List simulation
As an Admin, Senior Analyst, or Analyst, you can compare the performance of the current list to the list with a pending version.
To compare the performance:
- Open a list with a
and select
Test changes in the upper-right corner of the page. - A dialogue window opens. Check how the current list performs as compared to
the list with pending changes. If you select a big date range, you may see
that the simulation is
for some
time before the results are displayed.
Enter a different date range in the upper-left corner of the window if you want to test the list on different historical data.
List simulation dialogue
Let's have a look at this example to understand the information that shows the list performance.
| Column | Calculation | Description |
|---|---|---|
| Hit rate | Hit rate (%) = hit count / total transaction count | The number of transactions a list hit out of the total number of transactions registered in this date range. In our example, the hit rate is 0.17% for both lists. The current list hit 2659 out of 2575525 transactions, while the list with pending changes hit 2670 transactions. |
| Catch rate | Catch rate (%) = matched fraud count / total fraud count | The number of fraudulent transactions a list caught out of the total number of fraudulent transactions in this date range. In our example, the current list caught 117 out of 191 fraudulent transactions identified in this date range. The list with pending changes, however, caught 146 out of 191 fraudulent transactions. The catch rate is bigger for the list with pending changes – 76.43%. |
| Accuracy | Accuracy (%) = matched fraud count / hit count | Indicates how efficient a list version is. In our example, the current list caught 117 fraudulent transactions out of 2659 transactions it hit. The list is only 4.40% accurate, which makes it pretty inefficient. The accuracy for the list with pending changes is a bit better, but still pretty low. |
| False positive ratio | 1 : (hit count – match fraud count) / matched fraud count | Indicates the ratio between the fraudulent and legitimate transactions. In our example, the current list caught 1 fraudulent transaction for every 21.27 legitimate ones. The list with pending changed performed a bit better. |