> For the complete documentation index, see [llms.txt](https://www.degenbot.wiki/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://www.degenbot.wiki/performance-and-methodology/use-performance.md).

# Use /performance

Open the performance report and understand what its figures do—and do not—measure.

Send this message in the private bot conversation:

```
/performance
```

The command is the owner-reported way to view performance. The exact layout and available filters depend on the deployed release and your access. This documentation draft is not a live query of the bot's database.

### Read the context before the percentage

Look for the stream and version, report period, last update, eligible sample size, and outcome definition. If any of these are absent, the figure is less informative; request the missing context rather than assuming a definition.

| Item                         | Question it should answer                                         |
| ---------------------------- | ----------------------------------------------------------------- |
| Stream / version             | Which method produced this set of alerts?                         |
| Period                       | What dates or evaluation window does it cover?                    |
| Sample size                  | How many eligible observations were actually scored?              |
| Accuracy                     | What fraction satisfied the stated success rule?                  |
| Null / baseline              | What comparison method was used?                                  |
| Edge                         | How far above or below the baseline is the displayed rate?        |
| Pending / expired / excluded | What happened to observations outside the success/failure totals? |
| Updated at                   | How current is the report?                                        |

These are interpretation questions, not a guarantee that every field appears on screen.

### Three checks worth making

**Compare like with like.** A Fast Move report and a Consensus report may answer different questions. A higher percentage does not automatically make one better.

**Check the denominator.** A report of resolved observations may leave open alerts out of its accuracy calculation. That can be appropriate if clearly stated, but you should know how many remain and how expiries are counted.

**Separate alert results from account results.** An observation can resolve correctly without being executable at the reference price when a subscriber reads it. Fees, spread, slippage, and the subscriber's decisions are a different calculation.

### Empty or unchanged reports

An empty view can mean there are no eligible completed observations in that view, access is limited, or reporting is unavailable. It is not proof of a zero-percent strategy result.

An unchanged total can mean alerts remain open or the report has not refreshed. Check the period and update time; do not assume a missed result is a win.

### Historical screenshots

Keep the original date and version attached to a screenshot. Do not combine old and new methods, or reinterpret an older number after the resolution rules changed.

Continue with **Accuracy, null, and edge** for a worked example, or **Evaluating results fairly** for how a test should be interpreted.