How I Calculate Danger per Technique to Obtain Equal Portfolio Weighting
When operating a number of knowledgeable advisors or buying and selling methods in the identical portfolio, equal threat per commerce doesn’t imply equal publicity. Actually, utilizing a set threat per commerce throughout totally different methods nearly all the time results in imbalanced efficiency, the place some methods dominate the portfolio whereas others dilute returns.
The purpose of my threat mannequin is straightforward:
Each technique ought to contribute roughly the identical anticipated annual return to the portfolio.
To attain this, threat should be adjusted based mostly on:
Why Mounted Danger per Commerce Does Not Work
Let’s take a look at widespread errors:
1. Completely different commerce frequency
If each use the identical threat per commerce, Technique B will naturally have a lot increased publicity, even when it performs worse.
2. Completely different holding occasions
Even with the identical variety of trades per week, the technique with longer holding time has:
Due to this, you can’t equalize threat by:
Utilizing the identical proportion threat
Dividing threat by trades per week
Ignoring holding time and volatility
Step 1: Danger Should Be Based mostly on Volatility (ATR)
I base all threat on volatility, not stop-loss dimension or fastened percentages.
Particularly:
Danger is calculated utilizing 1 ATR (Common True Vary)
Often a day by day ATR, because it represents the market’s common day by day motion
This method:
Mechanically adapts to totally different devices (Foreign exchange, Gold, Indices, Crypto)
Adjusts for altering market situations over time
Avoids issues the place worth ranges change however volatility will increase
A 1% transfer at present will not be the identical as a 1% transfer 20 years in the past — volatility-based threat solves this.
Step 2: Outline the Portfolio Goal
Instance portfolio:
This implies:
Step 3: Backtest Every Technique at 1% ATR Danger
For every technique:
Revenue Issue is essential as a result of:
Step 4: Use a Reasonable Revenue Issue Baseline
Backtests typically exaggerate efficiency.
I assume a reasonable long-term revenue issue of 1.2.
That is:
Step 5: Scale Danger Based mostly on Revenue Issue Degradation
Instance 1: Robust Backtest, Wants Increased Danger
However:
Goal is 5% per yr, so:
Instance 2: Weak Backtest, Wants Decrease Danger
If PF improved from 1.1 → 1.2:
Goal is just 5%, so:
Closing Outcome
After adjusting threat this fashion:
Each technique is normalized to the similar anticipated annual contribution
Excessive-frequency methods now not overpower low-frequency ones
Lengthy-holding methods are correctly weighted
Portfolio conduct turns into smoother and extra predictable
If all methods find yourself with the identical real-world revenue issue, they may also produce the identical annual return.
That is the inspiration of a correctly balanced multi-strategy portfolio.
