When you run a single algo, you’re essentially relying on one trader with one strategy. Algo trading gives you the ability to run multiple traders at once, each with a different approach — and that’s where the real edge shows up: in how you combine them.
Here’s the framework I use:
1. Correlation — lower is better.
Check how correlated two algos’ entries/exits are. Uncorrelated algos rarely lose on the same days — one holds or wins while the other dips, and that’s what smooths out the overall equity curve.
2. Drawdown — look at the combined portfolio, not each algo alone.
Individual algos can each carry 20%+ historical drawdown and still combine into a portfolio drawdown well under 10%, when they’re genuinely uncorrelated. Worth noting: adding an algo can also temporarily move combined drawdown up before the next addition brings it back down — so it’s worth checking at each step rather than assuming every addition helps immediately.
3. Combination — keep weights equal.
Equal capital per algo. This keeps any single algo from dominating the portfolio’s behavior, so the diversification from step 1 actually plays out.
In the video I demo this live on the algo-provider platform, using each algo’s own historical execution data to show correlation numbers and how the combined portfolio’s drawdown and returns shift as algos are added one at a time — it’s meant to show how the mechanism works, not to claim those are my own trading returns (my actual live portfolio is a separate, smaller story mentioned briefly at the top).
This is a framework, not a portfolio to copy — apply your own risk capital and drawdown tolerance to it.
Full video: [youtu.be/daj7IcuGG0M]
Curious how others here think about this — are you running single algos or combining? How do you approach correlation and weighting?