Predict Bet

This is the main API call to check whether a bet is fraudulent or not.

All betting systems use internal IDs to manage users, sports, markets and more. You will need to use your internal ids to map the model. The model will analyze your internal id and make all the necessary assessments.

We always assume that at the moment you send us the prediction request the bet was placed at most a few seconds before. If this is not the case, from the Betraud -> Fraud AI -> AI Settings page, deselect the Use AI Datetime Evaluation parameter.

The Bet must consist of:

Parameter	Explanation
userId	The user's internal ID. This is a fundamental parameter for understanding and evaluating a user's history and the possibility of real fraud.
amount	Bet amount
sport	Your Sport ID
match	Your Match ID
market	Your Market ID
runner	Your Runner ID
spread	Your Spread ID. Can be null.
fixture	Fixture / Odd of bet

How does it work for a multiple bet?

Multiple bets are handled like a single bet, as multiple bets are a series of single bets grouped together usually with a multiplier. You will then have to divide the multiple bet into many single bets and send them to us. Don't worry, FraudAI will automatically detect that it is a multiple bet and will evaluate the bet as if it were unique, even without a multiplier.

We will create a specific endpoint for multiple betting in Q3 2024.

Predict Bet

POST /predict

<Description of the endpoint>

Headers

Name	Value
Content-Type	application/json
Authorization	Basic <token>

Body

Name	Type	Description
userId	number	Your internal userId
bet.amount	float	Bet amount
bet.sport	number	Your sport ID
bet.match	number	Your match ID
bet.market	number	Your market ID
bet.runner	number	Your runner ID
bet.spread	number	Your spread ID. Can be null.
bet.fixture	float	Fixture / Odd of bet

Body Example

{
  "userId": 4678,
  "bet": {
    "amount": 22.50,
    "sport": 4,
    "match": 390229384,
    "market": 3942,
    "runner": 12289,
    "spread": 783,
    "fixture": 2.18
  }
}

Body Example without spread

Response

200

{
  "userId": 4678,
  "prediction": "fraud"
}

400