Combined algos and correlation

Stratzy Algos
1

I combined few algos ( all are credit spreads and positive correlation ) and app was showing a drawdown of 8% and returns of 170% (Over a period of last 1 year). So I deployed them. Now Stratzy app says my portfolio of algos is very highly volatile since they positively correlated.

But when I tried to combine negatively correlated Algos, I was getting much worse numbers than mentioned above. Why it is so

2

Hi @nijinwilson

When creating an Algo portfolio, the objective should not be to simply combine Algos with positive or negative correlation. Instead, the focus should be on building a well-diversified portfolio where different strategies complement each other across varying market conditions.

Here’s a simple way to think about it:

• Positive Correlation (+1):
The Algos tend to move together. If one performs well or goes through a drawdown, the other is also likely to behave similarly. While such combinations may generate higher returns during favourable market phases, they can also experience larger portfolio-level drawdowns since the Algos react similarly to the same market conditions.

• Negative Correlation (-1):
The Algos tend to move in opposite directions. This can help reduce portfolio volatility, but it may also result in periods where the profits from one strategy are offset by losses from another, leading to relatively lower overall returns during certain periods.

• Low or Near-Zero Correlation:
This is generally considered a desirable characteristic while building a diversified portfolio, as the Algos behave more independently of one another. In such cases:

  1. They are less likely to generate profits or losses at the same time.
  2. One Algo may perform well while another goes through a drawdown, helping reduce concentration risk.
  3. The overall portfolio may experience a smoother equity curve and more balanced performance over time.

Coming to your observation, it is completely possible for a portfolio of positively correlated Algos to show higher historical returns than a portfolio of uncorrelated Algos. However, as you’ve already noticed, it also comes with higher portfolio volatility. On the other hand, introducing uncorrelated Algos may reduce the historical return over a particular period, but it generally improves diversification and can make the overall portfolio behaviour more stable across different market conditions.

Ultimately, portfolio construction is always a balance between risk and return. Rather than optimising only for the highest historical returns, it is equally important to consider how the portfolio behaves during drawdown phases and whether the overall risk aligns with your comfort level.

1 Like
3

@nijinwilson

I’m also sharing a video below that explains this concept. It covers how to build an effective Algo portfolio and how correlation should be interpreted while selecting Algos. I hope you find it helpful.