> ## Documentation Index
> Fetch the complete documentation index at: https://docs.egma.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Scores and thresholds

Each grader returns a score from `0` to `1`, or no score when it cannot grade
the evidence. A completed grade includes that score, its details, and the pass
threshold selected for the conversation.

## Read an individual result

| Grade                                          | Individual result |
| ---------------------------------------------- | ----------------- |
| Score is at or above the frozen pass threshold | Passed            |
| Score is below the frozen pass threshold       | Failed            |
| Score is unavailable because grading errored   | Errored           |

Suppose Expected behaviors checks four statements. Three pass, so the grader
score is `0.75`. With a pass threshold of `1`, that grader fails. With a
threshold of `0.75`, it passes. The failed statement remains visible in either
case.

A grading error means Egma could not decide or complete the check. It is
different from a score of `0`, which means the check ran and failed.

## Read the combined score

When every selected grader has a score, the **combined score** is their
unweighted arithmetic mean. Scores of `1` and `0.5` produce a combined score of
`0.75`.

This value helps compare conversations. It has no overall pass threshold.
Egma does not turn it into a pass/fail result for a simulation, test, suite, or
run. Read the individual grades to see which checks passed.

The combined score is unavailable when no grader was selected, grading is still
pending, or a selected grader has no score.

## Separate progress from quality

Grading state describes whether work has finished:

| API state       | Meaning                                                                             |
| --------------- | ----------------------------------------------------------------------------------- |
| `not_requested` | No grader was selected.                                                             |
| `pending`       | Grading is waiting to start.                                                        |
| `running`       | A worker is grading the evidence.                                                   |
| `complete`      | All requested graders returned scores, including scores that fail their thresholds. |
| `error`         | A grader or the grading work ended with an error.                                   |

`complete` does not guarantee that all agent evidence arrived. Grading can
proceed with the available evidence when the wait for agent evidence ends.
Check the simulation's `agentPovIncomplete` field and
[review the evidence source](/guides/start-and-follow-a-run#check-which-evidence-arrived)
when a result appears to omit something the agent did.

A simulation that fails or is canceled before producing a completed trace has
no grading work to wait for. Its API `gradingState` is `null`.

For release automation, define your own policy. The
[CI example](/guides/automate-tests-in-ci) requires every simulation to complete
and every selected grader to pass. It also fails on flagged missing agent
evidence, missing grades, grading errors, and timeouts.