Omega Point runs your portfolio through a set of machine-learning algorithms to generate proprietary insights that identify your portfolio's exposure to inflection points in factor movements. The **Mitigate Factor Drift** insight identifies your portfolio's factor exposures and analyzes each factor's **normalized return**. Based on certain threshold criteria, the insight suggests rebalances to your portfolio in order to mitigate exposure to **oversold** and **overbought** factors in the market.

### What Is Normalized Factor Return?

A normalized factor return is a standardization of a factor return index that uses several Savitzky-Golay filters to provide indicators on the short-term and longer-term levels of the factor return index. The final score is calculated by applying a standardized Z-Score on the difference between the short-term and longer-term filters.

### Overbought and Oversold Factors

A factor is defined as “overbought” or “oversold” if its normalized factor return is greater than 1 or -1, respectively.

A factor is defined as “extremely overbought” or “extremely oversold” if its normalized factor return is greater than 2 or -2, respectively.

Historically, when looking ahead to the next quarter of trading, an extremely overbought or oversold factor tends to revert to its long-term return trend more than 70% of the time. A reversion to the long-term return trend is defined as the normalized factor return moving from its current extremely overbought or oversold levels (i.e. greater than 2 standard deviations) to within 0.5 standard deviations of the long-term trend for at least 5 trading days within the selected period.

The below example shows the recent mean reverting pattern of the Axioma US4 Medium-Term Momentum Factor. As shown in the shaded region in Omega Point's factor profile screen:

- The factor hit extreme overbought levels (+2.00) on November 15, 2017 and peaked at a normalized return of +2.16 on November 21st, 2017.
- Subsequently, the factor reverted to +0.5 standard deviations by December 4th, 2017.

### How Does This Relate to the Mitigate Factor Drift Insight?

The “Mitigate Factor Drift” Insight does the following:

- Identifies
**overbought**and**oversold**candidate factors in your portfolio - Feeds your current portfolio into a factor targeting optimization that seeks to negatively tilt your exposure to
**overbought**factors and positively tilt your exposure to**oversold**factors.

### Factor Targeting Optimization

A factor targeting optimization is different from a traditional mean variance optimizer that purely aims to modulate the exposures in the portfolio, rather than reduce overall risk. For each targeted factor, the optimizer:

- Applies a loss function that seeks to minimize the difference between the current portfolio's factor exposure and a parameterized target factor exposure.
- Adjusts the loss function based on the underlying volatility of the factor. For example, a technical factor such as momentum or market beta will tend to have a higher volatility than a fundamental factor such as value or growth and the optimizer will take this information into account.

For any optimization in our system, we apply a set of default constraints laid out in the Portfolio Insights article.

### Optimization Output

The output of the Mitigate Factor Drift insight is delivered in our insights panel along with the Focus On Your Alpha insight and shows:

- The risk impact to your current portfolio (left pane)
- The historical theoretical performance impact to your portfolio (middle pane)
- The adjustments to the portfolio required to achieve the target exposures (right pane)