Get Ranked List for Contextual Bandit

Returns a ranked list of variants for a contextual multi-armed bandit (autotune) experiment. The ranking is based on predicted performance.

POST /v1/get_ranked_list

Get Ranked List for Contextual Bandit

cURL

curl --request POST \
  --url https://api.statsig.com/v1/get_ranked_list \
  --header 'Content-Type: application/json' \
  --header 'statsig-api-key: <api-key>' \
  --data '
{
  "configName": "product_recommendation_bandit",
  "user": {
    "userID": "user-123"
  }
}
'

200

[
  {
    "name": "variant_a",
    "value": {
      "recommendation_algorithm": "collaborative_filtering"
    },
    "rule_id": "rule_123",
    "score": 0.85
  },
  {
    "name": "variant_b",
    "value": {
      "recommendation_algorithm": "content_based"
    },
    "rule_id": "rule_124",
    "score": 0.72
  }
]

Authorizations

statsig-api-key · string · header · required

SDK API key (Server Secret or Client SDK Key)

Body

application/json

configName · string · required

Name of the contextual bandit (autotune) experiment

Required string length: 2 - 100

Pattern: ^[a-zA-Z0-9_\- ]+$

user · object

User object containing identification and attributes for evaluation. At minimum, provide at least one identifier.

Show child attributes

statsigMetadata · object

SDK metadata for tracking SDK type, version, and other diagnostic information

Show child attributes

Example:

{
  "sdkType": "js-client",
  "sdkVersion": "4.20.0",
  "exposureLoggingDisabled": false
}

Response

200

application/json

Ranked list of variants

name · string · required

Variant name

value · object · required

Variant parameter values

rule_id · string | null · required

Rule ID for this variant

score · number<double> · required

Predicted performance score for this variant