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

# Read analysis results

> Interpret factor group returns, IC statistics, and stability charts in Qube.

Qube keeps the selected analysis run, its progress state, parameters, and result payload together. Use the result timestamp to switch between historical runs.

## Start with the target group

The factor direction determines which group represents the intended signal:

* **Positive**: the highest factor-value group is the target group.
* **Negative**: the lowest factor-value group is the target group.

Factor return, Sharpe, annualized return, and maximum drawdown use the same directional target group. Do not assume that the first row in the group table is the target.

## Review the key metrics

| Metric            | What it tells you                                                                  |
| ----------------- | ---------------------------------------------------------------------------------- |
| Factor return     | Return of the directionally selected group or factor return series.                |
| Sharpe            | Return relative to volatility for the selected group.                              |
| Annualized return | Annualized performance over the selected period.                                   |
| Maximum drawdown  | The largest peak-to-trough loss for the selected group.                            |
| `IC_mean`         | Average Pearson information coefficient between factor values and forward returns. |
| `Rank_IC`         | Rank correlation between factor ordering and forward returns.                      |
| `IC_std`          | Dispersion of the IC observations.                                                 |
| `IC_IR`           | IC mean relative to IC variability.                                                |
| `IR`              | Information ratio for the selected return series.                                  |
| `t-stat`          | Statistical strength of the measured relationship.                                 |
| `P(IC<-0.02)`     | Share of observations where IC is below -0.02.                                     |
| `P(IC>0.02)`      | Share of observations where IC is above 0.02.                                      |

No single metric proves that a factor is useful. Check the sign, consistency, sample period, and economic explanation together.

## Read the charts and tables

* **Group cumulative return** shows how each group behaved over time.
* **Group return table** compares annualized return, excess return, drawdown, volatility, turnover, Sharpe, and information ratio.
* **Latest data** lists the latest symbols and factor values in the selected direction.
* **IC and Rank\_IC sequence** shows whether predictive strength persists through time.
* **IC and Rank\_IC distribution** shows the spread and outliers of the observations.
* **IC and Rank\_IC decay** shows how the relationship changes across forward horizons.
* **IC and Rank\_IC autocorrelation** shows whether the signal is persistent or rapidly mean-reverting.
* **3D group return and excess return** help you compare the shape of the group spread across time.

<Warning>
  Historical factor results are evidence about one code version, dataset, universe, and period. Validate the signal on a later or separate period before turning it into a strategy.
</Warning>

When the signal is ready for trading rules, use [Turn a factor into a strategy](/research/strategy/factor-to-strategy). Backtest the generated strategy from the Strategy workflow.