Allocation and rebalancing
Allocation determines how portfolio capital is distributed among the component strategies. Rebalancing determines when those allocations are reset. Configure both in the Portfolio Builder, or edit a saved portfolio's Configuration.
Fixed weights
Enter a positive percentage for each strategy, with a total of 100%.
- Equal weights gives every strategy the same allocation.
- Normalize rescales the entered weights to total 100% while preserving their proportions. For example, 20% and 30% become 40% and 60%.
- For a single strategy, use 100%.
Fixed weights are targets for each scheduled rebalance. Between rebalances, the strategies' values change with their returns, so their actual weights can drift.
Example: restore a 60/40 allocation
Suppose a portfolio allocates 60% to strategy A and 40% to strategy B. If A's value grows faster, its actual share can rise above 60%. At the next rebalance, the analysis redistributes the current portfolio value back to the 60/40 targets.
These allocations scale the component NAV series during portfolio analysis. They do not edit the source strategies' leg quantities. To change sizing inside a backtest, use Entry.QtyMultiplier.
Inverse volatility
Inverse volatility assigns more weight to strategies with lower measured return volatility. At each allocation date, MesoSim calculates each strategy's volatility over the lookback window, takes its reciprocal, and normalizes the results to 100%.
For example, if one strategy has twice the measured volatility of another, it receives half as much target weight: approximately 33.3% versus 66.7%.
Choose the estimator
| Volatility calculation | How it works |
|---|---|
| Standard deviation | Uses the sample standard deviation of simple strategy NAV returns in the lookback window. |
| EWMA (RiskMetrics) | Uses an exponentially weighted calculation, giving more influence to recent return observations. Lambda (λ) controls how quickly older observations lose weight; values closer to 1 decay more slowly. |
Lookback (calendar days) accepts 3–365 days and defaults to 60. EWMA defaults to a lambda of 0.94. The estimates use strategy NAV returns, rather than the underlying's Implied Volatility.
The opening lookback period is warm-up and is excluded from charts, statistics, and the other performance analyses. Each strategy needs enough return observations and a positive, usable volatility estimate. A constant NAV series cannot receive a valid inverse-volatility weight.
Choose a rebalance schedule
| Rebalance | Behavior |
|---|---|
| Initial | Allocates once at the start of the analysis, then allows weights to drift. |
| Daily | Rebalances on each subsequent portfolio valuation date. |
| Weekly | Rebalances on the selected weekday, or the first available valuation date on or after it. |
| Custom | Rebalances every configured number of calendar days, using the first available valuation date on or after each due date. |
Every schedule includes an initial allocation. The final valuation records the result without another rebalance. Custom periods accept 1–365 calendar days.
For inverse volatility, new target weights are calculated only when the schedule calls for an allocation. An Initial schedule therefore calculates the weights once; it does not continuously update them.
Inspect the effect
Select Analyze and review, then open Events → Reweighting activity. Inspect the weight before, target weight, and capital change for each strategy. Inverse-volatility events also show the volatility used to calculate the weights.
Compare one allocation change at a time using the results views. Portfolio rebalancing and delta-hedge checks have independent schedules; weekly rebalancing can be combined with daily or intraday hedging.