Statistics
Bayesian A/B test (Beta) API
Updates Beta posteriors for two variants from conversion counts and computes P(B beats A) by exact numerical integration plus expected lift and posterior means/variances. Answers 'What is the probability that variant B is better than A?', 'What is the expected lift in a Bayesian A/B test?'.
Price$0.1per request
MethodPOST
Route/v1/stats/bayesian-ab-test
StatusLive
MIME typeapplication/json
Rate limit120/minute
Cache0s public
statisticsbayesianab-testbetaposteriorconversionexperimentprobability
API URL
Integration docshttps://x402.hexl.dev/v1/stats/bayesian-ab-testExample request
{
"successesA": 100,
"trialsA": 1000,
"successesB": 120,
"trialsB": 1000
}Example response
{
"posteriorA": {
"alpha": 101,
"beta": 901,
"mean": 0.100798,
"var": 0.00009037
},
"posteriorB": {
"alpha": 121,
"beta": 881,
"mean": 0.120758,
"var": 0.00010586
},
"probBbeatsA": 0.923183,
"probAbeatsB": 0.076817,
"expectedLiftBoverA": 0.19802,
"prior": {
"alpha": 1,
"beta": 1
},
"rating": "B likely better",
"interpretation": "P(B > A) = 92.32%; posterior means A=0.1008, B=0.1208; expected lift 19.8%."
}Input schema
{
"type": "object",
"required": [
"successesA",
"trialsA",
"successesB",
"trialsB"
],
"properties": {
"successesA": {
"type": "integer",
"examples": [
100
]
},
"trialsA": {
"type": "integer",
"examples": [
1000
]
},
"successesB": {
"type": "integer",
"examples": [
120
]
},
"trialsB": {
"type": "integer",
"examples": [
1000
]
},
"priorAlpha": {
"type": "number",
"examples": [
1
]
},
"priorBeta": {
"type": "number",
"examples": [
1
]
},
"samples": {
"type": "integer",
"examples": [
4000
]
}
}
}Output schema
{
"type": "object",
"additionalProperties": true
}