout share can stand in 11 roles across the patterns this model declares. Pick one and the answer is drawn here, from the published graph.

• The calibration loop board → holds Metrics, Evidence Designs · the shape
• Setting board → holds Time Windows, Geographies, Audience Segments, Setting Bases, Units · the shape
• In what unit a parameter is measured board → holds Units · the question
• Which metrics a parameter sets board → holds Metrics · the question
• Which attribution models a parameter sets board → holds Attribution Models · the question
• Which evidence designs a parameter sets board → holds Evidence Designs · the question
• Which instruments a parameter sets board → holds Signals · the question
• Which purposes a parameter bounds board → holds Processing Purposes · the question
• Which tactics a parameter constrains board → holds Tactics · the question
• By which design a parameter is calibrated board → holds Evidence Designs · the question
• On what basis a parameter defaults board → holds Setting Bases · the question

Related— 4

• parameter of →Holdout experimentEvidence Designs
• measured in →PercentUnits
• defaults on →DecisionSetting Bases
• grounds on →MeasureEntities