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

# Review Results

> See an experiment's pass rate, find the items that regressed against the baseline, and see which judge or assertion failed and why.

After an experiment, the first thing to check is **Regressions**: items that passed before and fail now. Those are what your change broke. Then open any failed item to see which check failed and why.

Open an experiment from a dataset's **Experiments** tab, or from the **Experiments** tab on the **Datasets** page, which lists the experiments of every dataset.

## Experiments

The experiments list shows each experiment's number and name, its baseline, the agent and commit it tested, its mode, pass rate, regressions, status, and when it started. Filter by **Agent**, **Status**, **Mode**, and **Started by**, or search by name or number. On the **Datasets** page, filter by **Dataset** too.

| Status | Meaning |
| - | - |
| **Queued** | The experiment is waiting to start. |
| **Running** | Items are running. |
| **Completed** | Every item has finished. |
| **Cancelled** | You cancelled the experiment before it finished. |

Pass rate and regressions are final once the experiment stops. Open an experiment's **⋮** menu to **Open results**, **Cancel** it while it runs, or **Run again**.

## An experiment

The header shows the experiment's name and number, with **Cancel** while it runs and **Run again** once it ends. See [Run again](https://docs.zeroruntime.ai/agent-builder/datasets/run-an-experiment.md#run-again).

Below the header you see the agent and the branch and commit it tested, the judges, the iterations, the baseline, who started it and when, and what it cost. While it runs, a progress bar counts finished replays and conversations.

### Summary

| Card | What it shows |
| - | - |
| **Pass rate** | The share of items that passed, for example *18 of 20*, with how many were flaky, errored, not applicable, or cancelled, and the baseline's pass rate. |
| **Regressions** | Items that passed in the baseline and fail now. |
| **Fixed** | Items that failed in the baseline and pass now. |
| **Still failing** | Items that failed in both. |
| **No change** | Items with the same result in both. |

The comparison cards are filled in only when the experiment has a baseline. Without one they show 0 or —. Choose a baseline when you start the experiment, or use **Run again**, which picks one for you.

![Experiment #13 page with the Run again button, the agent, judges, iterations, baseline, start time, and cost, the Pass rate, Regressions, Fixed, Still failing, and No change cards, the Verdict, vs baseline, Kind, and Tags filters with the Show only regressions switch, and the results grid of three failed conversation items](https://cdn.zeroruntime.ai/zrt/docs-resource/agent-builder/experiment-detail.png)

### Verdicts

Each item gets one verdict across all its iterations:

| Verdict | Meaning |
| - | - |
| **Pass** | Every iteration passed. |
| **Flaky** | Some iterations passed, for example *2 of 3*. Counts as passing in the pass rate and the comparison, but it's worth a look: the agent handles this case only some of the time. |
| **Fail** | No iteration passed. |
| **Error** | The item couldn't run. Not counted. |
| **N/A** | The item doesn't apply to this run, for example an audio judge in a text run. Not counted. |
| **Cancelled** | The experiment was cancelled before the item ran. Not counted. |

An iteration passes when every judge and every assertion on the item passes.

### Every item, every judge

A grid with one row per item and one column per judge, plus an **Assertions** column. Each cell shows how many iterations passed, for example *2 of 3*. Each row also shows the item's kind, its verdict, and how it compares with the baseline: **Regression**, **Fixed**, **Still failing**, or **No change**.

Filter by **Verdict**, **vs baseline**, **Kind**, and **Tags**, or search. Turn on **Show only regressions** to see just what broke.

For a conversation item, select the **Add to dataset** icon at the end of its row to add the simulated call to a dataset. Judges that failed on it are filled in as the new item's checks. See [From a simulation](https://docs.zeroruntime.ai/agent-builder/datasets/items.md#from-a-simulation).

## A result

Select a row to open the item's result. The top shows its verdict and how it compares with the baseline. Pick an iteration to see:

* **Judges & assertions**: every judge's verdict with its reason, and every assertion with what it found, for example *got 4* for **Max turns**, or *The agent never called reschedule\_appointment*.
* **This run**: what the agent produced: its output, its tool calls, and its audio.

![Result panel for Reschedule a dentist appointment in Experiment #13 with a Fail verdict, no baseline result, and iteration 1 showing a passed Max turns assertion and a failed Tool called assertion, with the Open item, Open simulated session, and Add to dataset buttons](https://cdn.zeroruntime.ai/zrt/docs-resource/agent-builder/experiment-result.png)

Select **Open item** to see the item, or **Open simulated session** to open a conversation item's call in [Simulation](https://docs.zeroruntime.ai/agent-builder/simulation/review-results.md#a-conversation), with its transcript and recording. For a conversation item, select **Add to dataset** to add the simulated call to a dataset.

## Fix and re-run

**1. Find what broke**

Turn on **Show only regressions**, open a failed item, and read the judge's reason or the assertion's result.

**2. Fix the agent**

Update the agent in the builder, then [commit](https://docs.zeroruntime.ai/agent-builder/versioning/drafts-and-commits.md) and [deploy](https://docs.zeroruntime.ai/agent-builder/deploy.md) the change.

**3. Run again**

Select **Run again**. The new experiment uses this one as its baseline.

**4. Compare**

Check **Regressions** and **Fixed** on the new experiment to see what moved.

> [!TIP]
>
> An item that fails because the judge was wrong, not the agent, is a judge problem. Edit the item to change its checks, or tighten the judge's criteria in [Judges](https://docs.zeroruntime.ai/agent-builder/evaluations/judges.md).


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.