# How to configure surprise shorties

When running a **skill-based treatment (SBT)** session with the **Target** set
to **AUTOMATIC**, the **Target** CAB will be randomly chosen based on:

- the current
**SBT level**of the student, and - the
**{CAB}RewardProbability**expressions.

The chosen target will never exceed the current **SBT level** of the student.
But an earlier target (surprise shorty) can be chosen by evaluating each
**{CAB}RewardProbablity** expression in order from **cFCR** to the current
**SBT level**. If any of the expressions resolve to **true**, then that CAB
becomes the target of the next EO trial.

## Surprise shorty example

The current **SBT level** of the student is **CAB1** and the
**{CAB}RewardProbability** expressions resolve to the following values:

CAB | Resolved value | Target probability |
---|---|---|

cFCR | 0.333 | 0.333 |

TR | 0.500 | (1 - 0.333) * 0.500 = 0.333 |

CAB1 | 1.000 | (1 - (0.333 + 0.333)) * 1.000 = 0.333 |

In this example, **cFCR** will be randomly chosen as the surprise shorty 33.3%
of the time, **TR** will be randomly chosen as the surprise shorty 33.3% of
the time, and **CAB1** will be the default target 33.3% of the time.

## Default shorties

Using the default configuration for all **{CAB}RewardProbability** expressions,
the probability of rewarding each level based on the current level of the
student can be seen in the following chart: