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

# Performance methodology

Measure share returns and explain differences from the target strategy.

Measure what one share was worth at the start and end of the same period. Compare that return with the strategy target, then account for fees, trading costs, and missing observations.

## Actual vault return

Compare the value of the same unit, one share, at two points in time.

For a positive starting NAV per share, the simple period return is:

```
return = ending NAV per share / starting NAV per share - 1
```

Use contract-derived prices over a stated UTC interval. Report the observation blocks, fees already reflected in NAV, missing data, and any emergency or degraded periods. Deposit and withdrawal fees affect an individual investor's proceeds separately from this share-price return.

## Configured directional benchmark

The benchmark describes the chosen market view. It is a comparison, not a promised return.

Build the benchmark from basket weights, percentage returns, direction, and target leverage. Use the settings that applied to the period being measured. If the basket or leverage changed, split the period or document how the change is handled. Do not apply today's settings to earlier history without explanation.

This simplified benchmark excludes running costs and rebalance timing. See [Formulas](/reference/formulas.md).

## Tracking residual

The residual is the gap between what the vault returned and what the simple benchmark returned.

```
tracking residual = actual vault return - configured benchmark return
```

Subtract percentage returns on the same basis. The difference is in percentage points.

Residual can reflect the EVM buffer, funding, fees, spreads, slippage, rebalance timing, drift, pending actions, and compounding. It is not automatically an error.

## Drawdown

Drawdown measures the largest fall from an earlier high during the observed period.

Report maximum peak-to-trough NAV-per-share decline over the same observation set, together with sampling frequency and coverage.

## Fees and execution costs

A trading fee and a worse execution price are different costs. Report them separately when the records support that distinction.

Report observed exchange and contract-based builder fees separately. The EVM builder charge is recorded by vault fee events and liabilities; a zero or absent HyperCore fill `builderFee` field does not prove that the vault charged no EVM fee. Report slippage only when a fill is attributed to a recorded trade intent and reference price. Total execution cost remains partial when fills, contract fee events, or attribution are incomplete. A report containing only HyperCore fill fees may omit the contract-based builder charge.

## Sources and coverage

A reader should be able to trace a chart back to the observations used to build it.

Check the vault, network, time window, data source, and missing observations before comparing charts. Returns calculated from different periods are not directly comparable.

## Backtests and simulations

A simulation shows what a model would have done. It is not evidence of a live vault return.

Backtests and simulations must be labeled historical research and must never fill a gap in missing live protocol performance.