Upgrade to Pro
Last updated: August 20, 2025
The Fraud Detection Pro solution enables you to build more complex risk strategies and unlock new or upgraded services and features. For example:
- Backtesting
- Custom lists
- Custom and enhanced rules
- Risk scores
- Risk profiles
- Velocity rules
Information
To enable Fraud Detection Pro, contact your account manager.
For a detailed side-by-side comparison between Fraud Detection and Fraud Detection Pro, see Payment Fraud Detection.
Backtesting enables you to assess the performance of a test risk strategy by simulating how it compares to your current live strategy, using your past payments. This analysis helps you measure the test strategy's detection rates, false-positive rates, and potential impact on past fraud cases.
Backtesting is much faster than shadow testing. It usually takes minutes to simulate the data, whereas shadow testing can take between 24 hours and three months.
All payments within the selected time frame are assessed against your test and live strategies. The results are not compared to the past outcome, because your live strategy could have changed multiple times within that time frame.
You can use these analytics to determine if your test strategy is likely to improve profitability, considering what you already know about your business. For example, how much a fraudulent chargeback costs you.
You can only start backtesting from your client-level account. However, unlike shadow testing, it also assesses strategies at entity level.
Backtesting analytics include:
- Key performance indicators (KPIs), which display the difference in key metrics between your test and live strategies
- An overview of the Pre-Auth and Post-Auth processing decision changes
- The Amount and number of Transactions associated with each of the following processing decisions: Decline, Void, 3DS, Flag, Accept, Capture
Consider the following factors when implementing backtesting:
| Factor | Description |
|---|---|
Decline, trust, and custom lists | Decline lists, trust lists, and rule-leveraging custom lists all use the list entries from the date when the test occurs. They do not use the lists from the original date. This ensures your test and live strategies are compared as fairly as possible. |
Velocities | Velocities set frequency thresholds for certain attributes that appear in your patterns over a date range. All velocities start at 0 from the beginning of the date range – not the actual values when the payment occurred. |
Machine learning scores | Machine learning scores are not recalculated at the time of the backtesting. Instead, historical scores are reused. |
Decisions beyond Checkout.com | In some situations, Checkout.com may not know whether a payment that was originally declined, or that bypassed 3D Secure (3DS), will pass or fail 3DS. If we know the outcome, we apply it, otherwise we assume the following 3DS outcomes:
Similarly, a payment that the issuer declined is declined between pre-authentication and post-authorization. If we do not know if the payment was declined, we assume that authorization would succeed. For example, you have a payment that was declined but is now accepted. With backtesting, we assume that the payment would not have been blocked, and therefore we need to make assumptions about what happened to it. |
Comparing the results of previous backtests enables you to gain insight to inform your strategy development:
- Sign in to the Dashboard.
- Go to the Payments > Fraud > Strategy section.
- On the Strategy builder tab, go to the Outcome comparison section.
- Select the Backtesting tab.
- For the relevant test, select View results.
Information
If you have any feedback about the backtesting functionality, contact your account manager.
To create a custom list:
- Sign in to the Dashboard.
- Go to Fraud > Strategy > Lists > Custom lists.
- Select Create new custom list, and enter a name and optionally a description.
- Select Create.
- To import a template of values, select Import, select the relevant file, and then select Import.
- To add entries manually, select Add entry, enter the value, and select Add.
- To download the list, select Export.
You may want to leverage custom lists to help prevent payments to and from countries with a high risk of fraud. This high-risk countries list is then referenced in verified information rules.
If you have multiple entities, you can create separate high-risk countries lists for them.
You can build fully custom rules, or enhanced rules with multiple attributes and advanced properties. For example:
- Specific metadata – for example,
Product code is 14569 - Same customer attempts to pay with three different cards within one hour
- Same customer receives more than three insufficient-funds declines within 24 hours
- IP address contains the range
98.195AND the email domain is eithergmail.comorhotmail.com
| Rule name and syntax | Description |
|---|---|
Device has more than three disputes within 180 days
| Counts the number of disputes associated with a given device within 180 days across the Checkout.com network, and identifies if there are more than three disputes |
Card spends more than 10,000 USD within seven days
| Sums the total amount in |
A risk profile is a group of rules that determines outcomes based on rule scores and decision thresholds.
To create a risk profile:
- Sign in to the Dashboard.
- Go to Fraud > Strategy > Profiles.
- Select the relevant tab – Pre-auth or Post-auth.
- Select Create new risk profile.
- Enter the risk profile name, and then select Create.
- To add a rule to the profile, select Assign rule.
- You can add an unlimited number of rules.
- The order of the rules is not important.
- Select the rule from the Rule dropdown, assign a rule score, and then select Assign rule.
- Rule scores can be a negative or positive number.
- Under scoring decisions, use the toggles and sliders to configure your score decisions and ranges. These set the outcomes based on the total score.
- When you profile is complete, select Publish changes.
When the risk profile is published, payments are assessed against every rule. If a payment meets the rule's criteria, it receives the rule's score. The payment's risk score is the sum of all rule scores the payment met the criteria for. If the risk score falls within the range of a scoring decision, that outcome is applied to the payment.
For example, you may decide to decline all payments with a risk score above 90, and to require 3DS challenge for all payments with risk scores between 70 and 90.
Risk scores indicate Checkout.com's assessment of the level of risk for a given payment. They are based on our machine learning models, which are trained on all of our payments and fraud data.
Scores are provided in:
- The
risk.scorefield in the201payment request response - The
Checkout Fraud Scorefield in the Fraud Detection Report - The Payment details page in the Dashboard:
- Go to Payments > Processing > All payments.
- Select the relevant payment to open the Payment details page.
Payments receive a score between 0 and 100 to indicate the risk level as follows:
| Risk category | Score range |
|---|---|
Low risk |
|
Medium risk |
|
High risk |
|
Enhanced risk scores leverage more powerful machine learning models, return granular scores, and significantly reduce false declines.
Information
To enable enhanced risk scores, contact your account manager.
If you do not use Checkout.com risk scores, we can manage scoring for you automatically with no setup required. To enable this, contact your account manager.
To use enhanced scores, you must provide at least one of the following customer data points in your payment requests:
- Device data via the Risk.js package
- Email address
- Name
If you integrate with Flow, the required data is automatically provided for you.
Note
Payment requests that do not include this data do not receive a risk score.
To check if a payment received a score, use the following rule syntax: exists(:score:)
Enhanced risk scores use the following thresholds:
| Risk category | Score range | Estimated payments in category |
|---|---|---|
Safe |
| 10% |
Low risk |
| 40% |
Medium–low risk |
| 45% |
Medium risk |
| 4% |
Medium–high risk |
| 0.8% |
High risk |
| 0.2% |
Fraud Detection Pro enables you to use additional properties in your velocity rules. For example:
The Cumulative spend USD rule checks the total amount spent in US dollars (USD) on a specific attribute within a time frame.
For example, spend_usd (card_number, 24h, attempted) > 1000 identifies when attempted payments on a specific card exceed 1,000 USD within 24 hours.
The Relative rule checks how often one attribute occurs in relation to another attribute.
For example, relative_velocity (email_per_cardholder_name, 30d, attempted) > 10 identifies when the number of email addresses for a single cardholder goes above 10 in any 30-day period of attempted payments.
The Network velocity rule checks how often a specific attribute occurs across Checkout.com's network within a given time frame.
For example, network_count(device, 180d, dispute_raised) > 3 identifies when more than three disputes have been raised against payments associated with a single device within 180 days.